Doubling Cube Strategy

How the cube works is on the rules page. This is when to use it.

Most players lose more points to bad cube decisions than to bad checker plays. The cube is also the fastest part of the game to improve, because a handful of numbers cover most positions.

The one number to remember: 25%

You should take a double if you will win the game more than about 25% of the time. That looks generous until you do the arithmetic:

Drop, and you lose exactly 1 point, every time.
Take, and you play for 2. Winning a quarter of the time gives 0.25 × (+2) + 0.75 × (−2) = −1 point, the same as dropping.
Above 25%, taking is the cheaper option. Below it, drop and move on.

Two things move that number. Gammon risk pushes your take point higher, because some losses will cost 4 points instead of 2. Owning the cube after you take pushes it lower — often to around 20% — because you can double back and your opponent then has to decide.

When to double

Double when both of these are true:

You are clearly better — roughly 60% or more to win.
Your opponent still has a take. If they must drop, you collect 1 point and the game ends.

The trap is waiting. Backgammon positions swing violently, and a good roll can take you from “a little better” straight past the point where anyone would take.

Market losers

A market loser is a roll that would make your position too strong to get a take next turn. Count them: if a fair share of your rolls would leave your opponent with an easy drop, double now while they still have a reason to take. Losing your market is the most common and most expensive cube error there is.

Too good to double

Sometimes you are so far ahead that doubling is wrong. If you are likely to win a gammon — your opponent has checkers trapped behind a prime and has borne nothing off — doubling just lets them drop for 1 point. Play on, close them out, and collect 2 or 3.

The rule of thumb: play on for the gammon when you are winning and your opponent's position is dead; double when you are winning and their position is merely worse.

Beavers, redoubles, and cube ownership

When you take, you own the cube: only you may offer the next double. That ownership is worth real equity, and it is why taking is often right even in positions that feel uncomfortable. Doubling from a position where the opponent is about to own a live cube is worth less than it looks.

Match play changes the maths

In a match to a target score, points you cannot use are worth nothing. Two consequences follow:

Leading near the target: double aggressively when a win ends the match; the extra points beyond the target cost you nothing.
The Crawford game: the game after someone reaches one point short is played without a cube. In the game after that, the trailer should usually double at their first chance — there is nothing to save.

Where you will meet the cube

Tables in Gammon Rivals are single games, so the cube does not appear there. It belongs to match play, which the game engine supports and which multi-game formats use. The thinking above is worth learning anyway: take points and market losers are how strong players judge a position, cube or no cube.

Next: pip counting, the fastest way to tell whether you are the favourite at all.