Frequently asked questions.

Yes — that's the whole idea. Dix-de-Der is built to play offline against the AI: on the train, on a plane, in the countryside. Only multiplayer, the daily challenge and the leaderboards need a connection.

No. No tokens, no virtual currency, no aggressive storefront. The game is free, financed by discreet ads; an optional subscription removes them and unlocks a few extras. You'll never be blocked mid-game for lack of "credits" — in belote, you don't pay to play your cards.

In classic belote, trump is decided on a turned-up card: "je prends" or "je passe". In coinche, you announce a contract ("90 hearts") that the opponents can coinche (double) — and you play with declarations. Contrée is coinche's cousin; Dix-de-Der optionally offers its expert denominations No Trump and All Trump. The full details are in the rules.

Tierce (20), cinquante (50), cent (100) and carrés (100 to 200) are declared on the first trick and resolved on the second: only the side holding the highest declaration validates all of its own. The app announces and compares automatically — the full table is in the rules.

In classic belote, when takers and defenders finish in a perfect tie (81 to 81, or 91 to 91 with a belote), the taker's points are held in reserve: they go to whichever team wins the next deal. The defence marks its points straight away.

Yes: create a free account, add your friends and set up your table — empty seats are taken by the AI. Built-in voice chat lets you talk during the game, like in the living room. You can also share a deal ("beat my score") to challenge a friend on exactly the same cards.

Your rating shows how you play your cards, not whether you win. After each deal, the app replays the same deal with the AI in your seat — same cards, same partners. The AI's score shows what was achievable; your result is compared to it. So your decision counts, not your luck. You start at 5000.

The details are in the "Rating" chapter of the rules.

Absolutely not. Every AI decision rests solely on the course of play, the cards played, and what a player at the table can legitimately deduce. Nobody looks at your hand, and the shuffle is entirely random.

We have no interest in making you lose either: we'd rather have satisfied players who come back for a good game than players frustrated by a rigged one.

The three virtual players are instances of the same program: your partner is exactly as strong — or as weak — as your opponents. The opponents seem cleverer because they understand each other perfectly and never miscount cards. An unintended signal from you can also mislead your partner: it assumes you're playing correctly.

In Settings → full settings → AI brain, you pick your opponent. Tree is the classic brain: it follows fixed rules we wrote ourselves — always built in, always ready. Neural is a learning brain that trained itself by playing an enormous number of games. Both are serious opponents. See the "AI opponent" chapter of the rules.

Once. The neural brain is a set of trained model files: choose Neural and the app fetches them (a small one-time download); after that everything works offline. If they're missing, the app offers to fetch them again. When an improved brain comes out, the app lets you know.

Of course. The AI is demanding, not unbeatable: a good human player still beats it. Think of it as a solid club player. Better still: your well-played games can help us make it even stronger — see "Help improve the AI".

At Monkey and Beginner the opponents deliberately play weaker so you can learn in peace. At Pro the AI plays at full strength. Because measuring on the easy levels would unfairly inflate your rating, the rating only changes at Pro. (Your partner North always plays at full strength.)

That's the virtual players' chatter — short reactions to the game, purely for atmosphere. It has no influence on your rating and can be turned off in the settings ("Opponent chatter").

Some chat replies are AI-generated and can occasionally be off. See the privacy policy.

A voluntary option that helps us train the neural brain. It learns most from well-played games by good players. Switch it on (in the full settings) and we may collect your well-played games anonymously — no username or personal details, only how the game went (cards, tricks, score). Off by default; you can switch it off at any time. (Details in the privacy policy.)

Yes. When someone shares a deal, you play exactly the same cards against the same AI as they did: the comparison holds. (If the AI version differs, a small note may say the comparison is "indicative".)

A neural brain (a "neural network") is a program that learns from examples instead of following fixed rules. Compare it to learning a language: the classic brain (Tree) memorised a grammar book, while the neural brain picked the language up by listening and practising a great deal. That's why it often has a better feel for the right move in a given spot.

By playing an enormous number of games — against itself and against earlier versions — and remembering what worked each time. It improved step by step, without us dictating every good move. That's called machine learning. (And your well-played games can contribute — see question 15.)

No. It only sees what any player at the table sees: its own cards, the cards played, and the declarations. It makes an estimate based on the patterns of the game — it doesn't peek at your hand. (See also "Does the app cheat?".)

Tiny calculation differences between devices can tip a single move. It doesn't change how strong it is — it remains a tough opponent.

Your rating uses the AI as a yardstick: after your game it replays the same deal to show what was achievable. The yardstick is the same for everyone — that's what makes the rating comparable. (See "How does the rating work?".)

At an online table, empty seats are taken by virtual players. Each is run by one of the human players' devices and uses the brain that player chose — usually the classic Tree brain. The rating does not apply to multiplayer.

In the daily challenge, everyone plays against exactly the same AI — the classic Tree brain — with the same cards, so the worldwide ranking is fair. The neural brain is deliberately not used there: it needs a download and can vary very slightly per device, which would skew a global contest.

Open the app settings and use "Restore purchases" — it reactivates an already-paid subscription, for instance after switching devices. If the problem persists, write to us: support@dixdeder.fr.

We're always open to good ideas for improving the game — many suggestions from players of our sister app Klaverjas HD have made it into the app (backgrounds, sounds, game-pace settings, etc.). Got an idea? Write to us — support@dixdeder.fr.