Note: All ‘cash game’ problems assume the Jacoby Rule is in effect. That is, you can’t win a gammon unless the cube has been turned.

In this position, White launched a blitz after a few moves were made on both sides, and now Black is in serious trouble with two men up in the air facing a four-point board and more blots floating around. White’s 6-4 gives him several excellent options; his job now is not to get too excited, but to just find the most efficient way to bring the game home.

In this particular position White has three goals. Without trying to assess their relative importance, here they are:
(1) Close the 5-point. Closing the ace-point would also be nice, but White won’t have the ammunition in place to do that for some time. With a small double he’ll cheerfully switch points, but otherwise the ace-point will have to wait. (The most likely outcome of this position is in fact a well-timed ace-point game.)

(2) Pick up blots. Hitting more blots helps in several ways. The extra blots hit obviously lead to both more gammons and more backgammons. Hitting more blots also buys time to escape the back checkers before Black can get everybody in from the bar. Finally, hitting blots eliminates potential builders and slows down Black’s ability to put his position together once he enters.

(3) Escape the back checkers. The sooner White can extricate his back checkers, the sooner he avoids a potential disaster scenario: Black enters both checkers from the bar, and White is caught with blots in the outfield to pick up, as well as refugees in Black’s inner board that need to get out quickly.

Now let’s look at White’s possible plays. He has a bunch, some of which emphasize one goal, while others are more balanced.
Play A: 21/17* 24/18. The “Get-out-of-Dodge” play, this strongly favors escaping and hitting. White picks up one blot now, while generating a double-shot at the last outfield blot.

Play B: 13/7 13/9. This “Close-him-out-quick” play ignores the blots while creating three builders for the 5-point, as well as some numbers which might be used to hit on the ace-point if Black fans.

Play C: 24/14*. This play hits the blot that’s hardest to hit, while escaping the checker that’s hardest to escape.

Play D: 21/17*/11. Hit-and-Build #1 puts a third checker up while creating another builder for the 5-point.

Play E: 21/17* 13/7. Hit and Build #2 creates a different builder for the 5-point which also shoots at the ace-point. In addition, Black still has a direct shot at the blot on the 14-point.

Before we discuss the merits of the different approaches, we can prune our list a bit. It’s pretty clear that Play E, with a builder on the 7-point, dominates Play D, which puts a builder on the 11-point. The builder on the 7-point not only aims at the 5-point but also aims at the ace-point, while serving as a slot for the bar if Black starts to enter and White decides he needs a prime. So we can eliminate Play D.

Now let’s compare Play A and Play C. Both involve 24/18. Do we then prefer 18/14* or 21/17*? Since neither play puts a builder in position for the 5-point, 21/17* looks better, escaping a checker and getting some additional shots.

With these two eliminations, we’ve reduced our list of possibilities to three: Play A, the best for escaping the back men; Play B, the best for closing the 5-point; and Play E, the best for making progress on both fronts. Now what?

In the pre-bot era, the standard play here was Play B, 13/7 13/9. The idea was that closing the 5-point was crucial, after which White could pick up the blots at his leisure. While playing two checkers down didn’t actually give that many new numbers for closing the 5-point, it did make White a big favorite to hit loose if Black entered, after which Black would be an underdog to hit back, while White would then have multiple cover numbers if Black missed. In a high-stakes chouette in the 1980s or 1990s, this play wouldn’t even have generated much discussion – with two on the bar, building the board quickly was thought to be paramount.

One of the first lessons we learned from the early bots was that hitting was a big deal no matter how many men were on the bar. Every blot hit represented more gammons won, and if the gammon was in the bag then each blot hit started increasing the backgammon count. The problem with building plays is that blots are never as easy to hit as they appear to be. If you have a possible hit now but you pass on the hit to make a structural play, you may never get another chance. In this problem, for instance, bringing down more builders is always trivially easy, but hitting will never be.

If we eliminate the building play and focus on the two hits, then 21/17* 13/7 dominates 21/17* 24/18. While we don’t want to sacrifice hitting for building, we also can’t ignore building, because our life will get a lot easier if we prevent Black from making our 5-point. In addition, a blot on the 7-point is a fabulous multi-purpose blot: it’s simultaneously a builder for the 5-point, an attacker for the ace-point, and a slot in case Black makes our ace-point and we need the bar-point for control.

To see the power of hitting, take a look at the wins/gammons distribution after a rollout:

Overall Win %        G/BG Win %
21/17* 13/7             79.4 62.9
21/17* 24/18           79.1 61.8
21/17*/11                79.3 60.7
24/14*                     78.6 60.0
13/7 13/9                78.0 60.0

Although the numbers are relatively close, notice that the non-hitting play scores lowest in overall winning percentage and tied for last in gammon/backgammon percentage. Hitting is the road to winning more gammons, which is fairly obvious; but it’s also the road to winning more games, by keeping Black on the bar longer and allowing White to get his army home.