The first roll in backgammon is the only one whose correct play can simply be memorised. After that there is nothing to memorise: once the first move is made, positions become unique, and from there it is the plan that works, not the memory.
There are fifteen rolls in all, and learning them by heart is both feasible and worth doing. What follows is what is played on each.
One caveat belongs at the front, because it matters more than any single piece of advice: opening theory gives a head start, not a script. Every game is its own, and the position still has to be judged. One more honest note: there is no clear boundary between the opening and what follows.
The four rolls that make a point
If a single line survives from this page, let it be this one: the opening is played by putting the five-point above everything else. Making the five-point on the first roll is an immediate positional advantage, and the ranking of the rolls follows from it. Four rolls take a point on the very first move, and they are the strongest rolls of the opening; all four are played the same way — take the point. A fifth roll makes one too, 6-4, but the two-point is a case of its own — see that roll below.
3-1
3-1 is the best roll there is. It makes the five-point: a checker from the eight and a checker from the six, and you hold three points at once — the six, the five and the eight. It is the single best opening play in backgammon, without hedging.
The starting position and the best opening roll: the arrows show the play, the mark shows the point it is made for. Both dice lead to the same point, and it is held from the very first move.
4-2
4-2 is second, and its point is its own: the four-point, playing 8→4 and 6→4. The point is in the home board, and the same deuce clears a checker off the overstacked six — which is why 4-2 sits directly behind 3-1. Nothing else on this roll comes close.
The same device as 3-1, one point lower: the eight and the six give up a checker each, and the four is held. The six is left one checker lighter.
6-1
6-1 is third: the bar point, playing 13→7 and 8→7. It comes third, behind 3-1 and 4-2.
A six and a one make the bar point: one checker comes in from the outer board, the other from the player’s own row. The point is held, and the wall has begun.
5-3
5-3 is fourth. It makes the three-point, and it gives ground to the first three: the three-point is worth less than the five, the four and the bar point. What it offers instead is a plain play with no risk at all, and a home board that grows from the bottom. The second-best play, 13→8 and 24→21, is a clear step down.
Both checkers come from where 3-1 takes them from — the eight and the six — only they land lower.
Rolls that bring builders
These rolls make no point, but they bring down the checkers that will make one next turn.
5-2
5-2 is played two down, 13→11 and 13→8: an old play and still a strong one. The deuce could instead step a back checker to the twenty-two point, and that is the bad one: on double fives the opponent closes two points in his home board and both your back checkers land on the bar.
Two builders off the midpoint. No point is made, but next turn a great many rolls make one.
5-4
5-4 is the most underrated roll of the opening. The best play is the same shape as on 5-2: 13→9 and 13→8, both builders off the midpoint.
The same shape as 5-2, but the builders land closer to home: the nine and the eight.
Three plays are reasonable; the second is to split the back checkers and bring one checker down.
A back checker to the twenty-point — one step from an anchor — while the five brings a checker to the eight.
The running rolls
6-5
6-5 is the running roll and the simplest of them all: it escapes a back checker, and there is nothing to weigh. The play has a name of its own — the lover’s leap: 24→13, which carries a back checker half the way home safely. The rollout leaves no choice: the leap is the only way to play an opening 6-5. A lover’s leap can be answered with your own, though the more defensive options are often better.
This play has a name, and the board makes the reason plain:
The lover’s leap: a single back checker crosses half the board in one move and lands safely alongside its own.
6-3
6-3 is the awkward one, and running is not its only answer. The split is held to be stronger — 24→18 and 13→10: the back checker lands on the opponent’s bar point, the second brings a builder, and the stronger player is advised to take it.
One checker sits on the opponent’s bar point, the other becomes a builder. Both stand alone — that is what the position costs.
Running with 24→15 is rarely seen and perhaps should be: it gives up almost nothing in strength, but it is the harder and more dangerous play — the running checker is hit by thirteen rolls out of thirty-six.
The checker crosses nine points and stops under a direct shot from several numbers at once.
Rolls that split the back checkers
6-2
6-2 is played 24→18 and 13→11: the back checkers split while a builder appears in the outer board. Moving a back checker to the eighteen-point — the opponent’s bar point — is considered useful in itself.
Neither checker lands on a point, and both stand alone. The play is not about safety but about where they stand: the back one on a point the opponent wanted for himself, the front one where building is easy.
Why the eighteen-point in particular: 6-2 does not reach the middle of the board — only 6-5 escapes in one throw. With the escape out of reach, the best use of the six is to come out to the opponent’s bar point and wait there. The play even has a name: out and down. There is next to no argument about it: this is how 6-2 is played by default, and the old play, 13→5, is not played at all now.
The second way is to run to the sixteen-point.
One checker goes the whole way forward but lands in open country — three points short of its own.
4-1
4-1 is most often played 24→23 and 13→9, the most popular way.
The ace steps a back checker up one, the four brings a builder. The back checker leaves the direct shot and prepares to make a point.
The other way is to slot the five-point, and it is worse here: after 4-1 the checker on the nine-point already provides plenty of point-making chances, so the extra risk is not needed. And separately: slotting on 4-1 is more dangerous than on 2-1, because there are more ways to hit your blot on the nine-point.
The four brings a builder, the ace puts a checker on the five-point. One checker is exposed, but the point may close as soon as next turn.
How much worse is unsettled: the gap between slot and split is small, and what settles it is style rather than calculation. The slot goes to players who are stronger than their opponent and profit from a tangled game, the split to those who want a simpler position.
Slotting itself is a legitimate device — an accepted risk in many strong openings.
2-1
2-1 is the most argued of the five opening rolls, and it has two sound answers.
The slot: play the deuce down from the midpoint and put a checker on the five-point with the ace — the point is not held, but a great many rolls close it next turn, and the whole game goes to the attack.
The checker on the five-point stands alone and can be hit — but it is already there, and many rolls close the point next turn.
The split: a back checker steps up one while a second comes down from the midpoint — an improvement on both halves of the board at once.
There is still a blot, but it stands in the opponent’s home, where hitting it costs more.
The slot is the harder one to play against — which is why it is the one to choose. The cost is known in advance: if the blot is hit, it goes twenty points back. What you buy with it is that the opponent’s home board is still empty on the first roll, so you almost always enter straight from the bar — and the ace that slots also clears a checker off the overstacked six. Of the three rolls where slotting the five-point is reasonable at all — 2-1, 4-1 and 5-1 — this is the one where it costs least.
Where there is no agreement
A few rolls deserve their own section: they are played in different ways, and the best play is not settled.
First, about the format, or half of what follows hangs in the air. A single game — one win and it is over — is the simple case: there is no score in it at all, and the correct play is the same every time. In a match played to several points it depends on the score, because the value of a gammon changes and with it how much risk is worth taking. Leading, players avoid openings that risk giving up a gammon and keep things calm; trailing, the other way round. So where the play depends on the format it is said outright, and where it is not said, the format does not change it.
6-4
6-4. Plainly: nobody knows for sure which play is best here. The roll has three plans, and they stand about level — comparing them is hard precisely because each leads to its own kind of game.
24→18 and 13→9 is the most complicated of the three. Take it when you are clearly the stronger player: the more tangled the position, the more often the mistakes will be his.
Two lone checkers at opposite ends of the board: either can be hit, but you also have an answer to almost anything.
The two-point, 8→2 and 6→2, was long written off as weak — the point sits too far advanced to block much. Modern rollouts overturned that. But if you make it, steer away from a priming game and toward a blitz, a holding game or running, or the advantage has nowhere to go.
The point is held and nothing is exposed — but it sits too deep in the home board to stop the opponent’s checkers leaving.
Running with 24→14 gives simpler positions and shorter games. Of the running rolls 6-4 is the best one to run with: it is hit less often, and when it is not, you are further ahead in the race. It still falls short of 6-5 — one pip shorter, and the escape is not clean.
One checker crosses ten points and stops short of its own: a single pip is all this roll lacks to reach the middle of the board.
If you are not clearly the stronger player, the sound advice is to play whichever move you are most comfortable with.
3-2
3-2. Four plays are serious contenders, each with a name: down, split, reverse split, and up. The argument is mostly between the first two.
The split — 24→21 and 13→11 — is the most popular and, in a single game, ahead by a hair.
The three steps a back checker out of the worst of it, the deuce brings a builder. Two lone checkers again, and both stand where they are useful.
Down — 13→11 and 13→10 — is held to be its equal.
Both checkers come off the midpoint and land side by side. The back checkers stay put, but the outer board gains two builders at once.
The match score decides: well behind, play down — that game runs hotter, and the trailer is the one who needs heat; leading, split. The reverse split trails both by very little.
4-3
4-3. This roll is played two different ways, and the fashion between them has turned. For a long time two down — 13→10, 13→9 — was held to be the one poor choice of the four; now it is the one most often played.
The very play once held to be the worst of the four, and now the one played most often.
The choice is not about taste but about which game to play. Two down backs a point in the outer board or the bar point: both checkers work on the blockade. Splitting backs an anchor in the opponent’s home — the back checker goes to the twenty-one or twenty-point, the other comes to the nine, and from there you play a holding game, which is the easier one to play well.
Splitting with an anchor in mind: a back checker to the twenty-one point, the other to the nine.
There is a third way, to step both back checkers up at once; it is held to be sound when you lead the match.
5-1
5-1, in a single game, has two main plays: 24→23 with 13→8, or 13→8 with 6→5. Modern thinking gives the split the edge.
The five brings a checker to the eight, the ace touches a back checker.
The slot is slightly worse and risks a good deal more — the ace leaves a blot under a direct shot, and spending an ace that way is called a misuse of it.
The same five brings a checker to the eight, but the ace puts one on the five-point — and leaves it under a direct shot.
The slot is still chosen, and deliberately: it leads to complications in which a player expects to err less often than his opponent. Running to the eighteen-point with this roll, on the other hand, is not considered a contender in modern play at all.
Answering the opponent’s opening
A subject of its own, sketched only in outline: replying to an opening roll means following certain basic principles, and the specifics depend on what was played. One principle is almost always correct this early: hit the opponent’s blot if the roll allows it. The second is to answer what was actually played rather than a general plan: against a lover’s leap, for instance, the usual reply is to split the back checkers, keeping contact and raising the chances of hitting next turn. Learning every reply by heart, the way the fifteen opening rolls can be learnt, is out of reach — there are far too many positions, and what works here is principle, not memory. One exception is exact and is not about checkers at all — in a post-Crawford game the trailing player must double immediately, on the first move.
The match score changes the choice of opening too: when leading, avoid opening moves with a higher risk of giving up a gammon and favour calmer alternatives.
The overall picture of game plans is on the strategy page, and how the game ends on the bear-off page.
Comments