<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Orion Backgammon News</title>
    <link>https://orion.arthea.ai/news/</link>
    <description>Orion news and developer notes: how our provably fair dice work, what we are building, and what changes as we approach launch.</description>
    <language>en</language>
    <atom:link href="https://orion.arthea.ai/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Orion is out, and Season One opens 1 September</title>
      <link>https://orion.arthea.ai/news/season-one-opens-1-september/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/season-one-opens-1-september/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Orion Backgammon is out, and Season One, The First Crossing, opens 1 September 2026. What that means for trophies, the ladder and the Founding 1000.</description>
      <content:encoded><![CDATA[The game is out.

After a long build and a longer review queue, Orion Backgammon is downloadable. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now for the device you are holding, which is a more honest answer than anything we could freeze into a sentence here.

Season One, The First Crossing, opens on 1 September 2026.

When does Season One start?

1 September 2026. That is the day the ladder starts counting for the season: leagues, trophies, the season pass and the climb from Driftwood to Orion all begin together, for everyone, on the same board.

We chose a date rather than opening the season the instant the stores approved us, and the reason is fairness. A season that begins at approval begins at a different moment in every country, and the players who happened to be awake would have had a head start on a ladder that is supposed to start level. A named day removes that. Nobody is early.

What can I do before 1 September?

Everything except climb. Download the game, play real matches against real opponents, learn the doubling cube against someone who will punish you for it, visit the destinations, set up your profile and find your friends. The rules engine, the matchmaking and the provably-fair dice are all running today.

What is waiting is the seasonal layer: the ranked ladder for Season One, the pass and its tiers, the season's trophies. Those arrive on the first.

Do trophies carry over from before the season?

There is nothing to carry. Season One starts every player at zero trophies, on the same day, on the same board. That was true when we wrote it before launch and it is true now: the ladder is the record of the season, not of who installed the game first.

A ladder that starts level is the only kind worth climbing.

Is the Founding 1000 still open?

Yes, and it is the one thing that does reward arriving early. The first thousand players of the first crossing keep the Season One premium pass in full, a permanent Founder identity that stays on the account for as long as it exists, and an exclusive board made for the opening of the voyage. The board does not return to the store in a later season and it is not sold when the number is full.

It is a mark, not an advantage. Nothing a founder holds makes the dice kinder or the board easier, because a game with real competition cannot sell the competition. When the count is reached, the Founding 1000 closes for good.

How do I know the dice are fair?

You do not have to take our word for it, and you do not even have to install the game. Before every server-backed match the server publishes a SHA-256 commitment to a secret dice seed. Every roll is derived from that seed with HMAC-SHA256, and the seed is revealed when the game ends, so any roll can be recomputed afterwards and checked against the commitment.

The public verifier at orion.arthea.ai/verify runs that check entirely in your own browser, with no account and no download, and it ships with a demo replay so you can watch a real game be re-derived roll by roll. It has been live since before the game was.

What happens after the first crossing?

The voyage carries on. Twelve destinations are built and waiting, from Hassan's cafe in Tangier to the celestial table that gives the game its name, and Season One is the route that first connects them. Seasons after it will chart their own.

For now there is a board, an opponent somewhere in the world, and a date on the calendar. The door is at orion.arthea.ai/go.]]></content:encoded>
      <category>Season One</category>
      <category>Announcement</category>
      <category>Launch</category>
    </item>
    <item>
      <title>Is backgammon gambling?</title>
      <link>https://orion.arthea.ai/news/is-backgammon-gambling/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/is-backgammon-gambling/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Backgammon is a skill game played with dice, and it is only gambling when money is staked on it. What that distinction means, and how Orion is arranged.</description>
      <content:encoded><![CDATA[Backgammon is a game of skill played with dice. It becomes gambling when people stake money on the result, and it stops being gambling when they do not. The board does not decide that question. What is riding on it does.

That answer is unsatisfying only until you notice it is the same for chess, darts and golf. Any game can be wagered on. What separates a card room from a club night is not the equipment, it is whether anything of real value changes hands.

Is backgammon a game of chance or skill?

Both, in a proportion that changes with length. A single game turns heavily on the dice, and a beginner can beat a world champion in one sitting without anything strange having happened. Over a long match the dice average out and the stronger player wins almost every time, which is why serious backgammon is played as matches to a target score rather than as single games.

The skill is real and it is measurable: checker play, timing, the pip count, and above all the doubling cube, which is where most of the equity in the game actually lives. If backgammon were chance, the same names would not keep reaching the late rounds of every tournament. If it were pure skill, nobody would enter the early rounds against them. We wrote about the luck or skill question at length, because it is the most interesting thing about the game.

Is playing backgammon online for money legal?

It depends entirely on where you live, and this is not a question anyone should answer for you in a blog post. Some jurisdictions treat staked backgammon as a game of skill and permit it, some treat any dice game as gambling regardless of skill, and some draw the line at whether a house takes a cut. If you are considering a site that takes deposits, that is a question for the laws where you are.

The simpler route, and the one we took, is to remove the question. Where nothing can be deposited and nothing can be withdrawn, there is nothing to regulate as a wager.

Does Orion have real money betting?

No. There is no way to deposit money as a stake, no way to cash anything out, and no real money prizes. Orion is free to play. You can buy coins and gems the way you would buy anything in a game, and that is a purchase, not a deposit: the value moves one way and stops.

Nothing you win in Orion can leave Orion. That is the whole design, and it is deliberate.

We are not against the idea that people enjoy stakes. Backgammon has been played for something since the beginning, usually the next round of tea. But a game with real competition and real money in it has to spend its energy on policing the money, and we would rather spend ours on the game.

What are coins and gems actually for?

Coins are the table currency. Matches have buy-ins at a few levels, the stake is taken when the game starts, and the winner takes the pot. That structure is what makes a match feel like it matters, and it is why a daily allowance exists so that nobody is ever locked out of playing by a bad run.

Gems are the premium currency, for cosmetics and the season pass. Neither has any value outside the game, neither can be exchanged for money, and neither buys a better dice roll. That last point is not a matter of policy. It is a matter of arithmetic, because the dice are committed before the game starts and cannot be changed afterwards by anyone, including us.

How do I know the dice are not manipulated?

You check them yourself, which is the only answer that should satisfy anyone. Before a match the server publishes a SHA-256 commitment to a secret dice seed. Every roll is derived from that seed, and the seed is revealed when the game ends, so the whole sequence can be recomputed afterwards and compared against a commitment that was published before a single checker moved.

The verifier that runs this check is public, works in your own browser, and needs no account and no install. It ships with a demo replay so you can watch a real game be re-derived roll by roll before you ever download anything. It is at orion.arthea.ai/verify.

Is backgammon safe for younger players?

The game itself is one of the better things you can put in front of a child: it teaches counting, odds and graceful losing, and it has been taught grandparent to grandchild for centuries. The thing to check is never the game, it is the app around it. Ask whether real money can enter or leave, whether strangers can message freely, and what happens to a player who cannot pay to continue.

For our part: nothing can be cashed out, the daily allowance means a player who runs out of coins can always sit down again tomorrow, and the season pass is a set of rewards for playing rather than a way to buy a result. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>fairness</category>
      <category>rules</category>
      <category>trust</category>
    </item>
    <item>
      <title>What to check before you install a backgammon app</title>
      <link>https://orion.arthea.ai/news/choosing-a-backgammon-app/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/choosing-a-backgammon-app/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Nine things worth checking in any online backgammon app: dice fairness, what happens when you run out, ads, real money, bot detection and offline play.</description>
      <content:encoded><![CDATA[Every backgammon app on both stores says it has fair dice and a strong engine. Most say it in the same words. The store listing is not where you find out, so here is what to actually check, in the order that matters.

We build one of these, so read this as a checklist rather than as a recommendation. Every item below is something you can verify yourself, on any app, including ours.

Can you verify the dice, or only be reassured about them?

This is the one that separates claims from evidence. Most apps say the dice are fair. Some say they use a certified random number generator, which sounds stronger and is not: a certificate tests a generator in a lab, and says nothing about whether the numbers reaching your game came from it.

The version that can actually be checked is a commitment published before the game, revealed afterwards, so you can recompute the whole sequence yourself. If an app offers that, it can be caught lying, which is exactly why it is worth something. If it only offers reassurance, you are trusting a company, and you should know that is what you are doing.

What happens when you run out of coins?

Play the app until you lose a few games and find out. This is the fastest way to learn what a game actually wants from you. A daily allowance, or any reliable way back to a table, means the economy is there to give matches stakes. No way back except a purchase means the economy is there to sell you one.

Can real money enter or leave?

Buying coins is a purchase. Depositing a stake you can later withdraw is something else entirely, with different laws attached depending on where you live, and different consequences if you play more than you meant to.

Check both directions. Many apps that take no deposits still gate real competition behind a purchase, and many that look casual are one tap from a cash table. Neither is automatically wrong. You should simply know which one you have installed.

How many of your opponents are people?

Almost every app fills empty queues with bots, and almost none say so. There is nothing inherently dishonest about it: a game with nobody online is worse than a game with a good opponent. It becomes dishonest when the bot is presented as a person with a profile and a country flag while your coins go to nobody.

The tell is timing. A human hesitates unevenly, misclicks, takes eight seconds on an easy move and one second on a hard one. A queue that fills instantly at four in the morning, with opponents who always move in the same rhythm, is not a busy game.

Does it play the real rules?

Three tests catch most engines. Does it have a doubling cube, and does it let you use it properly? Does it score gammons and backgammons, or does every game count one point? Does it enforce the obscure legal-move rules, like being forced to play the larger die when only one can be played?

An app that quietly drops the cube is not playing backgammon, it is playing a race. That may be what you want, and it is worth knowing before you invest a hundred hours in it.

How does it handle a lost connection?

Phones lose signal in tunnels. The question is what happens to a game you were winning. A grace period and a resumed game is one answer. An instant forfeit, with your stake gone, is another, and on a mobile game it is a design decision rather than a technical necessity.

How much of the screen is advertising?

Check for interstitials between games rather than banners, because those are the ones that break the rhythm of playing. Then check whether the ad-free purchase is a one-time unlock or a subscription, and whether removing ads also removes the timers that were only there to sell you the removal.

Does it work offline, and does it need an account?

Two separate questions, both worth asking before a flight. An app with a local engine plays anywhere. An app that demands a sign-in before showing you a board is telling you what it values, and an app that lets you play as a guest first is telling you something else.

What do the recent one-star reviews say?

Not the average, which is bought and gamed everywhere. Sort by most recent and read the complaints. Reviews saying the game is beautiful and the dice are rigged usually mean a strong engine and a player on a bad run. Reviews saying a purchase never arrived, or an account was lost, mean something you can act on.

Then install two apps rather than one and play a few evenings on each. It costs nothing, and the difference between a good engine and a mediocre one becomes obvious faster than any list can tell you. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>fairness</category>
      <category>trust</category>
      <category>beginners</category>
    </item>
    <item>
      <title>Backgammon setup: where every checker goes</title>
      <link>https://orion.arthea.ai/news/backgammon-setup/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-setup/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>How to set up a backgammon board: the exact starting position, which way each player moves, the ten second check that proves it, and four common mistakes.</description>
      <content:encoded><![CDATA[The backgammon setup never changes. Whichever side of the board you sit on and whichever colour you play, fifteen checkers go onto exactly four points at the start of every game, and your opponent mirrors you. Learn it once and you will never think about it again.

Below is the position itself, then the check that tells you in ten seconds whether the board is right, then the four mistakes that account for almost every board that is wrong.

What is the backgammon starting position?

Number the points from your own home board outwards, so the point nearest your bottom right is 1 and the far corner is 24. Your checkers start like this:

Two checkers on the 24 point, your furthest point, sitting inside your opponent's home board. Five checkers on the 13 point, the long midpoint across the bar. Three checkers on the 8 point. Five checkers on the 6 point, inside your own home board.

Your opponent has the exact mirror image, so their two far checkers sit on the point you call 1, and their five midpoint checkers face yours across the board. Every game of backgammon in the world begins from this position, and it is the same in tavla, tavli, tawla, shesh besh and nard.

How do you check the board is set up correctly?

Count to fifteen. Two plus five plus three plus five is fifteen, and that is the only arithmetic the setup needs. If your four stacks do not add to fifteen, something is in the wrong place before a single die has been thrown.

Then run three shape checks, which catch the errors counting does not:

The two tallest stacks are five high, and they sit on opposite sides of the bar, six points apart on your side and thirteen on the far side. Your two loneliest checkers are as far from home as they can be, and they are staring at your opponent's five stack. Half the board is empty. Twenty of the twenty four points hold nothing at the start, and if the position looks crowded it is wrong.

Which direction does each player move?

Towards your own home board, and the two players move in opposite directions. Your checkers travel from 24 down to 1 and then off the board. Your opponent's travel the other way around the same circuit, which is why the two back checkers of each player have the furthest journey and meet the heaviest traffic.

Your home board is the quadrant nearest you, the six points where you will eventually bear off. It is not the quadrant you happen to be facing when you sit down, and swapping those two is the most common reason a beginner's first few moves feel wrong.

Twenty of the twenty four points start empty. Backgammon is a game about the space, not the checkers.

What are the most common setup mistakes?

Four, and all four are visible before the first roll if you look.

Swapping the 8 point and the 6 point. Three checkers belong on the 8 and five on the 6, not the other way around. This is by far the most frequent error, and it quietly ruins the opening. Setting up mirrored on the same side. Both players' five stacks should face each other across the bar, not sit side by side. Putting the back checkers on the 23 point instead of the 24. They belong on the very last point, inside the opponent's home board. Playing the wrong direction. Point at the quadrant where you intend to bear off before the first roll, and make sure your opponent points at the opposite one.

There is a fifth that is not a setup error but costs the same: not agreeing whether the doubling cube is in play. Set it to 64 in the middle of the bar if it is, and put it away if it is not.

Is the setup the same in every backgammon variant?

No, and this is worth knowing before you sit down with someone who learned the game elsewhere. The position above is standard backgammon, the game a Greek player calls portes and a Turkish player calls tavla. The other games played on the same board start differently: in plakoto and fevga all fifteen checkers begin stacked on a single point, and in fevga both players even run in the same direction.

If your opponent stacks all fifteen checkers in one corner, they have not made a mistake. They are setting up a different game, and the two of you should agree which one you are playing.

What comes after the setup?

Both players roll one die, and whoever rolls higher goes first, playing the two numbers that are already on the table. From there the rules are short enough to learn in a single evening, and the full walkthrough lives on the rules page.

On a digital board none of the above can go wrong. Orion sets the position for you, nothing slides when someone knocks the table, and the direction of travel is obvious from the first move. What is left is the roll, and because you cannot watch the hand that throws it, the dice are provably fair: the server commits to the whole sequence with SHA-256 before the game begins and reveals the seed when it ends, so anyone can recompute every roll at orion.arthea.ai/verify, with no account.]]></content:encoded>
      <category>rules</category>
      <category>beginners</category>
      <category>setup</category>
    </item>
    <item>
      <title>Bearing off in backgammon: rules and technique</title>
      <link>https://orion.arthea.ai/news/bearing-off-in-backgammon/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/bearing-off-in-backgammon/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>When you may start bearing off, the exact rule for a roll higher than your highest checker, what happens if you are hit, and how to avoid wasting pips.</description>
      <content:encoded><![CDATA[Bearing off is the last phase of a backgammon game: taking your checkers off the board entirely. It looks like the part where nothing can go wrong, which is exactly why more won games are lost here than anywhere else.

When can I start bearing off?

Only when all fifteen of your checkers have reached your home board, the six points nearest you. Not fourteen. If a single checker is still outside, or sitting on the bar, you may not take anything off until it is home.

That is the rule most often broken by accident, and it has a sharp consequence: if you are hit after bearing off has begun, the checker goes to the bar and has to re-enter and travel the whole way back before you may take another one off. A game that was over can restart.

What are the exact bearing off rules?

You bear off a checker from the point matching the number you rolled. A five takes a checker off the five point, a two off the two point. From there, two cases cover everything else.

If the matching point is empty and you have checkers on HIGHER points: you may not bear off with that number. You must play it as a move inside the home board instead. Rolling a five with the five point empty but checkers on the six means moving one from the six to the one.

If the matching point is empty and there is nothing higher: you bear off from the highest occupied point. Rolling a six with your back checkers on the four means a checker comes off the four point.

Doubles work as they always do: four numbers, so up to four checkers off in one turn. And you are never forced to bear off when a legal move exists instead, which is the whole basis of playing a safe bear-off against contact.

What happens if I get hit while bearing off?

The checker goes to the bar, re-enters in the opponent's home board, and has to run the full length of the board again. Everything you had taken off stays off, but you cannot bear off another checker until all fifteen remaining ones are home.

This is why the phrase used at the table is that you are never safe until the last checker is off. If your opponent still holds an anchor in your home board, you are in a contact bear-off and the shape of your position matters far more than speed.

How do I avoid wasting pips?

Wastage is what happens when a roll takes a checker off a low point that a smaller number could have handled. Bearing a checker off the one point with a six spends five pips for nothing, and in a close race those pips are the game.

Keep the position even. Filling gaps and avoiding tall stacks on the low points means fewer rolls are forced to overkill. Clear from the back. Taking checkers from the higher points first keeps your options open and avoids ending with a lone checker on the six. Do not create gaps you will have to jump. An empty point with checkers behind it turns future rolls into moves rather than bear-offs. Count before you commit. In a pure race, the pip count tells you whether you are playing for a win or playing to save a gammon, and those need different bear-offs.

What is the safest way to bear off against contact?

Differently from a pure race, and this is the distinction that costs beginners the most. When the opponent holds an anchor, your goal stops being speed and becomes never leaving a blot. Clear your points from the rear, keep an even number of checkers on the points the opponent is watching, and take the slower play that leaves nothing exposed.

The uncomfortable version of that advice is that you should sometimes decline to bear a checker off at all, playing a quiet move inside the home board instead, because taking one off would break a point and hand over a shot. Losing three pips is cheaper than losing the game.

Does bearing off decide gammons?

It decides whether one happens. If your opponent has not borne off a single checker when you finish, the game counts double as a gammon, and if they still have a checker in your home board or on the bar it counts triple as a backgammon. Those multipliers make the bear-off worth playing carefully even from a hopeless position.

So the losing player's job in the last phase is real: get one checker off, and the loss is halved. Our note on how scoring really works goes through what that is worth alongside the cube, and it is worth more than most players assume. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>rules</category>
      <category>strategy</category>
      <category>beginners</category>
    </item>
    <item>
      <title>The four backgammon game plans, and how to spot yours</title>
      <link>https://orion.arthea.ai/news/backgammon-game-plans/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-game-plans/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Priming, the blitz, the holding game and the back game: the four shapes a backgammon game settles into, what each one wants, and how to recognise yours.</description>
      <content:encoded><![CDATA[Beginners play backgammon one roll at a time. Stronger players are playing a plan, and most of the difference between them is not calculation. It is that the stronger player knows which of four games is actually on the board and stops making moves that belong to a different one.

