Site home

The Bold Play

Here is a position that came up during the October 29th 2007 club evening at Lord's.

It's a money game, and Green (Anthony Miller) has to play 31. The 3 must be played from the bar, but how should the 1 be played? Have a little think about it before reading on.

Playing 3/2 seems natural, preparing for the hit on the ace point if White (Tony Lezard) fails to escape. However, if White does escape, this is bad news for Green, who is trailing 106-64 in the race. Are there any alternatives?

Hitting with 2/1* is a bold play and puts White on the back foot. If White fails to come back on, Green can win instantly with the cube. (In fact it's almost - though not quite - too good to double.)

Positions like this illustrate two important points. First, the number one cause of mistakes in backgammon is instantly seizing on a good-looking move without considering the alternatives, and thereby missing the better play.

The second point to remember is that sometimes, the obvious play is the best! In this case, the conventional 3/2 gives an equity of 0.811, and the equity of 2/1* is only 0.506, the difference of -0.305 representing a serious blunder. It's true that 2/1* does win almost twice as many gammons (18% versus 10%) but it also concedes far more: 17% versus 5%. The correct play in this case is - overwhelmingly - the conventional one. (The third option of 13/12 comes in slightly behind 3/2.)

Why is the hit such a bad idea? The key problem is of course all those rolls that do hit - rolls containing a 1 or a 2, of which there are 20 out of a possible 36 - more than half. Moreover, White's board is at least passably good, with four points held and a good chance of saving the blot on the 6 point with the other half of his roll. Green has two other blots on the board as well, and if White picks these up, then gammons are a strong possibility.

This is not to say that it was daft to consider hitting. Sometimes, unconventional plays like this are called for. For example, consider the same position with White having taken some pieces off the board already:

Here, it's more important to put the brakes on White, and 2/1* is correct. Interestingly, the third alternative, 13/12, is now the worst of both worlds and a significant error.

Tony Lezard
30th October 2007