Openings, primes, blitzes, and the doubling cube — the ideas that turn a casual player into a winning one. Every position on this page is a link: click it and the board opens with that exact setup, ready to play against GnuBG.
Free, no signup, no ads. Beginner to expert bots, with hints and pip count.
Five ideas, each worth more than any amount of tactical cleverness. Get these right and you'll beat most casual players.
Computer rollouts have worked out the best play for every opening roll, and it comes down to three rules. If your roll can make a new point, make it. With 6-5, run a back checker to safety. Otherwise, bring a checker down from your midpoint.
Backgammon Opening Moves: the best play for all 15 rolls
3-1: Make your 5-pointPlay 8/5 6/5
The best opening roll. Your 5-point is the most valuable point on the board: it gives you a second point in your home board, blocks your opponent's back checkers and leaves nothing to hit.
Play it against GnuBG →
4-2: Make your 4-pointPlay 8/4 6/4
Nearly as good as 3-1. A second home-board point makes it harder for your opponent's back checkers to escape, and every checker you move lands safely.
Play it against GnuBG →
6-1: Make your bar pointPlay 13/7 8/7
Your 6, 7 and 8-points now form a three-point wall right in front of your opponent's back checkers, so they can no longer escape with a 6.
Play it against GnuBG →
5-3: Make your 3-pointPlay 8/3 6/3
Any new home-board point is worth making. The 3-point is deeper than ideal, but it is safe, and it leaves a checker you hit later fewer places to come back in.
Play it against GnuBG →
6-5: Run a back checker (lover's leap)Play 24/13
Run one back checker all the way to your midpoint, where it is safe. One of your two most exposed checkers escapes, and you move well ahead in the race.
Play it against GnuBG →Bring a checker or two down from your midpoint, often moving a back checker up with the other number. The top plays are close, so don't worry about picking the perfect one.
Four points are worth far more than the rest:
Your 5-point. The most valuable point on the board. Anchors your home board, blocks runners, and starts your prime. Make it whenever you can.
Your bar point (7). You already hold the 6 and 8, so the bar point completes a three-point wall. A back checker on your 1-point can no longer hop out with a single 6 — it needs a combination of 8 or more that clears all three points.
Your 4-point. With the 4, 5, and 6 held, an opponent coming off the bar needs a 1, 2, or 3 — half the entries gone.
The 20-point. Your opponent's 5-point, the "golden anchor." A safe landing spot deep in their territory and the place you rebuild from after being hit.
Play: Home board prime (4-5-6-7-8)The 20-point (your opponent's 5-point) is often called the "golden anchor". If you can't establish your own 5-point early, securing the golden anchor is the next best thing — it gives you a permanent safe haven deep in enemy territory.
Every safe-or-aggressive decision comes down to this table. Out of 36 dice combinations, here's how many reach a blot at each distance:
| Away | Chance of being hit | Which rolls |
|---|---|---|
| 1 | 31% | any 1 |
| 2 | 33% | any 2, 1-1 |
| 3 | 39% | any 3, 1-2, 1-1 |
| 4 | 42% | any 4, 1-3, 2-2, 1-1 |
| 5 | 42% | any 5, 1-4, 2-3 |
| 6 | 47% | any 6, 1-5, 2-4, 2-2, 3-3 |
| 7 | 17% | 1-6, 2-5, 3-4 |
| 8 | 17% | 2-6, 3-5, 4-4, 2-2 |
| 9 | 14% | 3-6, 4-5, 3-3 |
| 10 | 8% | 4-6, 5-5 |
| 11 | 6% | 5-6 |
| 12 | 8% | 6-6, 4-4, 3-3 |
Within direct range (1–6 points away), your blot is hit roughly 30–47% of the time. Beyond 6 points, the danger drops to under 17%. When you must leave a blot, leave it far away — or leave it where getting hit matters least.
Your pip count is how far all your checkers still have to travel. Both players start on 167. Every move subtracts the die you used; every hit adds them back on.
Ahead in the race? Break contact. Run your back checkers home and stop giving your opponent shots.
Behind in the race? Do the opposite. Hold your anchors, stay in contact, and wait for a shot — one hit can erase a 20-pip deficit.
Count by point: multiply the number of checkers on each point by the point number. The starting position: 2×24 + 5×13 + 3×8 + 5×6 = 48 + 65 + 24 + 30 = 167 pips. Practise counting mid-game — knowing the race lead shapes every decision.
Once all 15 checkers are home, position stops mattering and speed is everything:
Spread evenly. Fifteen checkers spread across all six home points bear off faster than the same checkers piled on the 4, 5 and 6, because fewer rolls go to waste.
Don't leave gaps underneath. Checkers stuck on the 5 and 6 with empty low points mean a 1 or a 2 does nothing but shuffle.
Don't build towers. Four or more on one point only come off together with doubles, and they shrink the number of rolls that help you.
Play: Ideal bear-off distributionAdvanced play is mostly pattern recognition. Every position is one of a handful of game types, and the skill is knowing which one you're in — and noticing the moment it turns into a different one.
A prime is a row of consecutive points you hold with two or more checkers. Four is strong, five is dominant, and a six-prime is unbeatable: nothing behind it can move until it breaks.
Building one. You start with the 6 and 8. Add the 5 and the bar point, then extend. The ideal wall runs 3-4-5-6-7-8.
Keeping it. Never break a six-prime while any legal alternative exists. Every turn it holds, your opponent loses a turn. Shuffle spares on the leading edge instead.
