You are too lazy to make endless calculations over the board? You don't want to memorize huge tables that you will need only a very few times? But what will you do in a position like this?
GNU Backgammon Position ID: YAAAAAMAAAAAAA Match ID : QYkEAAAAAAAA +13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg (Cube: 2) | | | O | OOO 0 points | | | O | OOO | | | | OOO | | | | OO | | | | OO v| |BAR| | | | | | XX | | | | XX | | | | XXX | | | X | XXX Rolled 11 | | | X | XXX 0 points +12-11-10--9--8--7-------6--5--4--3--2--1-+ X: wintom Pip counts: O 12, X 12
For these two-chequer positions you only have to memorize the following four rules: 1. Avoid to leave the two chequers on the same point. 2. If you can bear off one of the chequers, that's what you are supposed to do. 3. If you cannot bear off one chequer then put one chequer on the 5-point. 4. If you cannot put a chequer on the 5-point put one on the 4-point.
If you have a good look at one of those probability tables you will find out that in doubt a combination where there is a chequer on the 5-point is always best. There are only two exceptions: when you have to decide between 3-2 and 4-1 or when the decision is 6-1 or 4-3. In both cases the one with a chequer on the 4-point is better.
Applying this to the position noted above we now know that 6/5 6/3 is the best move.
robadams wrote: > I think one should memorize the 2 checker positions. Your rules > wouldn't help you play a 2,1 from that position correctly for instance.
As far as I know there are 10 combinations for 6,3 and 10 combinations for 5,4 so my rules would not be mistaken. To move 6/5 6/4 would be ok.
> with 9 pips 6,3 is better then 5,4
Though there is an equal number of combinations that bear off on the next roll there is a slight difference on the second roll: If you move to 6,3 even if you throw 21 there will be only one chequer left wich makes bearing off easier at that state (as stated in the second rule). That is the reason why 6,3 is supposed to be slightly better. A rollout with gnubg showed that the difference in equity is 0.003. Rather small. But thanks for pointing this out. I hadn't realized it before.
> That's about it. Not a huge chart.
Oh, for me that's big enough. You don't know how bad my brain works sometimes ;-) I guess I will stick to my rules loosing an equity of 0.003 in one of about 20 cases or I will just remember 6,3 as an exception :-/ But I understand, that this might disqualify my rules for common use :-(
|
