36 combinations of a 2-dice roll
| 1 - 1 | 2 - 1 | 3 - 1 | 4 - 1 | 5 - 1 | 6 - 1 |
| 1 - 2 | 2 - 2 | 3 - 2 | 4 - 2 | 5 - 2 | 6 - 2 |
| 1 - 3 | 2 - 3 | 3 - 3 | 4 - 3 | 5 - 3 | 6 - 3 |
| 1 - 4 | 2 - 4 | 3 - 4 | 4 - 4 | 5 - 4 | 6 - 4 |
| 1 - 5 | 2 - 5 | 3 - 5 | 4 - 5 | 5 - 5 | 6 - 5 |
| 1 - 6 | 2 - 6 | 3 - 6 | 4 - 6 | 5 - 6 | 6 - 6 |
Number of pips per dice roll: On average, we advance 8.2 pips per dice roll
| Dices | 1 | 2 | 3 | 4 | 5 | 6 |
| 1 | 4 | 3 | 4 | 5 | 6 | 7 |
| 2 | 3 | 8 | 5 | 6 | 7 | 8 |
| 3 | 4 | 5 | 12 | 7 | 8 | 9 |
| 4 | 5 | 6 | 7 | 16 | 9 | 10 |
| 5 | 6 | 7 | 8 | 9 | 20 | 11 |
| 6 | 7 | 8 | 9 | 10 | 11 | 24 |
Probability of hitting a blot
| Number to be made | Number of favorable rolls | Probability of hitting |
|---|---|---|
| 1 | 11 | 31% |
| 2 | 12 | 33% |
| 3 | 14 | 39% |
| 4 | 15 | 42% |
| 5 | 15 | 42% |
| 6 | 17 | 47% |
| 7 | 6 | 17% |
| 8 | 6 | 17% |
| 9 | 5 | 14% |
| 10 | 3 | 8% |
| 11 | 2 | 6% |
| 12 | 3 | 8% |
| 15 | 1 | 3% |
| 16 | 1 | 3% |
| 18 | 1 | 3% |
| 20 | 1 | 3% |
| 24 | 1 | 3% |
Probability of a checker re-entering
| Number of available squares | Number of favorable rolls | Probability of re-entering |
|---|---|---|
| 6 | 36 | 100% |
| 5 | 35 | 97% |
| 4 | 32 | 89% |
| 3 | 27 | 75% |
| 2 | 20 | 56% |
| 1 | 11 | 31% |
| 0 | 0 | 0% |
Probability of the last checker bearing off in a single roll
| Position | Number of favorable rolls | Probability of exit |
|---|---|---|
| Square 6 | 27 | 75% |
| Square 5 | 31 | 86% |
| Square 4 | 34 | 94% |
| Square 3 | 36 | 100% |
| Square 2 | 36 | 100% |
| Square 1 | 36 | 100% |
Probability of the last two checkers bearing off in a single roll
| Position | Number of favorable rolls | Probability of exit |
|---|---|---|
| Square 6-6 | 4 | 11% |
| __ 6-5 | 6 | 17% |
| __ 6-4 | 8 | 22% |
| __ 6-3 | 10 | 28% |
| __ 6-2 | 13 | 36% |
| __ 6-1 | 15 | 42% |
| Square 5-5 | 6 | 17% |
| __ 5-4 | 10 | 28% |
| __ 5-3 | 14 | 39% |
| __ 5-2 | 19 | 53% |
| __ 5-1 | 23 | 64% |
| Square 4-4 | 11 | 31% |
| __ 4-3 | 17 | 47% |
| __ 4-2 | 23 | 64% |
| __ 4-1 | 29 | 81% |
| Square 3-3 | 17 | 47% |
| __ 3-2 | 25 | 69% |
| __ 3-1 | 34 | 94% |
| Square 2-2 | 26 | 72% |
| __ 2-1 | 36 | 100% |
| Square 1-1 | 36 | 100% |
Here, you can manually enter a backgammon position from, for example, a real game.
This is a simple way to review your decisions after a game, understand what you could have played differently, and practice situations that give you trouble.
Simply enter the position point by point, the dice values, and click 'Play this position'.
This feature is designed exclusively for educational and post-game use. The goal is to help you improve, not to assist an ongoing game. MBT does not offer any automation for real-time online play and cannot be held responsible for how you use it.
Minimalist Backgammon Trainer (MBT) is a game situation simulator. Its goal is to improve your ability to choose the perfect move in realistic game situations. No stakes, no competition, no stress, just you and your skills, just thinking about the best move.