There are four, and almost every position is one of them or a transition between two.

What is a priming game?

Building a wall of consecutive made points in front of the opponent's back checkers. A checker cannot move more than six pips, so six points in a row is an absolute prison: nothing behind it can get out until the wall breaks.

Priming is the most reliable plan in the game because it does not depend on hitting anything. You are not asking the dice for a shot, you are asking them for points, and there are many more ways to make a point than to hit a specific blot. The cost is time: while you build, your opponent is building too, and a prime that arrives one roll late is a wall around an empty room.

What is a blitz?

The opposite temperament. You hit the opponent's blot, and instead of consolidating you keep attacking, making home board points while their checkers sit on the bar unable to answer. The aim is to close your home board entirely so that a checker on the bar cannot re-enter at all.

A blitz that lands wins quickly and often wins a gammon, because the opponent spends the whole game on the bar and bears nothing off. A blitz that stalls is the worst position in backgammon: blots everywhere, no structure, and an opponent who now gets to hit you back. Committing to one is a real decision, not an impulse after a lucky hit.

What is a holding game?

You are behind in the race, so you keep an anchor on a high point, usually the opponent's four or five point or the midpoint, and you wait. You are not trying to win the race. You are trying to still be there when the opponent has to break their position and give you a shot.

The holding game is patience as a technique. The mistake is impatience in both directions: running the anchor too early, which throws away the shot you were waiting for, or holding it far too long past the point where the shot can no longer help you. The pip count is what tells you which, and this is the phase where counting stops being an exercise and starts being the whole decision.

What is a back game?

The plan of last resort, and the hardest to play well. You hold two anchors deep in the opponent's home board, usually two low points, and you deliberately fall further behind in the race so that the opponent must eventually break down in front of you and leave a shot you can hit.

Back games have a reputation for glamour because when they work they win from nowhere. They mostly do not work. They need the right two points, they need timing you cannot always control, and if you hit that shot without a home board ready to contain the checker you have simply delayed the loss. Play one when the alternative is worse, not because it looks clever.

How do I tell which plan I am in?

Two questions answer it almost every time. First, who is ahead in the race? Second, does either side have checkers trapped behind the other's points?

Ahead in the race, opponent has checkers back: prime them, or race home if the way is clear. Ahead in the race, no contact left: it is a pure race, so count and run. Behind in the race, you hold a high anchor: holding game, and wait for the shot. Behind by a lot, you hold two deep anchors: back game, and manage your timing above all else. Even race, opponent leaves a blot in their home board early: a blitz is on the table if you can sustain it.

Can a plan change mid-game?

Constantly, and refusing to notice is the single most common intermediate error. A blitz that fails to close the board becomes a priming game. A priming game whose wall breaks becomes a holding game. A holding game that loses its anchor becomes a race you are behind in, and the correct response to that is usually to accept it and play the race rather than to keep hunting a shot that is no longer there.

This is also where the doubling cube belongs in the discussion, because each plan has its own moment of maximum equity, and that moment is usually earlier than it feels. A prime completing, a blitz landing, an anchor about to be forced: those are cube decisions, and recognising your plan is what lets you see them coming. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>strategy</category>
      <category>intermediate</category>
    </item>
    <item>
      <title>Match play and the Crawford rule, explained</title>
      <link>https://orion.arthea.ai/news/match-play-and-the-crawford-rule/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/match-play-and-the-crawford-rule/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Why serious backgammon is played as matches, what the Crawford rule is and why it exists, what post-Crawford means, and how the score changes the cube.</description>
      <content:encoded><![CDATA[A single game of backgammon is a coin toss with opinions. The dice decide too much of it for one result to mean anything, which is why competitive backgammon is not played in games at all. It is played in matches, first to a set number of points, and the match score changes almost every decision inside it.

How does backgammon match play work?

A match is played to a target: first to five points, to seven, to eleven, to twenty-five in a final. Each game is worth one point normally, two if it ends in a gammon, three in a backgammon, and all of that is multiplied by the doubling cube.

So a game played to four on the cube that ends in a gammon is worth eight points, which settles most matches on its own. This is the first thing to understand about matches: they are usually not decided by many small games, they are decided by one large one.

What is the Crawford rule?

When either player first reaches match point, meaning they need exactly one more point to win, the very next game is played WITHOUT the doubling cube. That single game is the Crawford game. After it, the cube returns for the rest of the match.

One game, once per match, on the first occasion the leader reaches match point. It is not applied again if the score later reaches match point in some other way, and it is standard in essentially every tournament in the world.

Why does the Crawford rule exist?

Because without it the trailing player has a free lever. Suppose the score is four to two in a match to five. The leader needs one point; the trailer needs three. If the trailer may double at the start of every game, they should, immediately and always, because a doubled game they lose costs them nothing they had not already lost, and a doubled game they win closes the gap twice as fast.

That is not a strategy, it is an exploit, and it makes the endgame of every match identical. Crawford removes it by taking the cube away for exactly one game, so the leader gets one clean chance to finish.

Crawford is the rule that stops a lost match from becoming a free lottery ticket.

What does post-Crawford mean?

Every game after the Crawford game, while the leader is still at match point. The cube is back, and now the logic that Crawford suppressed applies openly: the trailing player usually doubles at the first opportunity, because the match is over anyway if they lose.

There is one famous exception, and it is worth knowing because it is the sort of thing that separates a club player from a tournament player. When the trailer needs an even number of points, holding the cube for one roll can be right, since the leader may throw a set that lets the trailer double into a more favourable position. The general shape is simple: double immediately when you need an odd number, think when you need an even one.

How does the match score change my doubling?

Enormously, and this is the real reason match play is a different game from money play. In a money game every point is worth the same. In a match, points above what you need to win are worth nothing at all.

Leading by a lot, take fewer risks: a gammon you do not need cannot help you, and a doubled loss can undo the lead. Trailing by a lot, take more: variance is your friend when the safe path loses slowly and certainly. Near the end, count what you NEED rather than what you would gain. A two-point win and a four-point win are identical when two points takes the match. Beware the free drop: post-Crawford, when the trailer needs an even number, the leader can sometimes decline a double at no cost at all.

Serious players study this through match equity tables, which give the probability of winning from every score. You do not need to memorise one to benefit. Simply asking how many points would actually change the match, before you touch the cube, removes most of the errors.

Do I need match play to enjoy the game?

No. Single games are how most backgammon has always been played, over tea, one after another, and the doubling cube alone gives them plenty to argue about. Match play is what you reach for when you want the result to mean something, because it is the format in which skill reliably beats luck.

That is also why a ranked ladder is built the way it is: many results, rated, so that a season measures a player rather than an evening. Our note on how the leagues and trophies work goes through that side of it. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>rules</category>
      <category>doubling cube</category>
      <category>strategy</category>
    </item>
    <item>
      <title>Is backgammon good for your brain? An honest answer</title>
      <link>https://orion.arthea.ai/news/is-backgammon-good-for-your-brain/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/is-backgammon-good-for-your-brain/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>What the research on board games and cognition actually supports, what it does not, and which specific mental skills backgammon genuinely exercises.</description>
      <content:encoded><![CDATA[Search this question and you will be told, confidently and in a dozen places, that backgammon prevents dementia. That claim is not supported, and repeating it does the game a disservice, because what is actually true about backgammon and thinking is more interesting than the overclaim.

Here is the careful version.

Does playing backgammon make you smarter?

Not in the sense of raising general intelligence. What practice reliably improves is the thing practised, and the effects transfer much less widely than people hope. A chess player is not better at arithmetic because of chess; they are better at chess.

So the honest framing is narrower and more useful: backgammon gives sustained practice in a handful of specific mental operations, and those are worth naming individually rather than bundling them into a claim about the brain.

What mental skills does backgammon actually exercise?

Working memory, by holding a position and a plan in mind while calculating a move against it. Rapid arithmetic under pressure. The pip count is mental addition performed dozens of times per game, at speed, while someone waits. Probability estimation. There are 36 dice outcomes and you learn to feel the shape of them: how many rolls hit from six pips away, how many escape a prime, how a shot changes when a point is made. Decision-making under uncertainty, which is the real one. You choose without knowing the outcome, and then you find out whether the dice agreed. Few everyday activities give that feedback loop so quickly or so often. Emotional regulation, because you will lose games you played perfectly, and the game asks you to keep playing well anyway.

The last two are why backgammon is studied by people interested in judgment rather than by people interested in puzzles. Chess punishes calculation errors. Backgammon punishes bad process while sometimes rewarding it anyway, which is a far better model of most real decisions and a far harder lesson.

Do board games prevent dementia?

No study shows that, and no honest source will tell you otherwise. What exists is observational research finding that people who regularly do cognitively engaging activities, including board games, tend to show cognitive decline later than people who do not.

That is an association, and associations run in more than one direction. People whose cognition is already declining play fewer games, which produces exactly the same pattern in the data without the games having caused anything. Untangling that requires trials that are difficult and expensive to run, and the ones that exist are modest.

The safe version is worth saying plainly: playing backgammon is good for you the way a long walk with a friend is good for you.

Which is not nothing. Engagement, attention and company are genuinely associated with better outcomes as people age. It simply is not a treatment, and a game that has survived five thousand years does not need to be sold as one.

Is backgammon better for the brain than chess?

They are different exercises, and the comparison usually smuggles in a value judgment that neither deserves. Chess is a perfect-information game: everything is on the board and the discipline is calculation. Backgammon has hidden future information in the dice, so the discipline is estimating and committing.

Chess trains you to find the best move. Backgammon trains you to make the best decision and then live with the result, which is a genuinely different skill and the one most people use more often. Our note on whether the game is luck or skill goes into why both are true at once.

How long does it take to feel a difference?

For the concrete skills, quickly. Most players can hold a rough pip count within a few weeks of trying, and the sense for which rolls hit arrives faster than that, because the same situations recur constantly and the feedback is immediate.

For anything broader, nobody can tell you honestly, and be suspicious of anyone who does. The reasonable expectation is that you will get noticeably better at backgammon, that this will be satisfying, and that the arithmetic will stop feeling like arithmetic.

What is the best way to play for the mental exercise?

Against real opponents, in matches rather than single games, and with the doubling cube. The cube is where most of the thinking lives: it forces you to put a number on your own position and act on it, which is the single most transferable habit the game teaches.

And review the games you lost badly, which is the part almost nobody does. A game you can replay afterwards is a game you can learn from, and that is worth more than a hundred casual sessions. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>culture</category>
      <category>beginners</category>
    </item>
    <item>
      <title>Teaching a child backgammon, in the right order</title>
      <link>https://orion.arthea.ai/news/teaching-a-child-backgammon/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/teaching-a-child-backgammon/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Which rules to leave out at first, the order to add them back, what age it clicks, and how to handle the losing so a child wants a second game.</description>
      <content:encoded><![CDATA[Backgammon is one of the few serious games a child can beat an adult at honestly, on the first evening, without anyone pretending. That is the whole reason it works as a game to teach, and it is also the thing most explanations bury under the rules.

The difficulty is not the rules. It is the order you give them in.

What age can a child learn backgammon?

Around six or seven for the movement, once counting to twelve is comfortable and reliable. Around eight to ten for the shape of the game: why you would make a point, why leaving one checker alone is a risk. The doubling cube usually lands somewhere past ten, and there is no hurry at all.

Those ages are a rough guide and children vary enormously. The real signal is whether they are counting the pips out loud without being asked. When that starts, they are ready for more rules.

Which rules should I leave out at first?

Almost all of them. A first game should be moving checkers around the board and taking them off, and nothing else. Three things to withhold, in this order of importance.

The doubling cube. Leave it in the box entirely. It is the deepest part of the game and it is meaningless before a child can judge a position. Gammons and backgammons. Every game is worth one, and the game ends when somebody is done. Doubling the loss teaches nothing to someone still learning which direction their checkers go. Hitting, for the first game or two only. Let a checker land on a single checker without sending it home. This one comes back quickly, because hitting is most of the fun, but a first game where your pieces keep getting sent to the start is a first game that ends in tears.

What you keep from the start is the direction of travel, the roll and move, doubles giving four moves, and bearing off. That is a complete, winnable game, and it is enough for an evening.

What order should I add the rules back in?

Game two or three: add hitting and the bar. This is the biggest single jump in the game's depth and it explains itself in one move. A child who has been hit once understands blots forever.

Once hitting is comfortable: add making points: two checkers on a point mean the opponent cannot land there. Now the board has structure, and the idea of blocking arrives on its own.

When they start planning ahead: add gammons. The loss doubling if you have not borne a checker off gives the losing player something to fight for, which matters more than it sounds.

Much later, and only if they ask: add the doubling cube. Introduce it as a bet on the position rather than as a rule, and expect it to take several sessions before the take-or-drop decision means anything.

Should I let a child win?

No, and backgammon is the game where you do not have to, which is exactly its advantage over chess. The dice do the work. A beginner beats a strong player often enough that no charity is needed, and children detect charity immediately and resent it.

What you can do instead is play your own moves openly and explain them as you go. Saying why you are making a point teaches more than losing on purpose, and it keeps the game honest, which is the thing they will remember about learning it.

How do I handle the losing?

By naming the dice out loud when they decide something, in both directions. If a child loses to a lucky double, say so. If they win by one, say that too. What you are teaching is the distinction between a bad decision and a bad outcome, and backgammon separates those more cleanly than almost anything else a child will encounter.

Play the roll you got, not the one you wanted. It is the whole game, and it is a decent thing to learn at nine.

Keep the games short. A child's first sessions should end while they still want another one, which means stopping earlier than feels natural.

Is an app a good way for a child to learn?

For the rules, yes, and better than a person in one specific way: an app will not let an illegal move happen, so nobody has to be corrected. Learning which moves are even possible is faster against software that simply will not accept the wrong one.

For everything else, a real board and a real person is better, because the part worth passing on is not the rules. If you do use an app, check what happens when the coins run out and whether real money can enter or leave, which we wrote a checklist about. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>beginners</category>
      <category>culture</category>
      <category>rules</category>
    </item>
    <item>
      <title>Backgammon variants: plakoto, fevga and acey-deucey</title>
      <link>https://orion.arthea.ai/news/backgammon-variants-around-the-world/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-variants-around-the-world/</guid>
      <pubDate>Wed, 19 Aug 2026 00:00:00 GMT</pubDate>
      <description>Portes, plakoto, fevga, acey-deucey, nackgammon and narde: how the world&apos;s backgammon variants differ, and which one the international game is.</description>
      <content:encoded><![CDATA[Most people who say they play backgammon play one of a dozen related games. The board is the same, the checkers are the same, and then the rules fork: in one version you hit an enemy checker, in another you pin it, and in a third you never touch it at all.

Knowing which one is on the table saves an argument, and the differences are genuinely interesting, because each variant is a different answer to the same question about how much aggression a game should allow.

What is the standard international game?

The one played in tournaments, and the one the word backgammon means without qualification. Fifteen checkers each, opposing directions of travel, a lone checker can be hit and sent to the bar, and the doubling cube governs the stakes. Every variant below is best understood as this game with one rule changed.

It is also the game Orion plays. We considered offering the regional variants and decided that one game played by everyone, on a single ladder, was worth more than several smaller pools that could never meet.

What are the three Greek games?

In Greece the board holds three games, not one, and tavli means the set of them. They are played in rotation, usually best of five or seven, and a player who is good at one is not automatically good at the next. That rotation is the point: it rewards range.

Portes, which is the standard game: opposing directions, hitting allowed, the version an international player sits down to without needing anything explained. Plakoto, where a lone checker is not hit but PINNED: your checker lands on top of it and it cannot move until you leave. Trapping the opponent's back checker on their own one point can decide the game in the opening minutes. Fevga, where nobody is ever hit and both players travel in the SAME direction. It becomes a pure race for space, and a single checker is enough to block a point.

Plakoto is the one that surprises people most. Because a pinned checker is immobilised rather than sent home, a bad early position is not a setback you can recover from by re-entering. It simply sits there, and you play the rest of the game a checker short.

What is acey-deucey?

An American variant with a naval history, popular in the United States services through the twentieth century. Checkers start OFF the board and enter as they would from the bar, which turns the opening into its own contest, and a roll of one and two is special: you play the one and the two, then choose any doubles you like, then roll again.

The effect is a game with a much wilder swing than the standard one. That extra roll after a chosen double can undo a comfortable lead in a single turn, which is either the appeal or the problem depending on who you ask.

What are nackgammon and hypergammon?

Both are deliberate experiments with the starting position rather than the rules.

Nackgammon, devised by Nack Ballard, moves a checker so that each side starts with two checkers deep in the opponent's home board instead of one. Games last longer, back games are far more common, and the opening theory of the standard game mostly does not apply. Hypergammon gives each player three checkers instead of fifteen. It is short, sharp, and completely solved: the perfect play from any position is known, which makes it a strange and useful thing to study.

What is narde, and why is there no hitting?

Narde, played across Russia, the Caucasus and Central Asia, starts with all fifteen checkers stacked on a single point and forbids hitting entirely. Like fevga, it is a race decided by blocking rather than by contact, and like fevga it demands a completely different kind of patience from the international game.

The family resemblance across all of these is the clue to how old the game is. When a design survives five thousand years and crosses that many borders, it does not arrive anywhere unchanged. Each region kept the board and rewrote the one rule it cared most about.

Which variant should I learn first?

The standard game, without hesitation. It has the deepest published theory, the largest pool of opponents, and every concept it teaches transfers: the pip count, timing, the value of a made point, and how to use the doubling cube. Learn it first and plakoto will take you an afternoon.

Learn a regional variant first and the reverse is not true, because the standard game asks you to weigh a risk that fevga never presents and plakoto resolves differently. Start where the theory is, then go and enjoy the rest. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>rules</category>
      <category>culture</category>
      <category>tavla</category>
    </item>
    <item>
      <title>Can online backgammon be rigged?</title>
      <link>https://orion.arthea.ai/news/can-online-backgammon-be-rigged/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/can-online-backgammon-be-rigged/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Yes, it can be. A closed server can rig dice and results alone cannot prove it. What rigging looks like, and how commit-reveal dice make it detectable.</description>
      <content:encoded><![CDATA[Yes, it can be. Online dice are rolled on a server you cannot see, and nothing physically stops an operator from changing a number before it reaches your board. Results alone will never prove it either way, so the honest question is not whether rigging is possible, it is whether a game can prove it did not happen.

We make a backgammon game, so we are not neutral here. That is exactly why we want to answer this properly, including the uncomfortable parts.

Why do the dice feel rigged on every site?

