How to get better at backgammon: measure your PR, then lower it

PR stands for Performance Rating. It measures how well you played, decision by decision, against a strong engine. After a match, every move you made and every cube decision you took is compared with the best one. Each time your choice was worse, you gave away some equity, the value of the position in points per game. Your PR is the average of those losses, scaled to a readable number.

Lower is better. A PR of 0 means every decision matched the engine's. The higher the number, the more equity you gave away per decision.

What matters most is what the PR leaves out: the dice. You can play well and lose a match to lucky rolls, or play badly and win. Results mix skill and luck; the PR only looks at your decisions. That's why it's the most useful number to follow if you want to know whether you're getting better.

On Megammon, the formula is: PR = 500 × total equity lost ÷ number of counted decisions. Each part has a precise meaning:

• Equity lost on a decision is the gap between the equity of the best choice and the equity of the choice you made. It's never negative: a decision that matches the best one costs 0.
• Counted decisions are the non-trivial ones. A forced move, or a roll where every legal play is worth the same, says nothing about your skill, so it's left out. On Megammon, a checker play counts when the best and the worst legal moves differ by at least 0.001.
• Cube decisions count too: doubling or not, and taking or passing, count like any other decision, except obvious cube decisions you get right, such as not doubling when you're nowhere near a double.
• Losses are normalized to a cube of 1, so a mistake doesn't count eight times more just because the cube is on 8. In a match, they're expressed in money-game units, as normalized equity, so that mistakes at different scores can be compared.

A worked example

Equity is measured in points per game: +1.000 is a sure single win, −1.000 a sure single loss. Suppose your counted decisions in a match cost you 0.010 each on average, a hundredth of a point: your PR is 500 × 0.010 = 5. At an average of 0.020 per decision, it's 10.

An average hides the shape of your mistakes. A single blunder costing 0.200 adds 500 × 0.200 = 100 to the total before it's divided by the number of decisions: in a match with 100 counted decisions, that one mistake alone is worth a full point of PR. That's why cutting out big mistakes is the fastest way to bring your PR down.

There's no official table: each analysis tool uses its own wording for the levels. Megammon places every PR on its own scale, shown next to your PR in your matches and on your profile:

Read it as a guide, not a verdict. Your PR also depends on the positions you got: a long game full of contact offers many more ways to go wrong than a straight race. Compare your PR with your own earlier ones first, over many matches.

Your PR doesn't depend on whether your rolls were good or bad, but it still varies a lot from one match to the next. Two reasons:

• The dice choose your positions. Some matches hand you easy decisions, others a string of hard ones. Hard positions produce bigger mistakes, even for strong players.
• Short matches have few decisions. With fewer counted decisions, a single blunder weighs more in the average, as in the example above.

So judge your level over many matches. On Megammon, your profile computes your PR over all the decisions of your analyzed matches, not as an average of match PRs, and its chart shows each match's PR by date, so you can see the trend.

A PR is a total. To lower it, break it down. Three questions locate most of the equity you lose.

1. Checker play or cube?

Checker decisions come up every turn, cube decisions much less often. Cube errors tend to be fewer but bigger, so they can be quick points to win back. Compare your two PRs after every match.

2. Small slips or big mistakes?

Sort your mistakes by size. Megammon labels each decision by the equity it lost:

By definition, a blunder costs at least 0.080: twice as much as the smallest error, four times as much as the smallest doubtful play. Start with your blunders: they often come from the same idea going wrong again and again.

3. Which part of the game?

Openings, middle games full of contact, holding games, races and the bear-off each have their own kinds of mistakes. Your weakness profile on Megammon looks across all your analyzed matches and ranks your top three weaknesses by the equity they cost you, not by how often they occur, broken down by game phase and by cube decision.

Getting better at backgammon comes down to a loop: play, find your mistakes, understand them, and practice them until the right play comes naturally. Here's a plan you can follow at any level.