When you press Play, a unique game will begin without the Jacoby Rule and without the doubling cube. A random number of moves will be determined by GNU Backgammon (the engine on which MBT is based). You can choose between 'Mix', 'Opening', 'Mid Game', and 'End Game', which will affect the number of moves that will be played. By choosing 'Mix', MBT will play between 0 and 27 moves. By choosing 'Opening', MBT will play between 0 and 4 moves. By choosing 'Mid Game', MBT will play between 5 and 16 moves. By choosing 'End Game', MBT will play between 17 and 27 moves. For this configuration, the game will either be well underway or in a bear-off situation (These different intervals were chosen intuitively based on the analysis of my games).
Each turn, GNU will play what it considers to be the best move. The evaluation made by GNU Backgammon during the simulation to decide on these moves is based on a depth of 0 (ply 0). It's not the best move, but it's a sufficient analysis to propose realistic game situations. The final evaluation, which will serve as a rating for the user, is carried out at a depth of 3 (ply 3). The ply is the unit of measurement used to define the depth of the analysis. A 0-ply corresponds to an immediate, static evaluation of the position after the move, without considering dice rolls or future responses. A 2-ply analyzes your move and the opponent's best response move. A 3-ply analyzes your move, the opponent's response, AND your next move. You will never see a 1-ply because it is an approximation of a 2-ply, so GNU Backgammon intentionally ignores it.
At the end of the simulation, the game freezes in a realistic game situation, and the dice display their respective values for the next move. Now it is your turn to suggest what you think is the perfect move. If only one move is possible, a new game is started to avoid playing obvious situations.
Use the virtual keyboard to enter your answer. It should be in X/Y format, a combination that represents the movement of a checker. There will be as many combinations as there are moves. The moves can be condensed (6/1) or not (6/5, 5/1). These combinations will be separated by a space. X and Y will take the value 'Bar', 'Off', or a number between 1 and 24. You can also use your physical keyboard. Type 'b' for 'BAR', 'o' for 'OFF', the numbers, the / key, 'c' to switch between simulation types, 'g' to display Game Info, 'h' to display Hints and the Enter key to play a simulation or submit a move.
Then, submit your answer, and MBT via GNU Backgammon will tell you if it is the best move. If not, it will give you the equity of your move and the best move you should have found. Keep in mind, GNU Backgammon prioritizes purely mathematical precision, which may diverge from human intuition based on variance or psychology.
After, you can select 'Game Info' to see the list of moves played for the given simulation, along with the resulting pip. After submitting your move, you will also have access to 'Hints'. Here you will find the best moves as defined by GNU Backgammon. For each move, you will see in parentheses the level of ply used for the analysis, followed by the equity loss relative to the best move. You can consider you played an excellent move if you are better than 0.02 equity loss, you make a small mistake if you are between 0.02 and 0.08 and you make a big mistake (a blunder) if you lose more than 0.08 equity.
Finally, in the statistics tab, you will find your statistics for each game type, as well as your level based on your average equity loss across the three game types. Your level is expressed in points per move and is converted into an Error Rate (ER):
- Robot : 0 - 1
- World-Class : 1 - 4
- Expert : 4 - 7
- Advanced + : 7 - 11
- Advanced : 11 - 16
- Intermediate : 16 - 22
- Beginner : 22+
Note that because MBT evaluates only a single checker-play decision at a time, without cube actions and without full-game context, this ER is not equal to a full-match PR. However, the values are correlated, and you may consider that multiplying your MBT ER by approximately 1.35 will provide a closer estimate of your real PR.
MBT does not use any cookies or databases and collects no data. Your statistics and preferences are stored locally in your browser (via localStorage) and are erased in private browsing mode or when clearing browser data.
MBT is offered without any guarantees of any kind. Despite all the care taken during development and testing phases, it is still possible that you may encounter some issues. If this is the case, please be indulgent and feel free to report them to me :)
● GNU Backgammon 1.05.000 May 31 2021
Copyright (C) 1999, 2000, 2001, 2002, 2003, 2004 by Gary Wong
Copyright (C) 2015 by Gary Wong and the AUTHORS
● Porting GNU for the web by Theodore Hwa
● 'LE BACKGAMMON Stratégies et tactiques' by Henri Borentain / ISBN 2-85182-398-1
● Minimalist Backgammon Trainer developed by Alwin M. with the help of Gemini
● www.backgammon-trainer.ovh / backgammontrainer634@gmail.com
Version 1.4.2 - November 2025