Because backgammon variance is brutal and human memory is biased. There are thirty-six ways two dice can land, streaks are not a glitch, they are a guarantee, and somewhere tonight an honest server will roll four doubles in a row against someone who played every checker correctly. You remember the 5-5 that hit your blot. You do not remember the miracle shot that saved you last Tuesday. Losses stick, escapes evaporate, and after three bad games in a row the dice feel personal. They feel personal over a real board in a real cafe too. Ask anyone who has pushed wooden checkers for forty years.

Here is the trap. A rigged game would feel exactly the same. Streaks prove nothing in either direction, so if your evidence is how the dice felt, you cannot tell an honest server from a dishonest one. Neither can we. That is not a defense of online dice. That is the problem with them.

Could a site actually rig its dice?

Technically, yes, and it would not even be hard. The server sees the whole board before it rolls. An operator who wanted to cheat could:

Tip a few rolls per hundred at the moments they matter most: the shot that hits, the entry that fails, the double that swings a lost game. Keep matches artificially close, so sessions run longer and losses feel like near misses. Quietly favor one side of a game the operator has an interest in. Do it rarely enough that the dice still pass every statistical test a player could ever run.

We are not claiming any particular site does this. We are saying the architecture allows it. A closed server is a locked room: the operator rolls in private, reports a number, and asks you to take it on trust. From outside the door, honest and dishonest operators look identical.

Doesn't a certified RNG prove the dice are fair?

Read what the certificate actually covers. A lab tests a generator: an algorithm on a bench, producing millions of numbers that pass statistical checks. That certifies the component. It says nothing about delivery, meaning whether the number that reached your board came from that generator, at that moment, unmodified by anything in between. A certified scale in the back room does not prove your bread was weighed on it.

Certification is also a snapshot. Live servers redeploy all the time, and the code path between the generator and your game keeps changing long after the lab has gone home. So certified RNG answers a question nobody was really asking. The question that matters is not whether good dice exist somewhere on the premises. It is whether your game got them, every roll, tonight.

What does commit-reveal actually change?

It moves fairness from a promise to a proof. Before a game starts, the server draws a secret seed and publishes its fingerprint, a SHA-256 hash called the commitment. Every roll in the game is then derived from that seed, one HMAC-SHA256 computation per roll, in order. When the game ends, the server reveals the seed.

Now anyone can check two things. Hash the revealed seed: it must match the commitment published before the first move. Recompute the rolls: they must match every number that appeared on the board. If the server had altered even one die after seeing your position, the math would not line up. Rigging stops being a suspicion you cannot test and becomes a lie you can catch.

Dice that were locked before the game cannot react to it.

That is the heart of it. The useful kind of rigging is reactive: see the blot, roll the hitter. Commit-reveal takes the server's eyes away, because the entire sequence was fixed before the opening move and any tampering after the commitment is mathematically detectable. What it does not do is make the dice kind. The sequence was honest, not merciful. You will still sit on the bar at the worst possible moment, and now you will be able to prove it was fair.

How do you verify a game in practice?

This is how we built it in Orion. Every game played on our servers carries a commitment made before the first roll, the server validates every live move at source, and checking a finished game takes about a minute:

Open the verifier: Go to orion.arthea.ai/verify in any browser. Nothing to install, no account to create.

Enter the game ID: Every server game in Orion has one. Paste it in.

Let your browser do the math: The page recomputes the commitment and every roll using Web Crypto, locally, in your browser. Our word is not part of the check. If a single number had been changed, the recomputation would say so.

The verifier is live today. The game itself is finished, and the door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. Either way, the honest answer stands. Yes, online backgammon can be rigged. The fix is not a nicer promise. It is dice that show their work.]]></content:encoded>
      <category>fairness</category>
      <category>provably fair</category>
      <category>dice</category>
    </item>
    <item>
      <title>How to verify your backgammon dice were fair</title>
      <link>https://orion.arthea.ai/news/how-to-verify-dice-rolls/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/how-to-verify-dice-rolls/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Every Orion game commits to its dice with SHA-256 before the first roll. Paste the game ID at orion.arthea.ai/verify and your browser rechecks every roll.</description>
      <content:encoded><![CDATA[You verify backgammon dice with a promise made before the game and checked after it. Orion publishes a SHA-256 commitment to a secret seed before your first roll, reveals the seed when the game ends, and lets your browser recompute every roll at orion.arthea.ai/verify with just the game ID. If everything matches, the dice were fixed before play began and nobody, including us, steered them.

Every online player knows the feeling. You dance on the bar for three turns. Your opponent finds the one number that hits and covers. On a wooden board you would shrug, because you watched the dice leave the cup. Online, you are asked to trust a number you never saw generated. We did not want to ask for that trust. We wanted to make it checkable.

What does commit-reveal actually mean?

Picture a sealed envelope. Before you sit down, the house writes the whole evening's dice inside, seals it, and hands you a photograph of the seal. When play ends, the envelope is opened. If the contents match the seal, the rolls existed before the game did, and nothing could have been picked mid-game to hurt you. Commit-reveal is that envelope, built from arithmetic. The seal is a SHA-256 hash, and no one can produce a different seed that matches it. Four steps carry the proof from our server to your browser.

Before the game: The server draws a secret random seed and hashes it with SHA-256. That hash is the commitment, and it is attached to your game before the first roll. From that moment the whole game's dice are already determined. The server has published a promise it can no longer edit.

During the game: Every roll is derived from the seed with HMAC-SHA256: seed and roll number in, two dice out. The position, the score, the cube, who is about to win: none of it enters the computation. The server is not choosing rolls. It is reading out a list it committed to before you played.

When the game ends: The seed is revealed and stored with the game record. This is the envelope opening. Anyone can hash the revealed seed and compare it with the commitment that was published before the first move, and the two must be identical.

Verify it yourself: Open orion.arthea.ai/verify in any browser and enter the game ID. The page recomputes everything locally with Web Crypto, the cryptography built into the browser itself: first the SHA-256 of the revealed seed against the original commitment, then every roll of the game, die by die. No account, no app, about a minute.

A full match establishes three facts at once:

The revealed seed hashes to the exact commitment published before the first roll Every roll of the game derives from that seed by HMAC-SHA256, so nothing was rerolled or swapped The dice were never adapted to the position, the score, or the player

What would a mismatch prove?

A mismatch would not be a display bug. It would be evidence. If the revealed seed does not hash to the commitment, the dice were changed after the game began. If a single roll fails its HMAC check, the dice you saw were not the dice that were committed. Either failure is proof of tampering that anyone can reproduce from the same record, on any machine, forever.

This is also why you should never see one. The commitment is published first, before a single checker moves. A house that wanted to rig this game would have to publish the evidence of its own rigging in advance, then hand every player the tools to find it. We do not ask you to trust that we are honest. We built the system so you do not have to.

Nobody can aim a roll that was sealed before the game began.

If the dice are provably fair, why do I still get bad rolls?

Because fair dice are still dice. You will get closed out on the bar. You will need anything but a 2-1 and roll the 2-1. Over thousands of games the numbers flatten toward the true distribution, but one evening promises nothing. What verification removes is the question underneath the tilt. That miracle shot was fixed before either of you touched a checker, and it was not aimed at you. Knowing that will not soften the roll. It does let you stop auditing the universe and get back to playing the position.

Dice are the loudest suspicion, but not the only one. The same server that commits to your rolls also validates every live move at source, so a modified client cannot slide a checker where the rules do not allow it. Fair rolls and legal moves come from the same authority.

Can I try the verifier before the game is out?

Yes. The verifier is live today at orion.arthea.ai/verify, and it includes a demo replay, so you can watch a complete verification run on a sample game, commitment to seed to every roll, before you have a game of your own. For Orion itself, the door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.

Fair dice are the floor the whole voyage stands on. Twelve destinations, ten leagues, one rule underneath everything: when a game ends, you never have to take our word for it.]]></content:encoded>
      <category>provably fair</category>
      <category>dice</category>
      <category>guide</category>
    </item>
    <item>
      <title>Provably fair vs certified RNG: what protects you</title>
      <link>https://orion.arthea.ai/news/provably-fair-vs-certified-rng/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/provably-fair-vs-certified-rng/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Certified RNG asks you to trust an audit. Provably fair hands you a per-game receipt you can check yourself. Here is the difference, in plain terms.</description>
      <content:encoded><![CDATA[Certified RNG means an independent lab examined a random number generator at a point in time and judged its output statistically sound. Provably fair means the game hands every player a cryptographic receipt for their own match, so each roll can be rechecked without trusting anyone. One model audits the machine. The other proves your game.

Both phrases appear on the trust pages of online games that roll dice for you, and they sound interchangeable. They are not. They answer different questions, they fail in different ways, and the difference decides what you are being asked to believe when you sit down to play.

What does certified RNG actually certify?

A certification lab takes a generator, runs its output through statistical test batteries, and hunts for bias: numbers that arrive too often, patterns that repeat, sequences that drift. If millions of samples look indistinguishable from chance, the lab signs a certificate. That work is real and it takes skill. It establishes that the generator, as tested, on that day, produced sound randomness.

Notice what the certificate cannot see. It was issued in a lab, and your dice are rolled on a server.

Whether the code tested is the code deployed, months and many updates later. The wiring between the generator and your game: the code that turns a raw random number into your 6 and 4. Any individual roll in any individual match. Certification speaks about the generator in general, never about your Tuesday night game.

So certification carries a quiet second clause: trust that nothing changed since the test, and trust that the certified path is the one your game actually took. Most studios honor it. The point is that you cannot check. The model is an inspection of the machine at one moment, followed by faith in the deployment at every moment after.

What does provably fair mean?

Provably fair is a different bargain. Instead of pointing at an audit you never saw, the game commits to its randomness before play begins, then reveals enough afterward for you to recompute every roll yourself. The proof travels with the match. In practice it works like a sealed envelope.

Commit: Before the first move, the server generates a secret seed and publishes its SHA-256 fingerprint. The fingerprint reveals nothing about the rolls to come, but it pins the seed in place. Change a single bit of the seed later and the fingerprint no longer matches.

Roll: During play, each roll is derived from the seed with HMAC-SHA256, one computation per roll. The dice were decided the moment the envelope was sealed. A friendlier number cannot be slipped in mid-game without breaking the commitment already published.

Reveal: After the game, the seed is revealed. Anyone can hash it, compare against the commitment, recompute every roll, and set the results next to the dice that appeared on the board. If it all lines up, the game could not have been steered. If anything was tampered with, the math says so.

The trust moves from the company to the arithmetic. You no longer need to believe the audit was thorough or the deployment was faithful. You need SHA-256 to hold, which is the same assumption the rest of the internet leans on every day.

A certificate says the coin was fair the day we looked. A receipt says your flip was fair, and shows the math.

Does certification still matter?

Yes. Commit-reveal proves you received exactly the randomness that was committed. It does not prove that what was committed was good in the first place. A weak or predictable seed source would still produce receipts that validate perfectly: provable delivery of poor dice. Judging the quality of the source is exactly what audits and statistical testing are for, and in regulated industries a certificate is simply required.

The two models answer different questions. An audit asks: is the randomness any good? A receipt asks: did my game actually use it? A system with only the first runs on permanent faith in deployment. A system with only the second could faithfully deliver bad dice and prove it. The strongest systems close both doors.

Why does a public verifier matter?

Because a receipt no one can read protects no one. Some games publish commitments but give players no practical way to check them, or tuck verification inside the very app in question. That closes no loop. If the only tool that can vouch for a system belongs to the system, you are back to faith, now with extra steps.

The strongest setups publish three things. The commitment, before the game. The seed, after it. And an independent, public way to recompute the rolls: a plain web page, standard cryptography every browser already ships, no login in front of it. A skeptic who never played the match can audit it anyway.

That is the model we built Orion on. Every server-backed match commits to its dice with SHA-256 before the first roll, derives each roll with HMAC-SHA256, and reveals the seed when the game ends. The verifier at orion.arthea.ai/verify is live today. It recomputes the entire chain in your browser with Web Crypto, and it asks for no account, because a proof that requires permission is not a proof. The server also validates every move in live play at source. Honest dice deserve an honest board.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. Behind it, every match from the first carries its own receipt.]]></content:encoded>
      <category>provably fair</category>
      <category>certified RNG</category>
      <category>fairness</category>
    </item>
    <item>
      <title>Do backgammon apps cheat? What players can actually check</title>
      <link>https://orion.arthea.ai/news/do-backgammon-apps-cheat/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/do-backgammon-apps-cheat/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Mostly no: doubles land one roll in six and memory keeps a biased score. The real question is whether an app can prove its dice. Here is how to check.</description>
      <content:encoded><![CDATA[Almost always, no. The rolls that feel rigged are usually honest randomness read through human memory: doubles arrive one roll in six, streaks are normal, and the losses stay with you while the wins fade. The harder truth sits underneath. With a closed app you cannot verify any of this, so the real question is not whether an app cheats. It is whether the app can prove that it does not.

Every backgammon app hears the same two complaints, and players have made them for as long as dice have rolled. The opponent gets too many doubles. The opponent always finds the number they need. The loser said it in the cafes of Istanbul when the dice were bone and the table was wood. It deserves a serious answer, not a shrug. Here is one.

Why does the app roll so many doubles?

Because doubles are common, and the ones that beat you are memorable. Two dice can land 36 ways. Six of those are doubles. One roll in six comes up paired, and that single fraction powers almost the entire complaint.

A typical game runs 50 to 60 rolls across both players. Eight to ten doubles at the table is not a hot streak. It is the average game. Back-to-back doubles land one pair in 36. Play a few games in an evening and you should see it more than once. Three doubles in a row is one chain in 216. Play regularly and you will meet it. Arithmetic, not a glitch. Any single number your opponent needs shows on at least one die 11 times in 36, nearly one chance in three.

The bar tells the same story. Close two points of your home board and your opponent still enters about eight times in nine. Leave a blot within direct range and the number that hits it lands about one time in three before combination rolls add more. When the numbers they need are that available, "they always roll what they need" is simply what fair dice look like from the losing side of the board.

Why do the dice only feel unfair when you lose?

Because memory keeps a crooked ledger. Psychologists have measured this for decades: a loss weighs roughly twice as much as an equal win. So the 6-6 that wiped out your prime stays with you for a week, while your own 6-6 felt like justice and was forgotten by dinner. Nobody remembers the night the dice were average, and nobody posts about it either. That is why every rigged-dice thread reads the same, and almost none of them come with a tally.

This is not a flaw in you. Attention stores the roll that changed the game and discards the fifty that did not. Players were accusing dice of treachery when dice were carved from bone. The app inherited the suspicion the day it drew its first checker.

Could a backgammon app actually cheat?

Yes, and honesty requires saying so plainly. The tumbling cubes on your screen are an animation. The numbers come from code on a server you cannot see, and nothing physical constrains them. Most studios have no motive, since a rigged game is a dead reputation waiting to surface, and when players run counting tests on big apps the totals generally come back ordinary. But "probably not" is trust, not proof. With a closed app there is no experiment you can run that settles the question, because the architecture keeps the dice out of reach. So the honest division is not cheating apps versus honest ones. It is dice you can check versus dice you must take on faith.

You cannot audit a feeling. You can audit a receipt.

How can you test the dice yourself?

Before anyone reaches for cryptography, run the paper test. It settles most suspicions in a single evening.

Set the expectation first: Write down what fair looks like before you play: one roll in six is a double, for both sides. Deciding what counts as suspicious after you have lost is how bias wins.

Tally instead of remembering: For one full session, count every roll and every double, yours and your opponent's. Paper beats memory, because memory is the instrument that was misreporting in the first place.

Read the result like a statistician: In 120 rolls, expect about 20 doubles. Anywhere from the low teens to the high twenties is ordinary swing. Only a large gap that survives many sessions means anything at all.

Two things happen when players actually run this test. First, the opponent's doubles land near the base rate almost every time, and it is their own count that surprises them. Second, the test reveals its own ceiling. Totals can be clean while timing is dirty: a determined cheat could serve a perfectly fair share of doubles and still deliver the right one at the worst moment for you. Counting cannot catch that. Receipts can.

What does Orion publish for every game?

A receipt, in three parts. Before the first roll, our server commits to its dice: it publishes a SHA-256 fingerprint of a secret seed while the game is still unplayed. During the game, every roll is derived from that seed with HMAC-SHA256, one roll at a time, in order. After the last move, the seed itself is revealed.

That structure makes the dice checkable. Hash the revealed seed and it must reproduce the fingerprint published before anything was rolled, which proves the dice were fixed before either player moved. Re-derive the rolls and they must match the game you actually played, which proves nothing was swapped along the way. Change a single die anywhere and the chain refuses to close. The verifier at orion.arthea.ai/verify recomputes all of it in your browser using Web Crypto, with no account and nothing to install. And because the server validates every live move at source, an opponent cannot slip an illegal play past you while you are busy watching the dice.

We built it this way because "trust us" is not an answer we would accept from anyone else. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. The verifier went live first, on purpose. The proof should be waiting before the players arrive.]]></content:encoded>
      <category>fair dice</category>
      <category>provably fair</category>
      <category>backgammon apps</category>
    </item>
    <item>
      <title>Why a fair game server trusts no one, including us</title>
      <link>https://orion.arthea.ai/news/server-authoritative-multiplayer/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/server-authoritative-multiplayer/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>What server authoritative means: the server holds the board, validates every move at source, and commit-reveal dice keep even the house honest.</description>
      <content:encoded><![CDATA[Server authoritative means the server keeps the only copy of the game that counts. Your device sends the move you want to play, the server checks it against its own dice and its own rules, and only then does the board change. A modified app can claim anything it likes: the truth lives somewhere it cannot reach.

Backgammon earns this discipline. It is a dice game with a ladder and a pot, which means there are two ways to cheat: fake your moves, or bend the rolls. A fair server has to shut both doors. The second door includes us, and we will get to that.

What does server authoritative actually mean?

Every live Orion match exists in exactly one place: on our server. The board on your phone is a window, not the game. When you drag a checker, the app does not move it. It sends a small message that says, here is what I want to play. The server replays that intent against its own board, with its own rules engine, using the dice it rolled for that turn. Legal moves advance the game, and both players receive the same new position. Illegal moves never happened.

This is one server for everyone. The iOS app is native SwiftUI, the Android app is native Jetpack Compose, and both are windows onto the same backend and the same player pool. Different glass, same truth.

Why do games that trust the client get cheated?

Because your opponent's device belongs to your opponent. Anything running on it can be changed. A patched client can send moves that were never legal. A memory editor can rewrite a number in place, and the screen will faithfully display the lie. None of this is exotic. The tools are free, documented, and older than smartphones.

A game that trusts the client is asking every player to promise not to look under the hood. Some will look. So we made the client's opinion irrelevant. In Orion the app has exactly three jobs:

Render: draw the board, the checkers, the cube, the timer. Request: ask the server to play a move, offer a double, concede. Receive: display whatever position the server says is real.