1. Play matches, not just single games

Matches are how backgammon is played in clubs and tournaments, and the score changes the right decisions, especially with the cube (see match play). On Megammon, ranked matches are played to 1, 3, 5 or 7 points, with the cube from 3 points up.

2. Review every match, biggest mistakes first

Right after a match, while you still remember what you were thinking, go through the analysis. Look at your blunders first, then your errors. For each one, make sure you understand why the better play is better before moving on.

3. Name the cause

A mistake you can name is a mistake you can stop making. “I didn't count the pips”, “I broke my anchor too soon”, “I passed a take”: a handful of causes often explains much of a player's lost equity. Megammon's analysis names the main cause of each mistake, from “Back checker to escape” to “Wrong take”.

4. Drill your own positions

Reading about a mistake isn't the same as fixing it. Replay the positions you got wrong until you find the right play without hesitating. On Megammon, your errors and blunders become training exercises, and spaced repetition brings each one back right before you'd forget it.

5. Study one topic at a time

When your reviews keep pointing to the same area, read up on it, then go back to playing. Our guides cover the main ones: the opening moves, when to hit, anchors, the five game plans, the pip count, bearing off and the doubling cube.

Checker-play mistakes tend to come back in the same situations. These are the ones worth working on first:

• The opening. The first move comes up in every game, and there are only 15 possible opening rolls. Learn the best opening moves: it's equity you win back in every game.
• Hit or not. Decisions to hit, hit loose or play safe come up in almost every game. Our guide to when to hit covers the reasons on both sides.
• Choosing which blot to leave. When a blot can't be avoided, count the shots each candidate faces and what a hit would cost you, as in leaving the safest blot.
• The game plan. Many big mistakes aren't one bad move but the wrong plan: racing when you're behind, attacking without builders, holding on without timing. Check the pip count and choose a game plan before you look for the move.
• The bear-off against contact. A few careless moves at the end can lose a game that was won. See bearing off against contact.

Cube decisions are fewer, but each one can cost a lot. Four habits help:

• Think about the cube every turn. Before you roll, ask whether it's time to double. A double you never considered can't be a good one.
• Know the take point. In money play, leaving gammons and the value of owning the cube aside, a double is worth taking if you win at least one game in four. Beginners tend to pass too often: see the take point.
• Watch the gammons. Gammon chances weigh on every cube decision: they can turn a take into a pass, or make a position too good to double. See too good to double.
• Use the race formulas. In a pure race, the pip count turns cube decisions into arithmetic: see cube decisions in a race.

In a match, the score changes the right cube decision too. The Megammon Engine judges every cube decision by your match winning chances at the actual score, so your cube PR already accounts for the match.

Most mistakes aren't a lack of knowledge but a lapse of attention. These habits catch many of them:

• Count the pips at key moments: before a double, before choosing between running and holding, and when the race begins.
• Look at your opponent's next roll. Before you play, ask which numbers are good for them and how your move changes that.
• Spend your time on the hard decisions. Online on Megammon, every action starts with a 12-second delay before your time bank runs: routine moves fit in it, so keep your bank for the close calls.
• Judge decisions, not results. A good play that loses to a lucky roll was still a good play. Let the analysis, not the outcome, tell you whether you were right.
• Take a break after a bad beat. Frustration and fatigue lead to quick, careless moves. Stop, then come back.

• Judging your level by one match. A single PR swings a lot. Look at the trend over many matches.
• Playing without reviewing. Playing a lot builds habits, good and bad. Without reviews, the bad ones stay.
• Blaming the dice. Luck is real, but it's measured separately. If your PR was high, the dice weren't the only reason you lost.
• Memorizing without understanding. A position learned by heart helps once; the idea behind it helps in every similar position.
• Studying rare positions first. Fix what comes up in every game, such as the opening, hitting decisions and the cube, before exotic endings.

Every match you play on Megammon can be analyzed: online, against a friend or against the bot. You can play the bot with no sign-up, and with a free account the analysis is ready seconds after the match ends.

New to the game? Start with the rules of backgammon and the common backgammon mistakes, then come back to this plan.