Here's a position which confounded George Sulimirski, doyen of the London backgammon scene. The cube is in the centre and Black has a choice of two plays with 4,1: an attacking game plan with 11/7, 8/7, trying to contain White's last man; or a defensive position with 24/20, 21/20, making his opponent's 5-point. This is a difficult decision between two very different plans. The thinking might be like this:

For the attacking plan: my best idea is to restrain White's last man by making a prime. I can make a good improvement to my prime by making my bar-point with this 4,1 and then if I roll a 5 next turn I can make it a five point prime. I'm not too worried by White's potential for attack as all his men are in stacks and not very threatening.

For the defensive plan: it's never wrong to make your opponent's 5-point. This will stop him using those stacked men to improve his position if he rolls small numbers or a double. Meanwhile, even if White runs out with his last man I will still have a double shot at him; in any case, there will chances to make my bar-point later.

Most players opt for attack rather than defence. In this instance, however, the defensive play is correct because having set up a good defence the attacking possibilities on the other side of the board remain. The converse is not true as the game showed: after long thought, George opted for the attacking play. White rolled 4,4 playing 8/4(2)*, 6/2(2). George failed to enter. White rolled 4,2 playing 24/22, 8/4 and George stayed on the bar again. At this point White doubled and George dropped, thus never getting the chance to complete his attacking prime. The double is clear as is the drop - or is it? We'll see next week.