Rolling dice, judging legality, and moving a single coin are not on that list. Nothing that decides an outcome originates on a device.

What happens when I play a move?

You play.: Your app sends the intent: this checker, these points. Nothing on your device mutates the real board.

The server replays it.: Our rules engine runs the move against the server's board and the server's dice. If it is legal, the game advances. If it is not, it is refused at source, whatever app sent it.

Both sides get the same truth.: The new position goes out to both players. If your connection drops, the turn timer's reconnect grace holds your seat, and you come back to the board as it actually is, not as your phone last remembered it.

Stakes follow the same rule. When a match with a buy-in begins, the server escrows both antes. The pot lives in one ledger, on the server, and it settles there when the game ends. No client is ever asked to report what it thinks it won. Coins and gems are virtual, with no real-world value and no cash-out, and we still run that ledger like it holds something precious, because it does: your trust in the score.

What stops the server itself from cheating?

This is the right question, and most games hope you never ask it. A server that rolls dice in secret could roll them kindly for whoever it favors, and you would never see it. Every roll looks plausible. That is what dice do.

So we took that power away from ourselves. Before a game begins, the server generates a secret seed and publishes a SHA-256 commitment to it. Every roll is then derived from that seed with HMAC-SHA256, one roll at a time, in order. When the game ends, the server reveals the seed. Anyone can check that the seed matches the commitment made before the first roll, and that every roll in the game follows from it. In the cryptographic sense, your dice were fixed before either player touched a checker. We could not bend roll twelve without breaking the promise we published at roll zero.

You do not have to take our word for any of this. The verifier at orion.arthea.ai/verify recomputes the whole chain in your browser, using Web Crypto, on your machine, from the public record of the game. No account needed. No app needed.

How do the two halves lock together?

Server authority and commit-reveal dice are two halves of one design, and each covers the other's blind spot. A server that validates moves but rolls in secret asks you to trust the house. Committed dice on a trusting server ask you to trust every stranger's app. Together they close the loop. The client cannot fake a move, because the server replays every move at source. The server cannot rig a roll, because every roll is chained to a commitment published before the game. No one at the table, including the people who built it, holds unchecked power over the outcome.

The client cannot fake a move. The server cannot fake a roll. Everything else is decoration.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. The verifier came first, deliberately: the proof of fairness should exist before the first real pot is ever played.]]></content:encoded>
      <category>fair play</category>
      <category>engineering</category>
      <category>provably fair</category>
    </item>
    <item>
      <title>The doubling cube, explained properly</title>
      <link>https://orion.arthea.ai/news/backgammon-doubling-cube-guide/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-doubling-cube-guide/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Doubling cube rules in plain language: who may double and when, take or drop, redoubles, the 25 percent take point, and the mistakes beginners make.</description>
      <content:encoded><![CDATA[The doubling cube lets either player raise what a backgammon game is worth. At the start of your turn, before you roll, you may offer a double: your opponent either takes, and the game goes on for twice the stake, or drops, and concedes at the current value. Whoever takes the cube owns it, and only the owner may double again.

That one choice carries most of the game's depth. The checkers decide who is winning. The cube decides what winning is worth, and it asks you, several times a game, to put a number on your own confidence.

What is the doubling cube?

It is the big die that never gets rolled. Its faces read 2, 4, 8, 16, 32 and 64, and it starts each game in the middle of the board with 64 showing, which by convention means the game is worth one point and the cube belongs to nobody. It is a marker, not a die. It records two facts: the current stake, and who holds the right to raise it.

It is also young. Backgammon is thousands of years old. The cube appeared in New York in the 1920s and remade everything around one question: how sure are you?

Who may double, and when?

While the cube is in the middle, either player may make the first double. The moment is fixed: a double is offered at the start of your turn, before you roll, never after. The sequence never changes.

Offer: At the start of your turn, before rolling, turn the cube to the next value and place it toward your opponent.

Take or drop: Your opponent decides. Take, and the game continues at the doubled stake with the cube on their side of the board. Drop, and the game ends immediately. They concede the stake as it stood before your offer.

Ownership: After a take, the cube belongs to the taker. You may not double again. Only the owner may. If they later redouble and you take, the cube is yours once more.

Redoubles climb the same ladder, 2 to 4, 4 to 8, all the way to 64 in theory. In practice few games pass 4, because every level doubles the price of being wrong.

When is a take correct? The 25 percent idea

Set gammons aside and count. Drop, and you lose one point, every single time. Take, and the game is worth two: minus two when you lose, plus two when you win. Now play the same position four times with a one in four chance of winning. Dropping costs four points. Taking loses two three times and wins two once: the same four points. That is the break-even. Win more often than one game in four and taking is correct, even though you will usually lose that particular game.

This is the most freeing idea in backgammon. You are allowed to take doubles you expect to lose, and a correct take often feels a little wrong in the stomach. Gammons bend the number, since a doubled gammon costs four points, so be less eager when your opponent threatens one. But the anchor holds: you need a quarter of the game, not half of it.

You do not need to be the favorite to take. You need one game in four.

When should you double?

Doubling has a window. Double too early and you hand your opponent a comfortable take plus the cube, which they now own and will turn against you the moment the game swings. Double too late and they simply drop, and you collect a single point from a position that had grown worth more. The window sits between those errors: you are a clear favorite, the next exchange of rolls could move the game sharply, and your opponent's take is close to being wrong. Before choosing to wait, ask which rolls would make you regret it. If several sequences would carry you from strong to untouchable, double now, while they can still accept.

There is a ceiling too. If your position is so strong that a gammon is likely, doubling can be the second best play: your opponent drops, pays one point, and escapes a game that might have been worth two. This is called being too good to double. Play on, chase the gammon, and keep the option of offering the cube later if those chances fade.

What mistakes do beginners make with the cube?

Doubling from excitement. A thrilling roll is not a doubling window. Price the position, not the feeling the last roll gave you. Dropping too often. Every drop is a certain loss of a full point. Fold whenever you fall behind and good players will double you early and often, and the cube alone will beat you. Ignoring ownership. Holding the cube is quiet equity. Your opponent cannot raise the stakes, so you can play on through positions you might otherwise fear. A careless redouble hands that safety back. Never redoubling at all. When a game you took turns your way, the cube in your hand is not a trophy. There is a moment to use it.

How does the cube move the stake in Orion?

On Orion's live tables the cube works exactly as it does on wood, and the stake it moves is coins, which are virtual, carry no real-world value, and cannot be cashed out. When a double is taken, the raise is escrowed on both seats. The added stake is set aside from each player before play continues, and the escrow is re-verified at every redouble, so the number on the cube always matches what both seats have truly committed. The server validates every move of a live game at source, and the dice are committed before the game and revealed after it.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. The verifier at orion.arthea.ai/verify is live already, if you want to see how we treat a die before you ever have to trust one.]]></content:encoded>
      <category>doubling cube</category>
      <category>strategy</category>
      <category>rules</category>
    </item>
    <item>
      <title>Is backgammon luck or skill? Both, and that is the point</title>
      <link>https://orion.arthea.ai/news/is-backgammon-luck-or-skill/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/is-backgammon-luck-or-skill/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Both. One game is noisy, but across many games skill dominates: better players measurably win more. Why dice make backgammon better than pure-skill games.</description>
      <content:encoded><![CDATA[Both, and the blend is the design genius. One game of backgammon is genuinely noisy: a single roll can rescue a lost position or wreck a won one. Across many games the noise cancels and skill dominates, which is why the same names keep finishing at the top of every rated ladder in the world.

So the honest answer has two halves. Tonight, anything can happen. This year, the better player wins. Everything interesting about backgammon lives between those two sentences.

Why does one game feel so random?

Because the dice write the menu. Every turn you choose only among the moves your roll allows, and sometimes every option is bad. A blot survives or it does not. Double sixes arrive at the perfect moment or the worst one. In statistical terms, a single game is a high-variance event: the skill signal is real, but it is buried under honest noise. Even the strongest player alive drops single games to newcomers, and not rarely.

This is the part that fools people. They watch one game, see a lucky shot decide it, and file backgammon next to the games with no decisions at all. One game is simply too small a sample to show what is going on.

How do we know skill wins over many games?

Measurement. Modern analysis engines can grade every decision in a game against best play, and they find large, stable gaps between players: some give away a sliver of winning chances per game, others give away slices, and each player's error rate barely moves from month to month. Rating systems built on thousands of recorded games show the same picture from the other side. Strong players hold high ratings for years. If backgammon were a coin flip, every rating list would be soup. It is not.

The everyday proof is simpler. Sit a strong player against a casual one for a single game and the result is a real question. Make it a match to eleven points, then a season of matches, and the question quietly disappears. Skill in backgammon is a small current under choppy water: invisible in any single wave, undeniable in where the boat ends up.

Is backgammon more luck than chess?

In one game, yes, by design. Chess has no dice, so the stronger player wins almost every time, and a grandmaster against a club player is not a contest, it is a demonstration. That purity is chess's glory and its cruelty: if you are clearly weaker, you know exactly how the evening will go before you sit down.

Backgammon sits at a different point on the variance dial. The stronger player is favored in every game, often heavily, never certainly. Then the sample grows, and backgammon starts behaving like chess: over a long series, results converge to the true edge. Chess tells you the truth immediately. Backgammon tells you the same truth politely, over time.

Why do the dice make it better, not worse?

Here is the design genius. Because a beginner can win tonight, beginners keep playing. Nobody sits down at a backgammon board doomed. The dice hand new players real victories over stronger opponents, and those early wins are what keep someone at the table long enough to get good.

And because the edge only shows across many games, experts stay hungry. A pure-skill game goes quiet at the top: once the ranking is settled there is little left to prove. Backgammon never settles. The expert has to grind the edge out session after session, and every wild loss to a lucky shot is an invitation to prove it again. The luck is not a flaw in the design. It is the engine that keeps both ends of the table coming back.

The dice decide the game. The decisions decide the season.

Where does the skill actually live?

Everywhere the dice leave you a choice, which is nearly every turn. The decisions that separate players:

Opening play: the first rolls have known best moves, and small early mistakes compound. Checker play: building primes, timing hits, choosing when to run and when to anchor and wait. Risk shaping: which blots to leave, and which return shots you can afford to offer. The race: counting pips honestly, and playing a pure race differently from a holding game. The doubling cube: the deepest skill in the game, knowing when a lead is worth doubling and when a double is worth taking. The bear-off: bringing checkers home without leaving one last shot.

None of these choices guarantees a game. Each one moves your winning chances a little, and small edges compounded over hundreds of rolls are exactly what ratings detect.

How does rated play surface the skill?

This is why we built Orion around seasons rather than single games. Every player starts at zero trophies. Rated play carries you through ten leagues, from Driftwood at the bottom to Orion at the top, and the ladder runs in quarterly seasons. Season One is The First Crossing. A ladder like that is a long-run instrument: one lucky night barely moves it, while a real edge, applied across a season, carries you exactly as far as your play deserves.

The other half is trust, because a luck-and-skill game is only fair if the luck is honest. Orion's dice are provably fair: the server publishes a SHA-256 commitment to the dice sequence before the game begins, derives each roll with HMAC-SHA256, and reveals the seed when the game ends. Anyone can recompute every roll in a browser at orion.arthea.ai/verify, no account needed. So when a miracle roll beats you, and one will, you can confirm it was a genuine miracle.

Backgammon has kept its seat for five thousand years because the dice let anyone win a game and the skill decides who wins the year. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now, and behind it The First Crossing starts everyone at zero.]]></content:encoded>
      <category>luck vs skill</category>
      <category>strategy</category>
      <category>ratings</category>
    </item>
    <item>
      <title>Backgammon opening rolls: what to play and why</title>
      <link>https://orion.arthea.ai/news/backgammon-opening-rolls/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-opening-rolls/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>The standard play for all fifteen backgammon opening rolls: the point makers, the split or slot choices, and the runs, plus the principles behind them.</description>
      <content:encoded><![CDATA[Three opening rolls make a point on the spot, and you should always take it: 3-1 makes your five point, 4-2 makes your four point, 6-1 makes your bar point. With 6-5 a back checker runs straight to safety. The other eleven rolls are a choice between splitting your back checkers and bringing down builders, and the standard plays below cover all fifteen.

First, an anchor so the words mean something. We count points from your own side: your six point and eight point start with heavy stacks, your midpoint is the 13 point, and your two back checkers wait far away on the 24. The opening has two jobs, claim good points and put lazy checkers to work.

What are the best opening rolls in backgammon?

Four rolls make a new point immediately. Take them every time, no debate, no style. A made point is a wall your opponent must play around for the rest of the game.

3-1: make your five point, one checker from the eight point and one from the six. The best opening roll in the game. 4-2: make your four point, the same shape one step deeper, one from the eight and one from the six. 6-1: make your bar point, one from the midpoint and one from the eight. With your six and eight points already held, that is three walls in a row. 5-3: make your three point, one from the eight and one from the six. A little deep this early, but modern rollouts prefer the made point to shuffling builders around.

The five point is called the golden point for a reason. It cramps the two enemy checkers sitting in your home board, denies them the anchor they want most, and starts the prime that wins most games. When a roll offers it, take it.

Should you split or slot with 2-1, 4-1 and 5-1?

These are the awkward ones. The big number does honest work and brings a builder down from the midpoint. The one is the problem child, and there are two schools. Splitters move a back checker from the 24 point to the 23, giving the pair two ways forward and taking the sting out of the worst attacks. Slotters drop a checker from the six point onto the five, volunteering a blot on the golden point: survive one exchange and you cover it, and the best point on the board is yours by the second turn.

2-1: midpoint to the 11 point with the two, then split to the 23 or slot the five point with the one. 4-1: midpoint to the nine point with the four, then the same choice with the one. 5-1: midpoint to the eight point with the five, perfectly safe, then the same choice with the one.

Modern rollouts lean toward the split in a normal game and toward the slot when you need a gammon. Both are respectable. Choose one philosophy and learn its follow-ups rather than alternating at random.

What do you play with the middling rolls?

Four rolls neither make a point nor reach safety. Each one splits its work: one die reaches for ground on your opponent's side, the other brings a builder down. Playing both dice as builders is never far behind.

3-2: split to the 21 point with the three and bring a builder to the 11 with the two. Two builders down, to the 11 and the ten, is close behind. 4-3: almost a coin flip. Split to the 20 point with the four and bring a builder to the ten with the three, or play both down as builders. 5-2: midpoint to the eight with the five, then split to the 22 with the two. The quieter second builder to the 11 trails by a hair. 5-4: midpoint to the eight with the five and split to the 20 point with the four. The 20 is the anchor you want most, so start asking for it early.

When should you just run?

When you roll a six. A six is the only number that carries a back checker past your opponent's bar point in one hop, before any blockade exists. Only one of the four sixes lands safe; the others buy good ground with a blot.

6-5: the lover's leap. One back checker runs from the 24 point all the way to your midpoint. Completely safe, always correct. 6-3: run to the 18 point with the six and bring a builder to the ten with the three. 6-2: run to the 18 with the six and bring a builder to the 11 with the two. 6-4: a rare three-way tie. Run one checker to the 14 point, or run to the 18 with a builder to the nine, or make your two point. The first two keep everything working; the two point is real but deep.

Can you roll doubles on the opening roll?

No. A game opens with each player casting one die, and the higher number moves first using exactly the two dice showing. Equal numbers are cast again. So the first roll is always two different numbers, which is why there are fifteen opening rolls rather than twenty-one, and why doubles only enter the game from the second roll onward.

What principles cover every opening roll?

Strip the fifteen rolls away and three habits remain.

Make points. A made point never needs defending and never gets hit. Any roll that makes the five point, the four point or the bar point plays itself. Unstack. You start with towers of five on the six point and the midpoint. The third and fourth checkers in a stack are doing nothing; spread across the outfield they become threats. Do not bury checkers. Deep points like the two and the one are cheap now and useless later. Keep your checkers where the fight is going to be, from your four point out to your opponent's five.

The opening roll is not about this turn. It is about the shape your next six turns inherit.

However you open, the argument afterward should be about the play, never about the dice. In Orion every roll is committed to before the game begins and anyone can check the record at orion.arthea.ai/verify, which leaves the fifteen openings as the only thing left to argue about.]]></content:encoded>
      <category>Strategy</category>
      <category>Backgammon</category>
    </item>
    <item>
      <title>How to read a backgammon position like a player</title>
      <link>https://orion.arthea.ai/news/reading-the-board/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/reading-the-board/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Three questions strong players ask every turn: what kind of game is this, where are the blots, what does the pip count say. Learn to read structure.</description>
      <content:encoded><![CDATA[Reading a backgammon position comes down to three questions, asked in order before you touch a checker: what kind of game is this, where are the blots and what hits them, and what does the pip count say about the cube. Strong players run all three every turn, even the quiet ones. The dice tell you what you can do. The position tells you what you should want.

That is the real gap between beginners and strong players. A beginner sees 5-2 and hunts for somewhere safe to put it. A strong player already knows what the position wants, then checks whether this roll can deliver it. Rolls change every turn. Structure changes slowly, and structure is what decides games.

What kind of game is this?

Every position belongs to a family, and every family has its own plan. So name the game before you move. There are five shapes worth knowing.

A race: contact is broken or nearly broken, nobody can be hit, and the only plan is speed and a clean bear-in. A holding game: you keep an anchor, a made point in your opponent's territory, and wait for a shot as their checkers come home past it. A priming battle: both sides build walls of consecutive points and try to trap an enemy checker behind them. A blitz: one side attacks loose checkers in its own home board and tries to close the door before the defense finds an anchor. A back game: the last resort, two or more anchors deep in enemy territory, playing for a late shot after the race is long lost.

Then ask who prefers what, because the two of you almost never want the same game. The side ahead in the race wants contact gone: break away, run, turn everything into a sprint. The side behind wants contact kept: hold the anchor, build the home board, slow the game down and wait. And remember that games migrate. A blitz that fails becomes a priming battle. A holding game where the shot never comes fades into a losing race. Name the game you are in, and the game it is threatening to become.

Where are the blots, and how do you count shots?

A blot is a single checker on a point, and it is exactly as unsafe as the number of rolls that hit it. Not vaguely unsafe. Countably unsafe. There are thirty-six possible rolls. Any single direct number, six pips or fewer, appears on eleven of them, close to a third. A blot two pips away is hit by twelve rolls once double aces join in. At six pips away the combinations pile up and seventeen rolls hit.

Then comes the cliff. At seven pips away only six rolls hit. At eleven, two. So the working wisdom is short: a direct shot lands about a third of the time or more, an indirect shot a sixth of the time or less. Count shots for both sides every turn, the ones aimed at you and the ones you are aiming. Danger in backgammon is not a mood. It is a number out of thirty-six, sitting there waiting to be counted.