The catch. Priming games die of timing. Keep spare checkers in the outfield that you can bring forward without touching the wall.
Play: Six-prime trapping four checkersIn prime-versus-prime battles, timing trumps action. The player whose prime collapses first usually loses. Count how many non-destructive rolls each side has left — the player with more "timing rolls" has the advantage, even if they have fewer immediate threats.
A blitz is an all-out attack: hit everything you can while slamming your home board shut, so their checkers pile up on the bar with nowhere to enter.
When to blitz. Two or more enemy blots in your home quadrant, your 4- or 5-point already made, and builders waiting on the 8 and 13.
How to blitz. Hit everything in reach, even leaving blots of your own. Every home-board point you make removes another entry number, and the effect snowballs: with three points closed they still enter 75% of the time, with four it's 56%, with five just 31%. That outweighs the return fire.
The payoff. Close all six points with someone on the bar and their turns simply vanish. Close games become gammons this way.
Play: Five-point board blitz — 3 checkers on barBehind in the race, but not desperately? Hold an anchor — two checkers on one point in your opponent's territory — and wait for them to leave you a shot.
Where to anchor. The 20-point is best, the 21-point next. Anchors on the 22, 23, and 24 get progressively worse: harder to escape, fewer useful numbers.
While you wait. Bring everything else home safely and build your board, so the hit — when it finally comes — actually sticks.
When to let go. Only when you've pulled ahead in the race, you've hit and need to run, or your board is strong enough that hitting wins outright. Never for a small gain.
Play: Holding game with golden anchorThe back game is a last resort for when you're badly behind. You hold two anchors in your opponent's home board and build your own, betting that they'll have to leave a shot on the way in.
Which anchors. The classics are 1-3, 1-2, and 2-3. The 2-3 back game is strongest because it covers the most numbers as your opponent bears in.
Timing decides it. Back games lose by being forced to tear down your own board before the shot arrives. Hoard spares in the outfield. You want to hit late, with a closed board waiting.
The trap. Getting hit three times isn't a back game, it's just losing. A real back game means you're building a board worth hitting into. Without one, don't bother.
Play: 2-3 back game with strong home boardThe cube is the part of backgammon with no dice in it. Nothing else you learn will improve your results as much.
The 25% take point. Take any double where you win more than 25% of the time. Passing costs exactly 1 point. Taking a 25% game costs (0.25 × 2) − (0.75 × 2) = −1.0, the same. Above 25%, taking wins.
The doubling window. Double at roughly 70–75% winning chances. Earlier and you hand over the cube too cheaply; later and your opponent simply passes.
Woolsey's Law. If you can't tell whether your opponent has a take or a pass, double. Doubling a bit early costs you the cube; failing to double costs you the whole game's equity.
Gammons shift everything. If a fifth of your losses are gammons, your take point climbs from 25% to about 32%. When you're the one threatening gammons, double harder — their take is worse than it looks.
Owning the cube is worth something. Because you control the redouble, you can take a little thinner than the raw maths says — the practical take point is nearer 22% than 25%.
Two ideas that almost nobody teaches beginners, and that quietly decide most move choices.
Duplication. Arrange your blots so your opponent needs the same number for both threats. If a 4 hits one blot and a 4 hits the other, they only get one of them per turn (barring double 4s).
Diversification. The same trick pointed the other way. Set up so different numbers each do something for you: a 3 makes a point, a 5 hits, a 1 covers. Now most of the dice help.
In practice. Your opponent has a back checker on your 1-point and checkers on your 12-point (their midpoint). You have a blot on your 5-point — 4 pips from that back checker. If you must leave a second blot in the outfield, put it on your 16-point (4 pips from their midpoint) rather than your 17-point (5 pips). With both blots 4 away, 15 of 36 rolls hit something; at 4 and 5 away, 26 of 36 do.
Duplicate your opponent's good numbers, diversify your own. This single principle explains dozens of seemingly mysterious plays that computers and grandmasters make. When evaluating candidate moves, count how many "good" numbers each side has — the move that minimises your opponent's useful rolls while maximising yours is usually correct.
Timing is simply having harmless moves left. Good timing means spare checkers you can shuffle without touching your structure. Bad timing means every legal move damages something.
Wastage burns timing. Dumping a checker from the 6-point to the 1-point instead of making a point spends pips for nothing, and a stack of five has nowhere useful to go.
Recirculation. Getting hit can be a gift. When your prime is about to crack, a checker sent back buys you a dozen more turns before you're forced to break anything.
Stay in the zone. Points 7–12 are the reservoir. Checkers there can trickle in over several turns. Checkers already home are committed and burn timing fast.
In money play every game stands alone. In match play (first to N points) the score rewrites the value of everything.
The Crawford rule. When someone reaches one point from victory, the next game is played without the cube. Otherwise the trailer would double to 2 automatically, every time.
Free drop. In the game right after Crawford, the trailer should double at once. If the trailer needs an even number of points, the leader can drop and lose almost nothing — a spare point is worthless to the trailer — so the leader passes whenever the game has started badly. If the trailer needs an odd number, dropping is expensive and the leader should take.
Gammons change sides. Ahead in the match, an extra point often overshoots and barely helps, so play safe. Behind, a gammon is worth two, so take risks for it.
Match equity tables. Serious players memorise take points by score. At 3-away/3-away they're not far from money play. At 2-away/4-away they're very different.
Each of these has playable positions too: