Cash game. Center cube. White on move.

White to play 2-2.

The advent of the bots in the late 1990s enabled players to solve a myriad of small technical plays that recurred frequently but couldn’t really be tackled with hand rollouts or pure reasoning. Hand rollouts were so slow that they really had to be reserved for positions where the solution was unknown but the difference between plays was likely to be large and important. Players ignored what appeared to be small technical stuff, on the theory that solving these problems, even if possible, was most likely a huge waste of energy.

This position and roll is a perfect example. If you’ve practiced against bots a lot, you’ve seen this idea a few times and the answer probably seems trivially obvious. If you haven’t, it most likely seems non-obvious but also not terribly important.

First, let’s reject the clearly erroneous lines. White is holding the best possible defensive anchor, one that will cause Black a lot of problems. So we’ll reject any play that includes 20/18(2) or 20/16(2).

Let’s also reject the hit 11/3*. The 3-point has some value but it’s not nearly as important for White as the 5-point or the 7-point. In addition, the hit leaves the midpoint stacked and undeveloped, a situation which we’d like to fix.

Eliminating those ideas makes most of our play fairly clear. We’re going to use three deuces to play 13/9 11/9, making an excellent point which bears upon all the points we’d like to make next, and takes a further step toward building a good blockade. The problem then becomes the play of the last deuce. Do we play 6/4 or 13/11?

Both plays seem to have some obvious merits. Splitting off the 6-point looks like it unstacks a bit and creates a new builder for the 3-point. In addition, the play doesn’t create a new blot. But 13/11 creates another builder for the 5-point at a very small risk. What’s better?

The right idea, in almost all positions of this type, is 13/11. The real point of the play is that the two checkers on the 6-point belong where they are, taking aim at the 5-point. The structure consisting of the 4-point, 5-point, and 6-point (known as “the rack”) is much stronger then the 3-point/4-point/6-point structure, so White should keep as many checkers as possible aimed at the 5-point. Playing 6/4 removes a builder for the 5-point and aims it instead at the less-important 3-point. Even though the cost is small, that’s a mistake.

A large part of playing good backgammon consists in having hundreds of these little tactical ideas at your fingertips, which is why practicing against bots and reviewing each session afterwards is such a vital part of your development as a player.