What does the pip count say about the cube?

The pip count is the total distance each side still has to travel, added up checker by checker. Count yours, count theirs, and you know who is winning the underlying race. That number is the skeleton of your plan. Behind in the count, you should love contact and avoid racing. Ahead, you should be breaking contact and simplifying toward the sprint you are favored to win.

The count also runs the cube. In a straight race, a lead of around ten percent of your own count puts you in doubling territory, and somewhere past twelve percent the trailing side should start thinking about dropping. With contact on the board the count is only part of the story, since shots and home boards can outweigh it, but the habit does not change: every turn, ask the cube question in both directions. Would I double here? Would I take? Players who ask it every turn are never surprised by a double. Players who skip it are surprised constantly.

How do the three questions work in a real turn?

Picture a mid-game moment, no diagram needed. You are twelve pips behind. Your two back checkers sit anchored on your opponent's four point, your home board has three made points, and your opponent has escaped everything. Last turn their luck thinned, and they had to leave a lone builder five pips in front of your anchor.

Run the questions. One: this is a holding game and you are the holder. Twelve behind, you want exactly this, contact and patience, while they want to slip that builder home and turn everything into a race. Two: the blot sits five away, a direct shot. Eleven rolls carry a five, and with the middle points open, 4-1 and 3-2 stretch it to fifteen of thirty-six. Three: the count says you have no double yet, but it also says a hit rewrites everything. That checker would re-enter against your three-point board and travel the whole table again, squeezing past your anchor near the end. The race flips, and the cube starts leaning your way.

Now the dice, and here is the point of the whole exercise. Roll 5-2 and the read has already made the play for you: hit, tidy what you can behind it, and accept the return shots you have already counted. Roll 4-1 instead and you cannot touch the blot, and the read still makes the play: keep the anchor whole, work on the fourth home point, and wait. Two different rolls, one unchanged plan. The position never confused you, because you never asked the dice for a plan.

You do not play the roll. You play the position the roll landed in.

Ask the three questions for a week and boards start to organize themselves: game, shots, count, then the move almost picks itself. That habit is what we built Orion around: a voyage of twelve destinations where structure is everything, and dice you never take on faith, because every live roll carries a cryptographic proof and the public verifier at orion.arthea.ai/verify rechecks it in your browser, no account needed.]]></content:encoded>
      <category>strategy</category>
      <category>basics</category>
    </item>
    <item>
      <title>Pip counting: the skill that changes everything</title>
      <link>https://orion.arthea.ai/news/pip-counting-basics/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/pip-counting-basics/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>A pip count is the total pips needed to bear off every checker. Both sides start at 167. Learn the direct method, the shortcuts, and when to race or hold.</description>
      <content:encoded><![CDATA[A pip count is the total number of dice pips a player still needs to bring every checker home and bear it off. You get it by multiplying each point's number by the checkers on that point and adding the results. Both sides start the game at 167, and the lower count is winning the race.

Most players never learn to count. They feel the race instead of measuring it, and their biggest decisions float on mood. The pip count replaces that fog with one number. Once you have it, the position tells you what it is: a race to push, or a fight to hold. That is why we call it the skill that changes everything.

What is the pip count, exactly?

Every point has a number, counted from your own bear-off corner. A checker on your 6 point carries six pips. A runner back on the 24 point carries twenty-four. Your pip count is the sum over all fifteen checkers, the total distance your army still has to travel. The starting position is worth memorizing, because every early count is just an adjustment of it:

Two runners on the 24 point: 48 pips Five checkers on the 13 point, the midpoint: 65 Three on the 8 point: 24 Five on the 6 point: 30 Total: 167 for each side

How do you actually count, step by step?

The direct method is slow and honest. Do it this way first, before any tricks.

Number the points: Your home board is 1 to 6, your outer board 7 to 12, your opponent's side 13 to 24. A checker's point number is exactly the pips it still needs.

Multiply each stack: For every occupied point, multiply the point number by the checkers on it. Five checkers on your 6 point are 30 pips. Two on the 20 point are 40.

Keep a running total: Add as you go instead of memorizing a list. Sweep the board in one direction so you never wonder what you have already counted.

Count your opponent the same way: Their numbers run from their own corner, so their 6 point is six pips for them. Two honest totals, one comparison, and you know who is ahead and by how much.

What shortcuts do real players use?

Nobody multiplies twenty separate stacks under a turn timer. Counters lean on structure instead.

Clusters. Read a stack as one block: five checkers on the 6 point is one multiplication, 30, not five additions. Symmetry. Checkers spread evenly around a middle point count as if they all stood on it. One checker on your 5 point and one on your 7 count as two on the 6: twelve pips. Mental shifting. Sliding one checker a pip closer and another a pip farther leaves the total unchanged. Shift stragglers onto round numbers before you multiply. Count the difference. Pair your checkers with mirrored opposing ones and tally only the gaps. Late in the game you rarely need 132 against 141. You need: I trail by nine. Anchor on 167. In the opening, adjust from the start instead of recounting: subtract what each of your rolls moved, and remember that a hit sends a checker all the way back.

Why does the pip count decide strategy?

Because backgammon is two games wearing one board. Once all contact is broken it is a pure race, and the count is the whole story. While contact remains, the count tells you which game you are in. Ahead in the race, you want quiet: break contact, avoid loose blots, bring the army home. Behind in the race, you want noise: keep an anchor in their board, hold your ground, and play for the hit, because one hit swings the count more than any roll ever will.

Ahead in the race, race. Behind in the race, hold and hit.

The count also carries the doubling cube. In a pure race near 100 pips, an old guide says the leader can offer at about an 8 percent lead, redouble at 9, and the trailer should pass beyond about 12. The exact thresholds move with the length of the race; the habit does not. No count, no cube decision. Doubling on feel is how games quietly slip away.

How do you practice pip counting?

Like scales. Ten focused minutes beat an hour of vague intention.

Week one: own the start: Rebuild 167 from its parts, 48 plus 65 plus 24 plus 30, until the arithmetic is boring. Then play normally, and after your first two or three rolls adjust the count in your head.

Week two: one full count per game: Pick a single moment, the first cube thought is the right one, and produce both totals before you act. One honest count per game builds more than ten sloppy ones.

Week three: count endings: Late races are short counts with instant feedback. Count both sides, predict the winner, then watch whether the race ends the way the number said.

Week four: switch to differences: Count the gap directly with pairing and shifting, then verify with one full count per session. When your shortcut and your slow count agree, the skill is yours.

One more reason to trust the number. In Orion, the dice you count are provably fair: the server locks its rolls with a SHA-256 commitment before the game, reveals the seed after, and the verifier at orion.arthea.ai/verify recomputes every roll in your browser, no account needed. The count is math you do; the dice are math you can check.]]></content:encoded>
      <category>strategy</category>
      <category>pip count</category>
      <category>basics</category>
    </item>
    <item>
      <title>Gammons and backgammons: how scoring really works</title>
      <link>https://orion.arthea.ai/news/gammon-backgammon-scoring/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/gammon-backgammon-scoring/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Gammon doubles the game, backgammon triples it, the cube multiplies on top. Plain rules, worked examples, and why gammons move the cube.</description>
      <content:encoded><![CDATA[A gammon is a win where your opponent has borne off no checkers, and it scores double. A backgammon means they have borne off nothing and still have a checker on the bar or inside your home board, and it scores triple. The doubling cube then multiplies whichever result you land: win quality times cube, every time.

Most players learn the racing game and stop there. Then one night someone loses a single game by four points and asks what just happened. This is the part of backgammon scoring that turns a friendly race into a game of nerve, and it is worth learning properly, with numbers.

What is a single win worth?

One point, multiplied by the cube. You bear off all fifteen of your checkers first, and your opponent has borne off at least one of theirs. If nobody doubled, the cube sits in the middle showing 64, which by convention means one, and the game is worth exactly one point. Most games between careful players end this way: the losing side sees the race slipping and spends its last rolls saving the gammon, hurrying a single checker off the board so the game closes at its base price.

What is a gammon in backgammon?

A gammon is a win where your opponent has not borne off a single checker. It doubles the value of the game. Not the cube: the result. Win a gammon with the cube untouched and you score two points instead of one. Gammons are rarely accidents. They come from a blitz that closed your home board while a checker waited on the bar, or from a prime that held a runner until the race was gone. The rule exists to pay you for the size of the win, not just the fact of it.

What is a backgammon?

A backgammon is a gammon with one extra indignity. At the moment your last checker leaves the board, your opponent has borne off nothing and still has a checker on the bar or inside your home board. That triples the game. Backgammons are rare between experienced players, because the trailing side almost always has time to run its rearmost checkers home while you bear off. When one does land, it is usually the end of a crushed blitz: somebody dancing on the bar against a closed board while fifteen checkers file off on the other side.

How does the doubling cube multiply the score?

The win type and the cube are separate multipliers, and they stack. The whole system is one line of arithmetic: points equal win type, 1 for a single, 2 for a gammon, 3 for a backgammon, times the cube value. Worked through:

Cube in the middle, single win: 1 x 1 = 1 point. You double, they take, you win a single: 1 x 2 = 2 points. You double, they take, you win a gammon: 2 x 2 = 4 points. They redouble to 4, you take, you win a gammon: 2 x 4 = 8 points. Cube on 4, backgammon: 3 x 4 = 12 points, in a game that opened worth 1.

Notice what the stack does. A doubled gammon does not add a bonus, it multiplies one, so scores climb in jumps: 1, 2, 4, 8, 12. That is why experienced players fear the combination more than either part alone.

One footnote from money play: many tables use the Jacoby rule, where gammons and backgammons count only if the cube has been turned. It discourages grinding out a dead game for an undoubled gammon. In match play the rule is off, and every gammon counts in full.

Why do gammon threats change cube decisions?

Because the take point moves. The textbook number says you can accept a double with about 25 percent winning chances: win 2 points a quarter of the time and you roughly balance the 1 point you would concede by dropping. That arithmetic quietly assumes every loss costs the same. Gammons break the assumption. If a real share of your losses will cost 4 points instead of 2, a 25 percent game is no longer a take, and a position that looks playable becomes a clear pass. Strong players are not folding the checkers. They are folding the price.

The threat bends the other way too. When your game is so strong that gammons are genuinely coming, turning the cube can be the mistake. Double, and a sensible opponent drops at once and hands you a single point. Hold the cube, win the gammon, and the same game pays two. This is the zone players call too good to double: the win has grown so large that offering the cube would sell it at a discount. Seeing that line, and knowing when to cash the sure point instead, is most of the distance between a checker player and a cube player.

A double is a price, and gammon threats are what move the price.

How does Orion score gammons and backgammons?

We score it exactly by the book: quality times cube. A single pays the base value, a gammon pays double, a backgammon pays triple, and the doubling cube multiplies on top, the same arithmetic as every example above. The server validates each live move at the source and settles the result to your wallet exactly, so the number on the victory screen and the number in your balance never disagree. Coins in Orion are virtual, with no real-world value and no cash-out. What the scoring really raises is the temperature of the climb, ten leagues from Driftwood to Orion.

The dice underneath it all are provably fair: a SHA-256 commitment published before the game, the seed revealed after it, and a verifier at orion.arthea.ai/verify that recomputes every roll in your browser, no account needed. So when a backgammon finally lands on you, and one day it will, you will at least know it was honest.]]></content:encoded>
      <category>rules</category>
      <category>scoring</category>
      <category>cube</category>
    </item>
    <item>
      <title>Seven backgammon mistakes that cost beginners games</title>
      <link>https://orion.arthea.ai/news/common-backgammon-mistakes/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/common-backgammon-mistakes/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>The seven habits that lose beginner backgammon games, from buried checkers to a silent doubling cube, and the fix for each one.</description>
      <content:encoded><![CDATA[Beginners lose backgammon games in the same seven ways: stray blots left deep in the opponent's home board, checkers buried on the 1 and 2 points, a doubling cube that never moves, reflex hits, races run from behind, anchors abandoned too soon, and rolls forced to fit a plan the dice never agreed to. All seven are fixable tonight. None of them needs theory, just a bad habit swapped for a better one.

We watch a lot of games. Beginners rarely lose to one catastrophic move. They lose to the same small leaks, turn after turn, until the position quietly belongs to the other player. Here are the seven, why each one loses, and what to do instead.

1. Leaving loose blots in your opponent's home board

Your two back checkers start on the 24 point, deep inside your opponent's home board. Beginners nudge one of them a pip forward, or strand a third checker back there, and leave blots sitting under the opponent's builders. The pips you lose when one is hit are trivial. The structure you lose is not: your opponent makes a home board point on your head, you sit on the bar while their board grows, and every later hit becomes more expensive than the one before. Keep the back pair together. Move them as a pair toward an advanced anchor, and split them only when you can say the reason out loud.

2. Burying checkers on the 1 and 2 points

Stacking checkers on your 1 and 2 points feels safe because nothing can hit them there. Nothing can use them either. A checker on the ace point has finished its race before the game is decided: it will never make your 5 point, never join a prime, never cover a blot. Bury four checkers like this and you are playing eleven against fifteen. Keep your spares up in the outfield instead, where they can still become the 5 point, the 4 point, the bar point. The high points win games. The low ones mostly store the evidence that you were nervous.

3. Never touching the cube, or doubling on a rush

The cube has two beginner settings, untouched and reckless, and both cost points. If you never double, your best games count exactly the same as your narrowest escapes: one point each. If you double on the surge of a single lucky hit, a calm opponent takes gladly, because one hit is rarely a winning position, and now the cube sits on their side of the board waiting for the game to turn. The habit to build is quieter: double when you are clearly ahead and content with either answer, collecting the point if they pass, liking your game if they take. If the urge arrives with your pulse, wait one roll.

4. Hitting when a point was better

Hitting feels like winning, so beginners hit everything. But send a checker to the bar while your own board has only two points made, and it strolls back in next roll, no worse off, while the blots you stretched to make the hit are now targets. The quiet alternative, making your 5 point, would have improved your game for good. Before you hit, ask what the hit buys: real ground in the race, real tempo, or a real chance the checker stays out for a while. If the answer is none of those and a strong point is available, make the point. Structure outlasts drama.

5. Racing when you are behind in the race

When you are behind in the pip count, contact is your only way back. Beginners break away anyway, because running feels like progress, and turn a live game into a pure race the count has already decided. Do the opposite. Hold your anchor, leave checkers where they still threaten something, and spend the waiting rolls building your home board. You are not stalling, you are aiming. In a holding game the shot usually comes, and if your board is ready when it does, one hit is the whole game.

Ahead in the race, race. Behind in the race, hold.

6. Breaking your anchor too early

An anchor, a made point in your opponent's home board, is what lets you play the rest of the game bravely. While it stands you always have a safe landing, you cannot be closed out, and their checkers must come home past your guns. Beginners break it the moment it stops feeling useful, which is usually a few rolls before it stops being useful, and the breaking play itself often leaves a blot on the most dangerous point of the board. Leave your anchor when the race says leave: you are ahead or level, and the way out is clean. Restlessness is not a pip count.

7. Playing the roll you wanted instead of the roll you got

You planned to make your bar point. The dice gave you a 6 and a 2. So you play half of the dream, leave a checker hanging where a builder should be, and call it progress. This is the deepest habit on the list: forcing every roll toward a picture drawn for dice you did not get. Strong players start over every turn. Take the number you actually rolled, list the two or three things it does well, a hit, a point, a safe builder, an escape, and choose among those. Let the dice write the plan. Your job is to read them honestly.

Fix even three of these and you will feel it within a week: fewer games that drift away for no reason you can name. The dice owe you nothing in this game, and never have. In Orion we make them prove it anyway, every roll committed before the first move and checkable after the last, which leaves the seven habits above as the only thing left to blame.]]></content:encoded>
      <category>strategy</category>
      <category>beginners</category>
    </item>
    <item>
      <title>Backgammon terms, from anchor to the bar</title>
      <link>https://orion.arthea.ai/news/backgammon-glossary/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-glossary/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>A working backgammon glossary: point, pip, blot, anchor, prime, the doubling cube and more, each defined in one or two plain sentences.</description>
      <content:encoded><![CDATA[Backgammon's vocabulary is small, old, and precise. This glossary defines the terms you will actually hear at a board, grouped the way the game is played: the board itself, the race, contact, the doubling cube, and match play. We kept each entry to one or two sentences, plain enough to use tonight.

A note before you start. These words are not jargon, they are compact stories, most of them centuries old. The game will teach them to you faster than any page can, so use this as a reference for the first few evenings, then forget it.

What are the parts of a backgammon board called?

Start with the geography. Twenty four narrow triangles, a raised spine down the middle, six points of home on each side. Everything else in the game is described relative to these.

Checker: one of the fifteen pieces each player moves around the board. Players also say stone, man, or piece; they all mean the same thing. Point: one of the twenty four triangles. Two or more of your checkers on a point make it yours, and your opponent cannot land there. Bar: the raised strip dividing the board in half. A checker that gets hit waits on the bar, and its owner can move nothing else until it re-enters in the opponent's home board. Home board: the six points where your checkers complete their journey. All fifteen must arrive before the final phase can begin. Bear off: to remove a checker from the board once your whole army is home. The first player to bear off all fifteen wins the game.

What is a pip in backgammon, and what counts as a race?

Underneath the checkers, backgammon is a footrace measured in numbers. These are the words for the measuring.

Pip: one unit of distance, a single step from one point to the next. Your pip count is the total distance all fifteen checkers still have to travel; the lower count is winning the race. Race: a position with no contact left, where neither player can be hit. In a pure race only the pip count and the dice matter, which is why serious players count. Crossover: moving a checker out of one quadrant of the board and into the next. Counting crossovers is the fast way to see how many rolls of work remain before you can bear off.

What do contact terms like blot, anchor and prime mean?

Contact is where backgammon stops being a race and becomes a fight. Most of the game's famous vocabulary lives here.

Blot: a single checker alone on a point. Any opposing checker that lands on it sends it to the bar, and its whole journey starts over. Hit: to land on an opponent's blot. One hit can swing a game further than any other single event on the board. Anchor: a point you own inside your opponent's home board. It gives your rear checkers a safe base, and it keeps your chances alive when the rest of the game goes wrong. Prime: several consecutive points all owned by one player, a wall. A full six point prime cannot be passed by any checker trapped behind it. Blitz: an all-out attack on the opponent's home board, hitting blots and making points while their checkers sit on the bar. A blitz ends quickly, one way or the other. Back game: a defense built on two or more anchors deep in the opponent's home board, waiting for one late shot. Famous, dramatic, and harder to win than it looks. Holding game: keeping one strong anchor on a key point and waiting for the opponent to run past it. You hold the door; they have to come through.

A blot is not a word you learn. It is a word that happens to you.

How does the doubling cube work, and what is a take point?

The cube arrived in the 1920s and remade a game already thousands of years old. It turns every position into two questions: who is winning, and by how much.

Cube: the doubling cube, marked 2, 4, 8, 16, 32 and 64. It is never rolled; offering it asks your opponent to play on for double the value or concede the game as it stands. Take and drop: the two legal answers to a double. Take means play on at the new value; drop means concede the game at the old one. Take point: the minimum winning chance you need to correctly accept a double. The textbook starting figure is 25 percent, before gammons and cube ownership bend it. Beaver: a house rule, not a tournament one. A player who is doubled may instantly redouble while keeping the cube, turning a 2 into a 4 on the spot.

What do gammon, backgammon and the Crawford rule mean?

Some words only earn their keep when you play for points. These decide how much a win is worth, and when the cube sleeps.

Gammon: a win completed before your opponent has borne off a single checker. It counts double. Backgammon: a win while your opponent still has a checker on the bar or inside your home board, with none borne off. It is the game's own name for its harshest result, and it counts triple. Match play: playing to a fixed number of points rather than game by game. The score changes what the cube and the gammon are worth, which changes correct play. Crawford rule: in match play, the first game after a player reaches match point is played without the doubling cube. One game of shelter, then the cube returns.

That is the working set. There are deeper drawers, duplication, diversification, the Jacoby rule, but these entries carry nearly all of the talk at a real table, and each one will make sense on the day a position demands it.

One last term, ours: provably fair. In Orion the dice are committed before a game and revealed after it, and anyone can recheck a finished game in the browser at orion.arthea.ai/verify, no account needed.]]></content:encoded>
      <category>glossary</category>
      <category>rules</category>
      <category>learn</category>
    </item>
    <item>
      <title>Backgammon vs chess: which should you learn?</title>
      <link>https://orion.arthea.ai/news/backgammon-vs-chess/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-vs-chess/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Chess is harder to study, backgammon is harder to judge. We compare difficulty, learning curves, upsets, the cube, and which game fits your life.</description>
      <content:encoded><![CDATA[It depends on how you like to think. Chess is pure calculation: perfect information, nothing hidden, every loss your own. Backgammon is risk management: the dice add variance, upsets happen, and the skill is making better decisions than your opponent under uncertainty, roll after roll.

The two games have shared shelves and tables for centuries, and people compare them because they sit at opposite poles of strategy. One removes chance entirely. The other builds the whole game around it. That single difference decides almost everything else: how long a game lasts, how a loss feels, who stays at the table, and where each game fits in a life.

Is chess harder than backgammon?

Harder to study, yes. Chess is a perfect information game. Both players see everything, so the truth of every position is on the board, waiting for whoever calculates deepest. Centuries of players have mapped that truth: opening theory runs past move twenty in some lines, and serious improvement means learning a share of it. The rating ladder is unforgiving too. Give one player a modest rating advantage and the weaker side stops winning almost entirely. There is no alibi in chess. When you lose, you were outplayed, and the game can show you exactly where.

Backgammon is harder to judge. The rules take fifteen minutes, and your first real game can happen the same evening. But every decision after that is probabilistic. Two plays can look nearly identical and sit a full blunder apart. Do you leave a shot now to win the race later? Do you hit and expose a checker, or play safe and drift behind? Strong players think in averages: not what will happen, but what happens across all thirty-six rolls that could come next. That kind of thinking does not come naturally to anyone. Learning it is the game.

Can a beginner really beat an expert?

In backgammon, yes, tonight. The dice guarantee it. A well-timed double six beats a world champion, and nothing the champion does can prevent it. In chess that almost never happens: the gap shows in every single game, and a strong club player will beat a novice a hundred times in a row. Backgammon's luck is loud in one game and quiet over many. Across a long match the stronger player pulls ahead. Across a season there is nowhere to hide. Variance decides evenings. Skill decides careers.

Chess rewards whoever sees furthest. Backgammon rewards whoever prices risk best.

What does the doubling cube change?

It turns the game into a negotiation. Either player can offer the cube, doubling what the game is worth, and the other side must accept the new terms or concede on the spot. It is an offer pushed across the table, and reading the person matters almost as much as reading the position. Most of backgammon's depth hides here: knowing when your chances justify the offer, and knowing when pride is telling you to take a double you should drop. The cube is also why backgammon stays social. Chess has no equivalent. You cannot raise.

Which is faster to learn?

Both games have a short doorway and a long corridor behind it. The honest comparison looks like this.

Rules: fifteen minutes for backgammon, about an hour for chess. First competent games: days for backgammon, weeks of tactics practice for chess. Club strength: a season of checker play and cube basics for backgammon; years of openings, endgames, and pattern study for chess. Mastery: decades for both, and nobody arrives.

Why does backgammon suit phones and cafe tables?

Because the games are short and the silence is optional. A serious chess game asks for an unbroken hour and full concentration; conversation is an intrusion. A backgammon game lasts five to ten minutes, the thinking comes in bursts between rolls, and you can talk the whole way through. That is why the board has lived in cafes from Tangier to Istanbul for centuries, and why it moved onto phones so naturally. A game fits inside a coffee. A match fits inside an evening. Chess earns its cathedral quiet. Backgammon was built for rooms with noise in them.

So which should you learn?

We are not going to declare a winner, because there is not one. Choose by temperament. If you want a game where nothing is hidden and every result is fully earned, learn chess: it is the cleanest test of calculation ever put on a table. If you want to get better at deciding without certainty, learn backgammon: it teaches you to price a risk, hold a plan loosely, and lose a fair game without losing your judgment. Plenty of players keep both, the way a runner keeps two distances. They train different muscles, and both will still be teaching you in thirty years.

We build Orion on the backgammon side of that choice, and we start with fairness: every roll is committed before the game begins and can be checked by anyone afterward at orion.arthea.ai/verify, no account needed. The rest is the voyage, twelve destinations of it, behind the App Store and Google Play links at orion.arthea.ai/go.]]></content:encoded>
      <category>backgammon</category>
      <category>chess</category>
      <category>strategy</category>
    </item>
    <item>
      <title>How long does a backgammon game take?</title>
      <link>https://orion.arthea.ai/news/how-long-does-backgammon-take/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/how-long-does-backgammon-take/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>A casual game takes 5 to 10 minutes, a match to 5 or 7 points 20 to 45. What stretches games, what shortens them, and why the loop suits a phone.</description>
      <content:encoded><![CDATA[A single casual game of backgammon takes 5 to 10 minutes. Experienced players finish faster, often in under five. Match play to 5 or 7 points usually runs 20 to 45 minutes, and the doubling cube decides which end of that range you get.

Those numbers are the game's quiet advantage. Chess wants an hour of silence. A hand of bridge wants a table of four. Backgammon wants ten minutes and one opponent, and it hands you a complete story in return: an opening, a crisis, a race, a result.

How long is one game of backgammon?

A typical game lasts twenty to thirty moves a side, and each move is a few seconds of thought. Over a real board you add the rituals: setting the checkers, shaking the cup, the pause while someone recounts a pip. Online the board does all of that for you, which is why digital games sit at the fast end of every estimate.

Two experienced players over the board: 3 to 6 minutes A relaxed casual game: 5 to 10 minutes A first game with someone learning: 15 to 20 minutes, and worth every one of them

The dice set the floor. Even a perfect game cannot end before the checkers have travelled the board and come home, so almost nothing finishes in under two minutes. The ceiling is another matter. Strategy sets that.

How long does a match to 5 or 7 points take?

Plan on 20 to 30 minutes for a match to 5 points, and 30 to 45 for a match to 7. A match is first to the target, and a game can be worth one point, two for a gammon, more once the cube has been turned, so the number of games varies. Five points can arrive in two games or in seven.

The cube is the great compressor. A well-timed double often ends a game on the spot: one player offers, the other judges the position too far gone and passes, a single point changes hands, and the next game begins inside thirty seconds. Club matches to 11 points can fill an evening, and championship finals run whole afternoons, but those are the marathon end of a sport that mostly lives in sprints.

What makes a backgammon game run long?

Contact. As long as the two armies can still hit each other, the game keeps its tension and its length. Two classic structures are famous for it.

The back game: two anchors deep in the opponent's home board, playing for a late shot. Timing is everything, so both players slow down and count. The holding game: one anchor on a key point, waiting while the opponent tries to bring checkers home past it. Every roll near the anchor is a small crisis. The mutual hold: both sides sit on strong points and neither can safely break. These stretch the longest of all.

Add a checker dancing on the bar against a strong home board and a ten-minute game becomes twenty. None of it is wasted time. The long games are usually the ones you remember.

What makes a game finish fast?

Blitzes and early cube decisions. A blitz is an all-out attack in the home board: points slam shut, the defender dances on the bar, and the game can be over in a dozen rolls. The cube ends games even faster. If one side takes a clear lead by the fifth roll, a confident double and a correct pass close the game in three minutes, and nobody has played badly. Pure races, where contact breaks early, also settle quickly, because both sides simply run for home.

A whole game, with its opening, its crisis and its race, fits inside a coffee.

Does online backgammon play faster?

Yes, and not because anyone hurries you. The board sets itself up, the dice land the instant you roll, and nobody has to walk the checkers back to their points between games. What online play really adds is pacing you can trust. In Orion every live turn runs on a timer, so the game keeps moving and nobody can sit on a roll until you give up and leave. There is reconnect grace too: a Wi-Fi blip in a lift or a tunnel gives you time to come back instead of costing you the game. And the server checks every live move at source, so there is nothing to dispute and nothing to replay.

Why does backgammon suit a phone session?

Because the shape of a game matches the shape of a break. Most phone sessions are a few minutes long: a queue, a kettle, a train between two stations. Backgammon is one of the few games that offers a real contest with real decisions and still finishes inside that window. You are never abandoning something half-built when life resumes. And when a game ends with four minutes to spare, one more is an honest promise, not a trap.

That is the loop we built Orion around. A game is minutes long, but the frame around it is a voyage: twelve destinations, from a Tangier cafe to Orion, the celestial table, with ten leagues to climb and a new season every quarter. Short games. Long journey.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. Behind it, Season One, The First Crossing, sets off.]]></content:encoded>
      <category>basics</category>
      <category>game length</category>
      <category>match play</category>
    </item>
    <item>
      <title>Backgammon etiquette, at the table and online</title>
      <link>https://orion.arthea.ai/news/backgammon-etiquette-online/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-etiquette-online/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>The unwritten rules of backgammon: which side to roll on, how to offer the cube, how to concede, and how the same manners translate online.</description>
      <content:encoded><![CDATA[Backgammon etiquette comes down to a handful of habits. Roll the dice on your right side of the board, let both dice land flat or reroll them, keep your hands off the checkers while you think, offer the doubling cube cleanly, and concede a lost game instead of making your opponent grind it out. Online, the same spirit holds: play at a fair pace, never stall the clock, and use emotes for warmth rather than needling.

None of this is stuffiness. Backgammon has been played for something like five thousand years, most of them face to face, close enough to touch the same board. The manners exist to keep a fast, sharp, luck-streaked game friendly. They survived because they work.

Which side of the board do you roll on?

Your right. Both players roll into the right half of the board from their own seat, so the dice have a clear lane and the checkers in play stay undisturbed. You also wait your turn. Your opponent ends their move by picking up their dice, and rolling before that is the table equivalent of interrupting. In most rooms an early roll does not count. You take it back and wait.

The dice themselves have one rule: flat or nothing. If a die lands tilted against a checker, leans on the frame, or skips off the board entirely, that is a cocked die, and you reroll both. Nobody negotiates over whether a leaning die is mostly a five. The rule is strict precisely so no one ever has to argue.

Can you touch the checkers while you think?

In person, no. You think with your eyes and move with your hand, once. Hovering a checker over three points while you calculate is noise for the player across from you. The rules will not stop you: unlike chess, your move is not final until you lift your dice. That is exactly why the courtesy exists. Make the whole move, then lift the dice. Lifting them is the signature that ends the turn.

Online, this rule softens for a good reason. The interface is your hand, and dragging a checker to try a line, then sliding it back before you confirm, is how the software is meant to be used. The spirit survives the move: decide, commit, keep the game breathing.

How do you offer the doubling cube politely?

Cleanly, at the start of your turn, before you roll. Turn the cube to the next number, set it toward your opponent, and say one word: double. The cube is already the whole argument. It needs no speech, no raised eyebrow, no commentary on their position. If they take, play on as if nothing happened. If they drop, gather the checkers without a victory lap. Doubling is the most emotional moment in backgammon, which is exactly why the custom keeps it quiet.

How do you concede gracefully, and what is slow-rolling?

Concede when the game is truly gone. In cube play that usually means passing a double you cannot take. Either way, a gracious player does not force a decided game through every last roll. Slow-rolling is the winner's version of the same sin: stretching the ending, hovering over the obvious final play, making your opponent sit through your enjoyment. Win quickly, lose honestly, and say good game in both directions. Five seconds of grace at the end of a match is what makes the next one possible.

What are the etiquette rules for online backgammon?

The customs do not vanish when the board becomes a screen. They translate, one for one.

Play at a fair pace. What courtesy used to enforce, the clock now does: in Orion every turn runs on the same timer, with grace to reconnect when a tunnel eats your signal. The polite pace and the legal pace are finally the same thing. Never stall to force a timeout. Sitting on a lost position and letting the clock bleed is the online version of knocking the board over. Every community treats it as a real offense, because it steals time instead of winning points. Emotes are for warmth, not taunting. One emote after a ridiculous roll is the nod across the table. The same emote after every point you win is needling, and everyone can tell the difference. Report abuse instead of matching it. At a cafe table, a rude player answers to the whole room. Online, the report button is the room. Use it once and move on. Answering abuse with abuse seats two rude players at one board.

Why has backgammon etiquette survived five thousand years?

Because backgammon is a conversation. Two players trade short sentences, each reply shaped by the last, while the dice keep interrupting with news neither of you chose. A game like that runs on feeling, and etiquette is the grammar that lets the feeling move without breaking anything. You can teach a stranger the rules in ten minutes. The manners are what make the ten minutes worth having.

Etiquette is not decoration on the game. It is how two strangers agree to enjoy the same ten minutes.

That is the table we are building. In Orion the server checks every live move at the source, and any roll can be verified in the browser at orion.arthea.ai/verify, so the oldest table suspicion, blaming the dice, can finally retire and leave room for the conversation.]]></content:encoded>
      <category>etiquette</category>
      <category>tradition</category>
      <category>online play</category>
    </item>
    <item>
      <title>Five thousand years of backgammon, briefly</title>
      <link>https://orion.arthea.ai/news/history-of-backgammon-5000-years/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/history-of-backgammon-5000-years/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>From the Royal Game of Ur around 2600 BCE to the 1920s doubling cube: backgammon&apos;s real history, honestly dated, and what never changed.</description>
      <content:encoded><![CDATA[Backgammon is the living end of a family of race games about four and a half thousand years old, and its earliest known ancestor is the Royal Game of Ur, played in Mesopotamia around 2600 BCE. The game we would recognize took shape in stages: Roman tabula, Persian nard, the tables games of medieval Europe, an English name in the 1600s, a doubling cube in 1920s New York, and computer analysis since the 1990s. The costumes changed constantly; the idea, two players racing on dice in front of an audience, never did.

How old is backgammon, really?

Younger than the slogan, older than almost everything else people still do for fun. Nobody in Ur owned a doubling cube and no Roman played to our exact rules, so the five thousand year figure describes a lineage, not one unbroken rulebook. What the record shows is a continuous family of race games: two sides, dice, counters running a fixed track toward home. Games of that shape appear, mutate, borrow from their neighbors and refuse to die, from the Bronze Age to the phone in your pocket. Backgammon is the branch that reached the present with the most players. Calling it five thousand years old is like calling a family name ancient: true as inheritance, loose as fact.

What is the Royal Game of Ur?

In the 1920s, excavators of the royal cemetery at Ur, in what is now southern Iraq, lifted game boards inlaid with shell and lapis lazuli, dated to about 2600 BCE. The game they carried was a race. Pieces entered by dice throw, ran a marked track, and left the board to win. We even have rules for a late version, reconstructed from a Babylonian tablet written around 177 BCE, which means the game lived more than two thousand years on its own. Whether Ur's game is backgammon's grandparent or a great-uncle is honestly uncertain, and scholars still argue the family tree. What the boards prove is enough. People were rolling, racing, and drawing a crowd before most alphabets existed.

Did the Romans play backgammon?

They played its nearest ancient relatives. First came ludus duodecim scriptorum, the game of twelve lines, then a streamlined descendant called tabula, which by most reconstructions used twenty-four points, fifteen checkers a side, opposite directions of travel, hits, and re-entry. That is recognizably our game. The emperor Claudius was said to be so fond of it that he wrote a book on the subject, now lost, and had a board fitted to his carriage. Boards survive scratched straight into the pavement of forums and barracks floors. That detail is the one to keep. Tabula was never only a palace amusement. It was a street game, a tavern game, a soldiers' game.

What is nard, and is it backgammon?

Nard is the Persian branch of the family, and the branch that kept the game alive across the Middle East. Persian tradition credits its invention to the sage Bozorgmehr, in a tale where it answers the arrival of chess from India. That is legend rather than history, but it measures how seriously the game was taken. Nard traveled with the Islamic world from Morocco to Central Asia and never left. Its descendants are on the cafe terraces of Istanbul, Beirut, Cairo, and Tangier tonight, played fast, argued loudly, taught to grandchildren. When people say backgammon has never stopped being played, this is the river they mean.

The rules outlived every empire that carried them.

Where does the name backgammon come from?

Medieval Europe filed the whole family under one word: tables. The book of games compiled for Alfonso X of Castile in 1283 records more than a dozen tables variants side by side, which is what a family looks like before one member gets famous. The English word backgammon surfaces in print in the 1640s, and its origin is still argued. The common guess reads it as back game, since hit checkers must go back and come again. In 1743 Edmond Hoyle published a short treatise that settled the rules for the English-speaking world, and the modern game was, in every important way, on the table.

When was the doubling cube invented?

In the 1920s, in New York's clubs, by somebody whose name nobody wrote down. It is the finest idea added to the game in two thousand years, and by a neat coincidence it arrived in the same decade that dug the oldest boards out of Ur. The cube lets either player offer to double the stakes, and the other choose: play on at double, or concede now. At a stroke the race became a negotiation. Skill stopped being only about moving checkers and became about pricing a position, knowing when your chances justify the offer and when pride is simply paying twice. Nearly every serious game since has been played in the cube's shadow.

How did computers change backgammon?

Beginning in the early 1990s with a research program called TD-Gammon, neural networks taught themselves the game by playing millions of matches against their own copies, and reached world class with almost no human instruction. Analysis engines followed, and theory was rewritten in cold numbers. Opening plays that tradition had settled generations ago were remeasured, and some verdicts flipped. Today any player can roll a position out overnight and read the value of each play to three decimals. The game became harder and more honest at once. Your intuition is still yours; it can just be graded now.

What has stayed the same for five thousand years?

Strip away the names and the centuries, and the constants fit in a short list.

Two players, face to face. The family has cousins for more, but its heart is a duel. Dice, and the permanent argument about luck that arrives with them. A race. First one home wins, and anyone watching can see who is ahead. An audience. Tomb boards, pavement boards, cafe terraces: someone is always leaning in.

We build Orion at the near end of this line, a new table you reach through orion.arthea.ai/go, carrying the one thing no earlier century could offer. The dice prove their own honesty, and anyone can check a finished game at orion.arthea.ai/verify, in the browser, no account needed.]]></content:encoded>
      <category>history</category>
      <category>origins</category>
    </item>
    <item>
      <title>The cafe tradition: where backgammon lives</title>
      <link>https://orion.arthea.ai/news/backgammon-cafe-tradition/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/backgammon-cafe-tradition/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Backgammon lives in cafes: tavla in Turkiye, tawla in the Levant, shesh besh across the region. Tea, kibitzers, and boards handed down explain why.</description>
      <content:encoded><![CDATA[Backgammon lives in cafes. Across the Mediterranean and the Middle East, the board is not something you bring from home; it is furniture, stacked behind the counter with the tea glasses, handed over as naturally as a menu. The game is called tavla in Turkiye, tawla in the Levant, shesh besh across the region, and in every version it is public, loud, and played over tea.

The scene repeats from one end of the map to the other. In Tangier the board opens beside mint tea, with the noise of the medina coming through the window. In Athens the kafeneio keeps three or four boards, their corners gone soft as driftwood. In Istanbul the tea garden runs on tulip glasses and dice calls borrowed from Persian. In Beirut the table moves out to the sidewalk and shares its ashtray with the argileh. Different coffee, different alphabet, same sound: the clack of checkers, steady as rain on a roof.

Why is backgammon played in cafes?

Partly because the game and the coffeehouse grew up together. When the first coffeehouses opened in Istanbul in the middle of the sixteenth century, they needed things for people to do while they stayed, and the board was already ancient, portable, and cheap. Coffee spread through the Ottoman world and the board traveled with it, into Greece, the Levant, Egypt, and the cafes of North Africa. By the time the espresso machine arrived, the pairing was older than most nations on the route.

Partly because of pace. A game takes ten minutes, sometimes five, and talk runs straight through it. Nobody hushes a backgammon table. You roll, you move, you finish your sentence. Chess asks a room to be quiet. Backgammon asks it to lean in.

And partly because of the dice. Luck keeps the ranks close. A grandfather can lose to a twelve-year-old, the table laughs, and the rematch starts before the checkers are back in place. Skill still wins across a hundred evenings, and the regulars know exactly who is good. But on any single night, anyone can beat anyone. That is not a flaw in the game. It is the reason strangers sit down.

What do tavla, tawla and shesh besh mean?

They are mostly the same word. The Latin tabula, a board or table, became tavla in Turkish, tavli in Greek, tawla in Arabic. Across half the world the game is simply named the table, which tells you what it was for: not a possession, a surface you share.

Tavla, in Turkiye: fast and loud in tea gardens and kahvehanes, the rolls still called in numbers borrowed from Persian. Tavli, in Greece: three games played in rotation, portes, plakoto and fevga, best of five in the kafeneio. Tawla, in the Levant and Egypt: sidewalk tables, inlaid boards from Damascus, advice arriving from every chair. Shesh besh, across the region: a game named for a single roll, shesh from Persian for six, besh from Turkish for five.

That last name is the tradition in miniature. One phrase, two languages, passed from mouth to mouth until nobody remembers the borrowing. The game works the same way. It belongs to whoever is holding the dice.

How is the game handed down?

By standing, first. Around any good match there is a second ring of players who are not playing: the kibitzers, the parliament of the cafe, narrating blunders and offering moves nobody asked for. Children start there, carrying tea, watching hands. You learn the rules long before you touch a checker, the way you learn a language years before you can write it.

Then one day an elder waves you into the empty chair, and the real lessons begin. Slow down. Hit when it costs him more than it costs you. Never count your luck out loud. The old players teach the game the way it was taught to them, one proverb at a time, and the losses sting exactly enough to be remembered.

The boards are handed down too. A good one outlives its first owner. The felt goes, the hinge gets replaced, the points wear pale where fifty years of checkers have landed. In Damascus, workshops still set mother-of-pearl into walnut for boards that are bought once and inherited twice. Nobody hangs these boards on a wall. They stay on the shelf, in service.

Why is backgammon social infrastructure?

Because it gives strangers a script. Sitting alone in a cafe is waiting; sitting at a board is an invitation. The game supplies the opening line, the shared object, the reason to stay another half hour. It costs nothing, needs no batteries, never updates, and works for two people who do not share a language. Very little in public life does all four.

An open board means the chair across from it is never really empty.

That is what social infrastructure means: the ordinary equipment that makes company possible. Benches do it, barbershops do it, the tea itself does it. The board does it best, because it brings just enough drama to gather a crowd and keeps every loss small. Losing costs a little pride, and the rematch refunds it.

This is why Orion begins its voyage in a Tangier cafe. We could have started anywhere among our twelve destinations, and Season One, The First Crossing, charts a route far past the medina. But the game feels most at home at a small table with tea on it, so that is where the first board waits. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now.]]></content:encoded>
      <category>culture</category>
      <category>tavla</category>
      <category>tradition</category>
    </item>
    <item>
      <title>Twelve tables: the destinations of Orion&apos;s first voyage</title>
      <link>https://orion.arthea.ai/news/twelve-destinations-of-orion/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/twelve-destinations-of-orion/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Every destination in Orion Backgammon, in voyage order: Tangier cafe to the celestial table, and how you reach each board by playing, not paying.</description>
      <content:encoded><![CDATA[Orion has twelve destinations, and you meet them in order: a Tangier cafe, Athens, Monte-Carlo, Lisbon, Dubai, Istanbul, Santorini, Kyoto, Paris, New York, Gstaad, and finally Orion itself, the celestial table the game is named for. Each one is a place with its own board, its own light, its own sound. You reach them by playing, not by paying, and this is the tour.

We built the voyage this way on purpose. A lobby is a queue, and a queue gives you nothing to look forward to. A ladder of places does. Some of the finest boards in the game are not behind a purchase button at all: they are at the far end of the climb, and the only road there runs through every table in between.

Backgammon has always been a game of rooms as much as rules. The board you learned on had a window, a smell, a certain chair. We wanted the climb to feel like that, so every destination is a place you have to get to, and remember getting to.

The Mediterranean opening

The voyage opens in a Tangier cafe: worn wood, a slow ceiling fan, mint tea going cold beside the board. Everyone who loves this game learned it somewhere, and this is where you learn Orion.

Athens is marble and hard light, a table set among stones that were old before the game had its modern name. Dice have been rolling in this part of the world for thousands of years, and you can feel every one of them watching your opening play.

Monte-Carlo is evening on the water, white hulls in the harbor, a quiet room above it. The pace slows, the doubling gets braver, and the only stake that matters here is pride.

Outward

In Lisbon the table waits in the Alfama, up steep streets, against a wall of blue and white tiles. A song drifts out of a doorway below, and the checkers land softly, the way this city does everything.

Dubai plays high over the marina, glass on every side, towers turning gold at dusk. One of the oldest games on earth sits perfectly at home in one of the newest skylines on it.

Istanbul sets its board on the Bosphorus, ferries crossing between two continents behind your opponent's shoulder. Backgammon is not a pastime here, it is a first language, and the whole city speaks it.

The climb

Santorini is white walls and a caldera that falls away just past the edge of the board. The light is so clean you feel you could read your opponent's whole plan from across the table.

Kyoto is a moon garden: raked gravel, a lantern lit early, the first cool air of the evening. The game goes quiet here and shows you what it has always been underneath, two minds taking turns.

Paris is a palais, tall windows, parquet, gilt catching the last of the daylight. The board sits where it has always deserved to sit: in a beautiful room, taken seriously.

Above it all

New York plays at night, the skyline burning on the far side of the glass. A turn-based game in a city that waits for no one, which is exactly why it works.

Gstaad is snow to the windowsills and a fire that has been going since morning. The chalet table is warm wood and long games, the kind where you stop noticing the time.

And then, at the end of everything, Orion: the celestial table, the one the whole voyage is named for. You do not arrive here early, and we will not describe it too closely, because some tables you should see for the first time in person.

You do not buy your way to the last table. You play your way there.

How do you reach the next destination?

By playing. The voyage moves in order, and every table ahead of you is reached the same way, one match at a time. Trophies start at 0 for everyone, the ten leagues climb from Driftwood all the way to Orion, and the seasons turn quarterly. Season One is called The First Crossing, which is exactly what it is.

And the road is shared: iOS and Android play on one backend, in one pool, so the opponent across the table in Kyoto might be holding a different phone entirely.

There is no shortcut through the shop. Some boards cannot be bought at any price: they are earned, like the exclusive board reserved for the Founding 1000, the first thousand players through the door. The same rule runs through the rest of the game: card packs are earned by playing, never bought. In Orion, the room you sit in says something true about the road behind you.

One more thing the twelve tables share: the dice. They are committed before every game and revealed after it, and the verifier at orion.arthea.ai/verify recomputes each roll in the browser, no account needed.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. The Tangier cafe is the first table behind it.]]></content:encoded>
      <category>destinations</category>
      <category>voyage</category>
      <category>worlds</category>
    </item>
    <item>
      <title>Why Orion is landscape first</title>
      <link>https://orion.arthea.ai/news/why-landscape-first/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/why-landscape-first/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Orion is landscape first because the board is wide. What turning the phone buys: the full board, both home boards, and the cube where hands expect it.</description>
      <content:encoded><![CDATA[Orion is landscape first because the backgammon board is wide. Twenty-four points in two facing rows make a shape that is wider than it is tall, and a portrait screen can only crop that shape or shrink it. So we turn the phone instead of shrinking the game.

There is a small ritual in it. You rotate the phone, the screen stops being a feed, and the device becomes a table. Every real board you have ever played on sat wide between two people. We wanted the first second of every match to remember that.

Why is a backgammon board wide?

Open any board and look. Twelve points on the far side, twelve on the near side, the bar running down the middle. The checkers race around a horseshoe toward home, and the story of the game reads left to right and back again. Height barely matters. Width is everything: primes stretch sideways, blockades stretch sideways, the race itself stretches sideways.

This is not a screen convention. It is thousands of years of tables. The game's ancestors were played on wide boards in Mesopotamia, in Rome, in Ottoman coffeehouses, and the shape survived because it fits two people facing each other with a table between them. The board was wide before screens existed. It will be wide after ours are gone.

What does portrait do to the game?

It offers two bad options. The first is to shrink the board until it fits a tall frame, and now a stack of five checkers is a smudge. You squint to count, and counting is the game. The second is to crop or fold it: hide the far half behind a swipe, stack the board in pieces, make you scroll a position you should see whole.

A backgammon position is read whole or not at all. Should you double? The answer lives in both home boards at once, in the shape of two primes, in a race counted across the full width. Portrait asks you to judge a table through a keyhole. We were not willing to ask you that.

Portrait is also the shape of feeds, and feeds are built for passing through: skim, flick, move on. A match is not passed through. It is sat at. The rotation is a threshold, the same gesture as pulling out a chair.

What does landscape actually buy?

The whole board at readable size: all twenty-four points on screen, no folding, no scrolling, nothing hidden behind a gesture. Both home boards visible at once, so races and cube decisions read at a glance. Checker stacks tall enough to count without squinting. The doubling cube where your hands expect it, on the table between you, not buried in a menu. Tablets treated as first-class tables, with the wide stage growing into the screen instead of stretching a phone layout.

The cube deserves the emphasis. Doubling is the deepest decision in backgammon, and it is a whole-board judgment: your structure, their structure, the race, the threats. Offered on a full wide board, the cube feels like what it is, a question asked across a table. Crammed into the corner of a tall screen, it feels like a notification.

A feed is tall. A table is wide. We built a table.

Doesn't landscape make typing worse?

Yes. That is the honest cost, and we will not pretend otherwise. A landscape keyboard eats half the screen, thumbs travel farther, and nobody enjoys writing sideways. We knew this when we chose.

So the design absorbs it. Typing lives at the edges of Orion, never in the middle of play. A match runs on fingertips, taps and drags from the opening roll to the last checker borne off, and nothing in a game ever waits on a keyboard. The words worth typing are short and rare. If a screen forced you to write a paragraph sideways, we would call that a design failure and fix the screen, not your patience.

What does landscape first mean for tablets?

A tablet in landscape is the closest a screen has come to a real board. Because the wide stage is the design, a bigger screen does not stretch a phone layout; it lets the table become more itself, nearer the size your hands know from wood and felt. We build natively on both platforms, SwiftUI on iOS and Jetpack Compose on Android, and everyone plays in one shared pool. The person across the table might hold a phone in Lisbon while you sit at a tablet that almost feels like furniture.

What does respect the board mean?

It is the principle underneath every screen we make. When the interface and the game disagree, the interface adjusts. The board has been wide for thousands of years; portrait phones have existed for about twenty. That was never a close call, and the same rule settles smaller arguments every week. If a panel would crowd the points, the panel moves. If an effect would cover a checker, the effect goes.

All twelve destinations are built on that promise, from a cafe table in Tangier to the celestial table called Orion, every one of them wide. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. When you step through, turn your phone sideways. That is the whole ceremony. The table does the rest.]]></content:encoded>
      <category>Design</category>
      <category>Behind the game</category>
    </item>
    <item>
      <title>Seasons, leagues and trophies: how the ladder works</title>
      <link>https://orion.arthea.ai/news/seasons-and-trophies-explained/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/seasons-and-trophies-explained/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>Trophies come from rated play only. Everyone starts at zero and climbs ten leagues, Driftwood to Orion, in quarterly seasons. Nothing you buy helps.</description>
      <content:encoded><![CDATA[Trophies in Orion come from rated play and from nowhere else. Everyone starts at zero and climbs ten leagues, from Driftwood at the bottom to Orion at the top, through seasons that run quarterly. Nothing you can buy adds a trophy, skips a league, or makes the dice kinder.

That is the whole system in three sentences. What follows is the reasoning, because a ladder is a promise about what a number means, and promises are worth spelling out.

How do you earn trophies?

You win rated matches. Rated play runs on our servers, where every move is validated at source and every roll is committed before the game begins, so a result is a fact before it becomes a number. When the match ends, the ladder moves. There is no other door into the count: not a purchase, not a pack, not a promotion. If a profile shows a thousand trophies, a thousand trophies' worth of backgammon happened.

Not every game is rated, and that is deliberate. Play for the joy of it and the ladder holds still. Rated matches are the ones where both players sat down to be counted.

What are the ten leagues?

The climb passes through ten leagues. It begins in Driftwood, the shore where everyone washes up, and it ends at Orion, the celestial table the game is named for. We named leagues like places instead of ranks because the whole game is built that way: twelve destinations, one voyage, and progress that reads as distance travelled rather than a spreadsheet row. A league tells you where you are on the crossing. You are not assigned to one. You arrive.

Why does everyone start at zero?

Because a number you were handed is not a record. Many games seed a new profile partway up the ladder so the first session feels generous. We understand the instinct and we refused it. If the count opens at three hundred, then three hundred means nothing, and every number above it inherits the doubt. Zero is the only honest starting line. Your first trophy is the first true sentence on your profile, and you are the one who wrote it.

Zero also settles an old argument before it starts. Nobody can ask which part of your standing was a gift, because none of it was. The record is the seasons you played. Nothing more, and nothing less, which is exactly what makes it worth reading.

How long do seasons last?

A season runs a quarter of the year. The first is called The First Crossing, and the name is literal: nobody has crossed from Driftwood toward Orion before, and once this season closes, nobody will ever be first again. A season is a chapter with a closing date, and the date gives the climb its shape: enough weeks to matter, a horizon near enough to see. Quarterly is a deliberate tempo. Fast enough that a season never goes stale, slow enough that a climb is a story rather than a sprint.

What resets when a season ends, and what stays?

The rule is short. The season resets. You do not.

Resets: the climb. Standings close, the chapter is archived, and the next season opens a fresh ladder for everyone. Stays: your name, your history, and everything you earned along the way. Earned things remain yours. That is what earned means. Permanent: founder identity. The first thousand players, the Founding 1000, carry that mark through every season that will ever run.

So the ladder is seasonal and the record is cumulative. Years from now, the interesting thing on a profile will not be a trophy count from week two of some quarter. It will be the shelf of seasons behind it: which crossings you made, and how far each one carried you.

Can you pay to climb faster?

No, and we mean the strong version of no. Coins and gems are virtual, they have no real-world value, and there is no cash-out. Nothing in the shop touches the ladder. Card packs are not even for sale: they are earned by playing, only by playing. And the dice cannot be bribed at any price. Every rated roll traces back to a SHA-256 commitment published before the game starts, each roll is drawn by HMAC-SHA256, and the seed is revealed when the game ends. You can recompute any finished rated game in your browser at orion.arthea.ai/verify, without an account. Kinder dice are not on the menu, because nobody can reach into the rolls. Including us.

A ladder you can buy is a receipt. A ladder you climbed is a record.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. Behind it, The First Crossing begins the way every season after it will: all of us at zero, ten leagues ahead, and dice anyone can check.]]></content:encoded>
      <category>seasons</category>
      <category>leagues</category>
      <category>trophies</category>
    </item>
    <item>
      <title>A season pass without the dark patterns</title>
      <link>https://orion.arthea.ai/news/the-season-pass-philosophy/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/the-season-pass-philosophy/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>How Orion designs its season pass: cosmetics and pace, never power. No anxiety timers, no pay-to-win tier, and a free track worth playing on its own.</description>
      <content:encoded><![CDATA[A fair season pass sells two things: how your table looks and how fast the rewards arrive. It never sells power. On Orion the premium pass cannot touch your dice, your matchmaking, or your opponent's luck, and the free track is a real journey rather than a row of consolation prizes.

This is a design position, so we will state it plainly. Before we built our pass, we wrote down the ways season passes go wrong. Then we built against the list, and we are publishing the rules so you can hold us to them.

What do season passes usually get wrong?

Most battle passes are not designed around the game. They are designed around the moment you almost stop playing. The craft that should go into the play goes into the pressure instead, and it usually takes three shapes.

Timers tuned to anxiety. The countdown is not there to structure a season. It is there to sit in your chest on a Sunday night, until a purchase feels like relief. Paid tiers that gate power. The stronger piece sits behind the price, and every match quietly becomes a comparison of receipts. Prices built to obscure value. You buy a currency to buy a bundle to unlock a tier, until nobody can say what anything costs. The fog is the point.

