Backgammon

When deciding whether to take a double, the basic rule is that you must expect to be able to win 25 per cent of the games. That figure is valid only in cases where you do not expect to lose many gammons, but let's not complicate things for the moment. Because the next roll after a double is usually crucial, a good way to evaluate a position is to look at a cross-section of 36 games (the number of possible dice rolls) and estimate whether you can win nine of them.

Take this position from the recent Biba tournament at Brighton. Double Fives aficionado Julian Fetterlein was playing White against a weaker player when he ventured a double. His opponent should have reasoned as follows:

"How many games out of 36 can I win? Julian has a direct shot at a blot which, if he hits it, would guarantee winning the game and maybe a few, but not very many, gammons. So he has 17 good numbers (all 6's, 22, 33, 42, 24, 51, 15). But what of the race? Before the roll I lead 91-100. If Julian doesn't hit the shot the race is likely to be about even with me on roll. On nine of his numbers (11, 12, 21, 13, 31, 14, 41, 23, 32) I will get a shot at his blot, and he still has a man trapped in my board. "Of the 19 numbers where Julian doesn't hit my blot I am a substantial favourite, especially as I own the cube. I estimate that I could win 14 or 15 of those 19 games. I only need to win nine games to be able to accept the double, therefore I take."

Sadly for him, his reasoning went awry and he dropped, the fear of the direct shot dominating his thinking. Full marks to Julian for an aggressive double which paid dividends. It is only by evaluating the position methodically that you can reach the correct decision. Jellyfish, incidentally, evaluates the position as No double/ take, preferring to double out its opponent if it hits the shot.