by Phil Simborg
26 May 2009

I am not a bot. I am not Snowie, nor GNUBG, nor any other backgammon computer program. While I envy their skills, I am not able to play as well as the bots. Fortunately, most of my opponents are not as good as the bots either.
The problem is, however, that when I attempt to learn the best checker and cube decisions, I pretty much have to rely on the bots. Even when I am fortunate enough to get the advice of one of the better players in the world, unless the answer is obvious, they, too, pay homage to the bot and insist on seeing a full rollout before providing assurance of an answer.
| The rest of this article (17.31 K) is premium content. |
One retort to Positon #3 about Ray saying he may be able to use the cube later to punish you, when? How? Your take point is 11% at that score as you rewhip for the match and assuming any ELO difference your TP may be even higher. Ray would have to get to one hell of a position to ever use it 'as a weapon against you'. More likely is he'd hold off on cubing when it's possible he could cube/cash because he'd rather not put the match on the line.
Good point Jacob (Stick). It's true that Ray has little recube vig, but a major point is that once I give the cube, I can never use it to end the game at a point where I could and should cash, and even with an 11% takepoint, as it happens, we certainly did get below that before the game ended!
Hi Phil... I don't really remember emphasizing my recube vig so much... I probably wouldn't redouble until you were dead, as Stick correctly points out.
You correctly figured that it was an easy take OTB.
I have found that a more effective match strategy against LEADERS in matches is to give them closer decisions and lead them to PASS a TAKE in trying to protect a lead. You want to win gammons of course... But this position is hardly more gammonish than say, the starting position. One ought to be optimistic as the trailer. As the trailer you expect the LEADER to be more conservative and maybe you can take advantage of that. What player is so conservative that he will actually PASS this cube that will take him to the Crawford game at least if he wins??? I don't know, but it isn't a "taker" like me!
I was glad to get the cube here as you have explained in this article and here is why:
It is only "correct" to double by .014. In this type of position with so much complexity I can state unequivocably that even the best player in the world, (insert your name here), will not play the position well enough to justify doubling. 0.014 is just too thin a margin for error.
Therefore, NOBODY should double this unless they KNOW they are playing a person who is capable of PASSING. I don't know if any open player is going to pass this, and if he does, he should go back to the intermediate division!
A great player shouldn't, and wouldn't, double this as the trailer... but it would be a better idea for him than for the weaker player to do it. If an expert makes 5 errors to 10 for an average player, his equity would amount to more than 0.014 from this particular position.
If you are playing someone better than you, you may expect him to play better... The longer the game goes, (and this will be a long one), his advantage will multiply GEOMETRICALLY. You may get lucky sometime and win a gammon, but then you will have a slight lead in a match against a better opponent which is nothing more than a 50/50 proposition. The expert is glad to risk that in exchange for playing this complex position out until the bitter end, all the while enjoying ownership of the cube!
To make a long story short:
Nobody is good enough to get the equity the bot assigns to this position out of it. A bot should double, but a human should NOT.
Good article, Phil. I just can't figure out which 15% of the open players would take the first position's cube.... 5-10% maybe.... and I still think 10% is high.
3 point board, 2 checkers on your 1, still behind a 4 point prime.... ghastly...
You must be signed in to post comments.

