FIBS Board backgammon forum

OK, lets actually discuss the positions! Position 1.
The first thing to say about this position is that it is not yet a backgame. Granted Black has six men back and can certainly enter a backgame here if he wants to, but as he owns the 20pt and can make his own 5pt here, then surely this is still a positional struggle. Black is in quite good shape for that and should aim to recycle checkers through his high anchor so that he can make points in his own outfield. I don’t think that the massive backgame option is necessarily a non starter, in fact it may do as well as the simpler game plans, but I don’t see how you could know that over the table. 24/20, 8/5 for me, the play of snowie and gnu, although they both think that 24/20, 13/10 is close.
Here’s what Bill Robertie has to say about back games. “Well-timed back games can give you a chance to save a bad game, but that’s all. Don’t seek them out.” Again, “Remember that the back game is the last resort when all other plans have failed, so look for opportunities to break out of the backgame by hitting shots”. Who is Bill Robertie? The first two-time world champion and one of the truly great theorists and teachers. His books are of somewhat variable quality, but I strongly recommend, if you can find them “Reno ‘86”, “Dwek v. Genud” and “Modern Backgammon”. His Advanced Backgammon Vols 1 and 2 will also provide you with a terrific grounding in checker and cube play.
Some points to note. Losing a gammon here doesn’t cost much, but how much is not much? Losing your free drop costs you about 1.5% of match equity, i.e. you are 51.5% to win the match at 1-away, 2-away post Crawford. Losing a backgammon is of course a disaster and it will cramp your style more than somewhat if you get down to the hard end of a backgame.  You will have to run to save the bg in positions that are a big stay normally.
So here we have a choice of two game plans. Make the 5pt and try to play a nice solid positional game that should be fairly straightforward, or go for a big back game which will undoubtedly be hard to play. At this score it seems to me to be an easy choice.

In position 2 we can see what can happen down the back game road. This isn’t a bad roll, it’s just very tough to play but you may like to ask yourself how would you have played any other rolls. This is a rocky road from here to the end! 24/20, 13/11(Snowie’s play) looks reasonable and 24/22*, 13/9 also has its merits. Hitting the blot isn’t an attempt to go forward. The idea is to make the 22pt if possible which will work quite well with any of the other available anchors. 24 and 23 is a poor game at best and neither of those points works well with the 20pt.
Other plans are probably quite reasonable too, but avoid playing 5/3 or 8/6 with the 2. You must keep all your checkers in play and as flexible as possible. Thus 13/7, or 13/9, 13/11 or even 24/20, 24/22* are not too far from the right idea.

Now we move on to position 3 and look at a White play for a change. Oh dear, how has Black got a checker onto his ace point? Now he will have to play with 14 men and the dead checker can’t be part of a winning prime. In some variations it will get recycled, but the move that took it there just had to be wrong. Anyway, should White hit Bar/17* or try Bar/23, 21/15? The other two sixes aren’t worth considering, as the midpoint still has a function connecting with White’s back checkers and burying the spare from the bar point can’t be right. Spares mean choices and flexibility, very important. Snowie will hit here, but the rollout clearly shows that this is wrong. The quiet play wins 60%, the hit 55%, a huge difference. Although White will take a gammon if she can get it and will benefit hugely from a bg, her prime aim is just to win the game and this play is clearly best for that.
Now we move on to position 4. The over-riding concern here is to keep the 23 and 22pts. Black is just about committed to his back game and these are probably the best combination that he can get, only 24 and 22 come close as an effective force. With this in mind the threes have to come from the midpoint. The rollout comes down in favour of Bar/22, 13/10(3), but Bar/22, 13/10, 13/7 is close. Snowie will play Bar/22, 23/20, 13/10(2), trying to avoid committing to the true deep backgame, but that play doesn’t do well in the rollout. We hear a lot about how “Bots don’t know how to play backgames”, but the fact of the matter is that they are extremely tough to play, from both sides and that however badly the bots do or don’t play them, humans are worse!
Finally position 5. Black picked out bar/22, 13/10(3), nice play and once again White has to choose between hitting or not. Sometimes, it doesn’t matter much either way and this is one of those times. The rollout barely distinguishes among half a dozen different plays.
You could try 21/16, 21/17 and 17/12*, 13/9 tries to make another blocking point. Take your pick.
I hope that I have demonstrated that back games are to be avoided if possible. Their outcomes are uncertain, the play extremely difficult. In position one, the computer assigns almost equal equity to the straightforward play that makes the 5pt and the more complex 24/20, 13/10, inviting a backgame. However, you must bear in mind that as humans our chances of attaining the theoretical equity in any difficult position are next to none. I believe that only a true World Class player, of whom there are probably only a hundred or so in the world, will get near to the theoretical equity of position 1 when choosing a back game, whereas an ordinarily strong player should do quite well with 24/20, 8/5. However, any play that hits on the 4pt is clearly a blunder and should be avoided at all costs! I note that one correspondent suggests that gnu hits in this position, but I don’t know why, mine doesn’t!
Enjoy the game!

Quote from: dorbel on July 16, 2008, 08:30:34 PM
I hope that I have demonstrated that back games are to be avoided if possible. Their outcomes are uncertain, the play extremely difficult. In position one, the computer assigns almost equal equity to the straightforward play that makes the 5pt and the more complex 24/20, 13/10, inviting a backgame. However, you must bear in mind that as humans our chances of attaining the theoretical equity in any difficult position are next to none.

Thanks for the analyzes, dorpel.  I have to say that because backgames are so hard to play that means that the opponent, another human, will make errors too.

There is some mix up in the position 1 also. It is white's turn, white on bar, it is not possible to hit to 4 point or play 24/20 8/5??

I see. if I could read a diagram correctly I'd be a great analyst! With white on roll it's pretty clear to hit from the bar, mostly to try and stop the opponent from making his 5pt.
There is a tendency to think that hitting a man with  six checkers back must be wrong. it can be, but it isn't always.