None of this is accidental. These systems are tuned by talented people against precise metrics, and the metrics do not measure joy. That is why the fix has to be a rule written down in advance, not a mood we hope to keep.

What does a fair season pass sell instead?

Cosmetics and pace. That is the whole list. The premium track in Orion carries the look of your table and the identity you bring to it, and it moves you through the season's rewards faster. It does not carry a single point of playing advantage.

The deeper promise is what the pass cannot do. It never makes the dice kinder. It never softens your matchmaking. It never buys a stronger checker, a longer clock, or a gentler opponent. Two players at the same board are equals, whatever their receipts say.

The dice do not know what you paid. Neither does the matchmaker.

You do not have to take that on faith. Every live roll in Orion comes from a commit and reveal scheme. The server publishes a SHA-256 commitment before the game begins, draws each roll with HMAC-SHA256, and reveals the seed when the game ends. The verifier at orion.arthea.ai/verify replays the whole sequence in your browser, no account needed. A pass that bent the dice would be exposed by our own tool.

Is the free track actually worth playing?

Yes, and this is where most passes give the game away. A free track built as an advertisement teaches you that playing is worth less than paying. We built ours as a season you could complete happily without spending anything: real rewards at an honest rhythm, earned the only way anything should be earned at a backgammon table, by playing. If a free player finishes The First Crossing proud of the trip, the pass did its job.

Card packs make the rule visible. In Orion, packs are earned only. They are never sold, at any price, inside any bundle. Coins and gems are virtual, with no cash value and no way to cash out. The marks of a voyage are things you did at the board, not things you found in a shop.

What is the Founding 1000?

When the doors open, the first thousand players to arrive receive the first premium pass free, along with a permanent founder identity and an exclusive board. It is a mark of arrival, not a discount lever. Season One is called The First Crossing, and the people who make that crossing first should be able to point to it forever.

Then it closes. One thousand means one thousand: no extended window, no second wave, no anniversary reissue. We are told this leaves money on the table. It does. Scarcity is only honest when the number holds, and a founder mark only means something if we refuse to mint more of it.

Why is restraint the brand?

Because backgammon does not need the machinery. The game is around five thousand years old and has outlived every empire that ever hosted a board. Something with that much gravity does not need a countdown tuned to your pulse. It needs a table worth sitting at, in a cafe in Tangier or by a window in Kyoto, and time enough to enjoy it. Our seasons run quarterly, a quarter of real life, so the pass is something you live alongside rather than something you sprint through on the final weekend.

And because trust compounds. We built dice you can verify, a server that validates every live move at source, and a public page where anyone can check a game. All of it would be worth nothing if the shop played by different rules than the board. Restraint is not a limitation we accepted. It is the product.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now, and The First Crossing is waiting on the other side.]]></content:encoded>
      <category>season pass</category>
      <category>fair play</category>
      <category>design</category>
    </item>
    <item>
      <title>Shipping backgammon in 43 languages</title>
      <link>https://orion.arthea.ai/news/building-43-languages/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/building-43-languages/</guid>
      <pubDate>Sat, 01 Aug 2026 00:00:00 GMT</pubDate>
      <description>How we shipped a backgammon game in 43 languages: 884 strings per locale, RTL mirrored by the layout system, a CI guard for gaps, fonts tested per script.</description>
      <content:encoded><![CDATA[Orion ships in 43 languages by treating translation as engineering, not decoration: 884 strings per locale live in catalogs, never in code, and a CI guard fails any build that carries an untranslated line. Right-to-left screens mirror themselves through the layout system, and every font has to prove, on pixels, that it can draw every script we ship. Here is how that holds on two native platforms, and why an English-only backgammon game would have been a contradiction.

Why localize a backgammon game at all?

Because backgammon is not from the English-speaking world. It grew up around the Mediterranean and the Levant: tavla over tea in a Turkish cafe, tavli by a Greek harbor, shesh besh in Beirut, takhteh in Iran. The game was old in Arabic, Persian, and Greek before it had an English name. Orion's destinations follow the same map, Tangier and Athens and Istanbul among them, so an English-only release would have painted the cafe and locked its door. If the game belongs to those places, it has to speak their languages. That decision set the number: 43 languages, right-to-left scripts included, on both platforms.

How do you manage 884 strings across 43 languages?

By refusing to let a single one live in code. Every label, button, rule text, and victory line is a key, and the keys resolve through each platform's native catalog: a String Catalog on iOS, string resources on Android. Multiply 884 strings by 43 locales and you get 38,012 lines of text to keep true. At that scale, copy pasted straight into a screen file is a time bomb: no tool can see it, and no translator will ever be shown it. Keys keep the whole surface visible. When a sentence changes, one key changes, and 43 languages queue up behind it. The iOS app is native SwiftUI, the Android app is native Jetpack Compose, one shared backend sits behind both, and the same key tells the same truth on every phone at the table.

How does right-to-left support work in a game?

Through the layout system, not by hand. We never built mirrored copies of screens for Arabic. Every layout is written in terms of leading and trailing instead of left and right, and SwiftUI and Compose do the rest: when the locale reads right to left, the interface flips itself. Navigation runs the other way, back arrows reverse, text sits on its natural edge. The real work is discipline about what you type. Write the word left once, because it looks fine on your English phone, and you have planted a screen that lies in Arabic. Hand-mirrored screens rot the first week nobody checks them. Mirroring owned by the layout system cannot drift, because there is nothing to forget.

What stops an untranslated string from shipping?

A guard that fails the build. Missing translations are quiet by design: platforms fall back to English without raising an error, so an English-speaking team can stare at a broken Turkish screen and see nothing wrong. Our CI counts translation gaps across all 43 locales on every build, and the number must read zero. One missing key in one language fails the build, loudly, before any store ever sees it. That sounds severe until you picture the alternative: a lone English sentence in the middle of an Arabic season screen, found by exactly the player it disrespects.

How do you test a font against every script?

A font that silently drops glyphs is a lying font.

Strings fail toward English. Fonts fail toward nothing: a blank box where a letter should be, or a silent swap to some fallback that changes the look mid-word. A display font can be gorgeous in Latin and empty in Arabic, and its coverage table will not volunteer that. So we test pixels. We render real strings from every one of the 43 locales in the fonts we actually ship, look at the output, and keep the fallback chain we have watched draw every script with our own eyes. Trust what renders. Nothing else counts.

Why 49 store listings for 43 languages?

Because the stores keep their own maps of the world. Both stores split some languages by region: Portuguese becomes Brazil and Portugal, Spanish becomes Spain and Latin America, English appears more than once. By the time both catalogs were covered, 43 in-game languages had become 49 store listings, each with its own description and screenshots. It is unglamorous work, and it is also the first thing a player ever reads. The listing should greet you the way the game will.

Four rules carried the whole effort:

No string lives in code. Every line has a key in the platform catalog. No hand-mirrored screens. Right-to-left comes from leading and trailing, enforced by the layout system. No silent gaps. CI fails the build if one key in one locale is untranslated. No trusted fonts. Every script is rendered and checked on pixels before it ships.

The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. Step through it and the game greets you in whichever of its 43 languages your phone already speaks.]]></content:encoded>
      <category>localization</category>
      <category>RTL</category>
      <category>engineering</category>
    </item>
    <item>
      <title>The First Crossing</title>
      <link>https://orion.arthea.ai/news/the-first-crossing/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/the-first-crossing/</guid>
      <pubDate>Sat, 25 Jul 2026 00:00:00 GMT</pubDate>
      <description>Season One, The First Crossing, opens on 1 September 2026. Twelve destinations across the Mediterranean and beyond, one board, one long voyage.</description>
      <content:encoded><![CDATA[Season One has a name. The First Crossing.

It opens on 1 September 2026, a named day rather than the hour each store happened to approve us, so that nobody starts the climb early. It begins where backgammon has always felt most at home. On the Mediterranean, in the late light, at a table someone has already set for you.

Twelve destinations

The voyage starts in Tangier, at Hassan's cafe, where the oud plays and strangers become rivals before the second game. From there it moves outward: marble in Athens, velvet and brass in Monte-Carlo, tiled alleys in Lisbon, gold on the water in Dubai.

Then it climbs. Tea on the Bosphorus. Sunset over Santorini. Raked gravel in a Kyoto garden at moonrise. Chandeliers in Paris. Ninety floors above a New York avenue. Snow on the peaks at Gstaad.

And at the end of it, above all of them, Orion. The celestial table. You do not arrive there early.

Each destination has its own host, its own room, its own light. We wanted them to feel less like levels and more like somewhere you had arrived.

Why places

Because backgammon has always belonged to places. You remember the cafe where someone first taught you. The old board in a relative's house. The friend who always doubles too early. The sound of a checker landing on wood.

A lobby cannot hold that. A destination can. So we built twelve of them, and a reason to cross from each one to the next.

What the crossing asks of you

Ten leagues form the ladder, from Driftwood to Orion. Your place on it comes from rated play, so the climb is not decoration. Trophies start at zero for everyone, on the same day, on the same board.

Some of the finest boards in Orion are not behind a purchase button. You reach them. That is the point of a voyage: at the end of it, the things you carry should say something about where you have been.

Season One is not the game. It is where the game starts.

At the opening

The tables are set. The hosts are waiting. Orion is already in the sky above them, and the first crossing begins.

Roll under the stars.]]></content:encoded>
      <category>Season One</category>
      <category>Announcement</category>
    </item>
    <item>
      <title>The Founding 1000</title>
      <link>https://orion.arthea.ai/news/the-founding-1000/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/the-founding-1000/</guid>
      <pubDate>Sat, 25 Jul 2026 00:00:00 GMT</pubDate>
      <description>There is only one first crossing. The first thousand players to join Orion keep a permanent mark of the season they arrived in.</description>
      <content:encoded><![CDATA[There will only ever be one first crossing.

Every game that lasts long enough eventually has a group of people who were there at the start. Usually nobody thinks to mark it until years later, when it is too late to prove.

We are marking it now.

What a founder receives

The Season One premium pass, in full. A permanent Founder identity that stays on the account for as long as it exists. An exclusive board, made for the opening of the voyage.

The board does not come back. It will not appear in the store in a later season, and it will not be sold when the number is full. It belongs to this moment, and only to it.

Why scarcity, and why only here

Scarcity is easy to fake and easy to abuse. So we are keeping it to one thing: the record of when you arrived. Nothing a founder holds makes the dice kinder or the board easier. It is a mark, not an advantage.

That distinction matters to us. A game with real competition cannot sell the competition. It can only sell what the competition looks like.

Years from now, if you find one across the table, you will know exactly when that player arrived.

One thousand, then closed

The count is the first thousand accounts of Season One. When it is reached, the Founding 1000 closes for good and the voyage carries on without it.

The first crossing begins on 1 September 2026, and the game is downloadable before it. After that day, everyone who joins is arriving somewhere that already exists. Which is its own pleasure, but it is not the same one.]]></content:encoded>
      <category>Season One</category>
      <category>Founders</category>
    </item>
    <item>
      <title>How Orion proves every dice roll is fair</title>
      <link>https://orion.arthea.ai/news/provably-fair-dice/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/provably-fair-dice/</guid>
      <pubDate>Sat, 25 Jul 2026 00:00:00 GMT</pubDate>
      <description>Orion commits to the dice before a game starts and reveals the seed when it ends, so any player can recompute every roll and prove it was never changed.</description>
      <content:encoded><![CDATA[In most online dice games you are asked to trust the operator. You roll, a number appears, and the only evidence that it was not chosen for you is the operator's word. Orion is built the other way around: the dice are locked in before the game starts, and when the game ends you get everything you need to check them yourself.

The commitment comes first

Before your first roll, the server generates a secret seed and publishes a commitment to it: the SHA-256 hash of that seed. The hash reveals nothing about the seed, but it pins it down. Once that commitment is published, the server cannot change the seed without the hash no longer matching.

Before the game: The server publishes sha256(serverSeed). The seed itself stays secret while you play.

During the game: Every roll is derived with HMAC-SHA256 from that same sealed seed and the roll's position in the game. Nothing is sampled at the moment you tap.

After the game: The seed is revealed. Anyone can hash it, compare it to the commitment published before the first roll, and then recompute every die.

Why this is checkable, not just claimed

A commitment scheme only means something if outsiders can actually run the check. So the verifier is a public page, it needs no account, and the cryptography runs entirely in your own browser using the Web Crypto API. Our servers are not asked whether a game was fair; your machine does the math and tells you.

Every finished Orion game has an ID. Paste it into the verifier and you can replay the game move by move, see the commitment that was published before the first roll, see the revealed seed, and see each die recomputed from it.

What this does not do

Provable fairness proves the dice were not tampered with. It is not a claim about matchmaking, opponent skill, or luck evening out over a session. Those are separate problems and we would rather name the boundary than blur it.

Where Orion is today

The fair-dice system, the public verifier and the game itself are built and running. The door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available on each right now. The verifier is live and you can inspect a real finished game with it today.]]></content:encoded>
      <category>Provably fair</category>
      <category>How it works</category>
    </item>
    <item>
      <title>What we are building: Orion, a backgammon voyage</title>
      <link>https://orion.arthea.ai/news/what-we-are-building/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/what-we-are-building/</guid>
      <pubDate>Sat, 25 Jul 2026 00:00:00 GMT</pubDate>
      <description>Orion is a premium backgammon game built as a voyage across twelve destinations, with provably fair dice, leagues and clubs. Here is what exists today.</description>
      <content:encoded><![CDATA[Orion is a backgammon game made with the care usually reserved for far bigger genres. The rules are the ones you already know. Everything around them, the table you sit at, the rival across from you, the reason to come back tomorrow, is what we have spent our time on.

A voyage, not a lobby

Instead of a single anonymous table, Orion is a journey across twelve destinations, each with its own host, its own room, and its own stakes. You begin at Hassan's cafe in Tangier and climb toward Orion itself, the table at the end of the voyage.

Twelve destinations, from Tangier Cafe to Orion, each with a resident host and their own board. Leagues that run from Driftwood to Orion, so a ladder always sits above you. Clubs and friends, with private one-on-one challenges. Provably fair dice on every server-backed table, checkable by anyone.

Fair by construction

The dice are the heart of a dice game, so they are the part we refused to ask for trust on. Orion commits to its dice before a game begins and reveals the seed afterwards, and a public verifier lets anyone recompute the rolls in their own browser.

Where we are

Season One, The First Crossing, opens on 1 September 2026. That date is ours and it is fixed. Per-store availability is not ours, so we do not freeze it into a sentence: the download door at orion.arthea.ai/go carries the App Store and Google Play links and always shows what is available right now. The verifier is already public, and this news section is where changes get recorded as they happen.

The dice are locked in before the game starts, so nobody, not even us, can change them.]]></content:encoded>
      <category>Announcement</category>
      <category>Behind the game</category>
    </item>
    <item>
      <title>One board, many names</title>
      <link>https://orion.arthea.ai/news/one-board-many-names/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/one-board-many-names/</guid>
      <pubDate>Fri, 24 Jul 2026 00:00:00 GMT</pubDate>
      <description>Tavla, tavli, nard, shesh besh, gamao, backgammon. The same twenty-four points, carried across empires and languages. Orion ships in 43 of them.</description>
      <content:encoded><![CDATA[Backgammon has crossed empires, borders and languages without ever needing to explain why it should still exist. Boards have been carved into wood, stone and leather. Games have unfolded in homes, cafes, clubs, ports and hotel rooms from Istanbul to London, from Beirut to Paris.

The names changed along the way.

Tavla. Tavli. Nard. Shesh besh. Gamao. Backgammon.

The board endured.

A game with no single home

Ask three players where the game belongs and you will get three answers, all of them sincere and none of them wrong. A Turkish cafe at midday. A Greek terrace. A living room in Tehran. A club in Beirut. A kitchen table in a family that has played it for four generations without once calling it a hobby.

Different names, different cities, different customs at the table. Some double freely. Some consider it rude. Some play to five, some to seven, some until someone has to leave.

Same twenty-four points.

What that meant for how we built it

A game this widely held cannot ship in one language and call itself finished. Orion is built for 43, including right-to-left layouts where the interface has to mirror rather than merely translate.

That work is invisible when it is done properly. A player opens the app in Arabic or Hebrew and the board is simply where it should be, the panels read in the right direction, and nothing about the game feels like it was made somewhere else and shipped over.

The people who call this game backgammon are only one part of its story.

The point of twelve destinations

Orion's voyage runs through Tangier, Athens, Monte-Carlo, Lisbon, Dubai, Istanbul, Santorini, Kyoto, Paris, New York and Gstaad before it reaches the celestial table.

That route is not a travel brochure. It is an acknowledgement. This game was never ours to reinvent. It arrived already loved, in a dozen languages, and the most useful thing a studio can do is build it a room worthy of the company it has kept for five thousand years.]]></content:encoded>
      <category>Behind the game</category>
      <category>The world</category>
    </item>
    <item>
      <title>The board remains sacred</title>
      <link>https://orion.arthea.ai/news/the-board-remains-sacred/</link>
      <guid isPermaLink="true">https://orion.arthea.ai/news/the-board-remains-sacred/</guid>
      <pubDate>Fri, 24 Jul 2026 00:00:00 GMT</pubDate>
      <description>We changed much around backgammon and nothing inside it. Fifteen checkers, twenty-four points, two dice, and a cube that still ruins comfortable positions.</description>
      <content:encoded><![CDATA[For everything we changed around backgammon, there was one thing we never intended to touch.

Backgammon itself.

There are still fifteen checkers. There are still twenty-four points. There are still two dice entirely capable of ruining an excellent plan.

A blot is still a risk. A prime is still a wall. The doubling cube can still turn a comfortable game into a test of nerve, and the deeper you go the more the game gives back.

The board is the game

Most mobile games treat the play area as one component among many. We went the other way. Every checker lift, hit, roll and bear-off was built to feel physical and immediate, on warm wood, felt and stone rather than a flat interface with a board pasted into the middle.

Nothing floats over the felt that does not need to be there. When the match is running, the season, the quests, the ladder and the collection all recede. Two players remain.

Where the ambition went instead

Into the parts you are not supposed to notice. A rules engine built and tested separately from the interface, against the details that make real backgammon awkward to implement: bar entry, forced moves, higher-die obligations, bearing off, gammons, backgammons and the cube.

Into online matches that are settled by the server rather than trusted to the device in someone's hand. Into dice you can verify yourself after the game rather than take on faith.

Most players will never see any of it, and that is exactly right. The engineering exists so that when the cube slides across the board, the only thing on your mind is whether you are going to take it.

We are not building a retention system that happens to contain backgammon. We are building backgammon that gives you reasons to return.]]></content:encoded>
      <category>Behind the game</category>
      <category>Design</category>
    </item>
  </channel>
</rss>
