I admit that doubling decisions are a weak point for me, also because cube reference positions are not a major source of interest to me. I am satisfied when I am not more than 0.1 off on average, and hope that this equity loss will be made good by proper checker play and psychological assessment.
In order to make up for my (and other's) laziness in this domain, in this thread me or someone else is going to post a position, and the intention is to give the arguments for the cube decision. Hopefully we will get sight of rough guidelines also for complex positions. Any ideas or commentary are welcome.

Position ID: bdsBCwC3zQYYAA
Match ID: MAEAAAAAAAAE


Does black have no double, a reasonable double or a good double?

No double. Black behind in the race. If white brings his men to safety you're glad you didn't double, and if white has to leave a shot, you can then double if you want. You only regret not doubling if you throw something like 66 and "lose your market".

The diagram actually says 'White on roll' although the text underneath says Black. White has to be on roll here, else it's a trivial question.

I think it's a clear double for white:
- 14 out of 36 rolls (39%) hit the blot.
- White is ahead in the race
- If white misses, black has to run with his blot, and white probably then has at least two safe rolls, since there are enough stones on his 6 and 7 point to play awkward rolls, whereas a black 6 then means further trouble for black.

I would double as white. I'm not sure it's a take for black.

On pip count alone it's a drop, surely with the added risk of having a checker exposed and a 5 point trap the situation is decidedly gloomy to my eyes.

I am somewhat flabbergasted, because gnu bugs keep teasing me. I posted this position because of conflicting information. However, using our common sense it is no decision problem at all. My apologizes.

Position ID: 2m4UADi2bSEBUA Match ID: cAkAAAAAAAAE


What are the cube decisions?

Switch to XG. It's a great program. You'll love it.

This is one of those situations where I think a double is correct, but I don't have the courage OTB to pull the trigger. Very silly, but my lack of confidence in my decisions, and the knowledge that if you are wrong, you are very, very wrong lead me to hesitate in situations like this.

My reasons for doubling: the exposed checker on the 2 point, as well as entry on the 1 point mean we are pretty likely to get out. Hitting white at this point would be positively fantastic. Dancing isn't the end of the world, without a 6 white is going to crunch pretty soon. If we get out, our chances of covering the 7 point are high on the same or next throw, and we have a couple of throws before we run out of play.

Assuming the double is correct, it's a take for white. Black may dance and then we are in with a chance.

well said,

I run ubuntu. I don't love windows.

I don't 'love' windows either. But it's a very useful tool and since almost all third-party software runs under windows, I use it.

I can't say that I 'love' hammers, but if I need to drive a nail, a hammer is an excellent choice, regardless of any aesthetic flaws in its design.

BTW has anyone tried XG with wine?

I'll join the chorus of XG praise. If I had a mac and I couldn't get XG running well on it, I'd probably buy a cheap Windows laptop to just run XG. I just wish it would stop telling me how badly I handle the cube .

As to the cube problem, I already checked the bot so I won't say too much, but it's a great position to try to break down the rolls -- 16 dances, 12 hits, 8 other numbers. Assume double and take. Try to estimate how many of each category White wins after accounting for gammons. 2 gammons = 1 extra loss/win. So if you think White wins 2 out of the 12 hits (if you think he's 1/6 to win after being hit), but loses 6 gammons out of the 12 (if you think he's 50% to be gammoned after being hit), that would count as -1 wins for White in those 12 games.

And then see if White wins more or less than 9 games for the take/pass decision. And if it's lower than 9 but pretty close, or there's some question about what side of 9 it falls on, you would know it's a double.

When coming up with numbers for the three categories, remember that White has the advantage of owning the cube so you have to adjust upwards a bit on your win numbers if you are thinking cubelessly.

Off topic, but how do I use the Position ID in XG? I presume it can be entered somewhere which saves having to manually set the position up?

As my topic is up for grabs in anyway, I can tell you as well that pressing ctrl+c in the one screen and ctrl+v in the other screen has a great chance of succeeding.

Position ID: 2m4UADi2bSEBUA Match ID: cAkAAAAAAAAE

White - Pips 134

Black - Pips 120
Black on roll. Cube action?

Back when Bill was doing the PotW, it was usually a safe bet that the solution to one of his cube positions was Double/Take. It seems odd to say it, but when Bill tried to insert some No-double and Pass positions into the mix, they were much easier to spot, and hence, less interesting than his regular fare. My guess is that we will discover the same thing holds in this thread.

In Position 1, if White gets trapped in an ace-point game with 4 checkers back, and his board crashes, I recall his winning percentage falls to something like 5%-6%. If that is right, the equity from his ace point game may be worth 2 wins out of 36 games. If White mangages to hold onto his board, then his winning chances rise to something like 15%-18%. Under either scenario, White loses a gammon 55%-60% of the time. With two other targets outside, the gammon number could go much higher. If Black scoops up all of White’s blots, there will also be a few backgammons to consider.

When Black hits on his first try, therefore, White will be under enormous pressure to respond immediately. Otherwise, the lights may go out faster than he can blink. From the bar, White will have 11 entering numbers. 61 or 11 is the joker he really wants, but if Black still has two blots in White’s board, 41 or 31 will do. With the latter pair, White will hit loose again, and hope for a dance.

The other two entering numbers White has, 51 and 21, are potentially worse than dancing. After he plays either of those, Black will have—in theory—27 numbers that complete his prime. If that happens, White will be well on his way to a crashed board. With some of Black’s rolls, however, Black may decide he should jump out instead of priming, especially if he can hit a fly shot. In addition, 52 and 53, which do not complete the prime, comprise four of the hitting numbers. Either way Black plays it, White will be in trouble.

When White dances, of course, then all of Black’s priming and hitting options will be available.

Black can also enter without hitting. 61 and 41 are particularly hurtful to White. With his other aces, 51 and 31, Black can simply lift the exposed blot on his bar point. All 20 entering numbers play quite nicely for Black.

As bleak as that sounds for White, things could get just that dark for Black. When Black rolls one of his 16 dancing numbers, then White will have eleven 6s that hit, rolls of 3, 7, and 8 (20 numbers, not counting 6s) to make his 2pt, plus 41 (and a couple switching doublets) to make his 1pt. Just as with White, Black would be under immense pressure to respond immediately. Could he take a cube or recube if he danced with two up?

Ultimately, getting a 6 is the real key for White. Otherwise all his checkers are headed for the lower points of his board. Closing Black out won’t help if White cannot launch a runner soon.

To be honest, all of this is very confusing to me. This looks like a highly volatile two-way gammon position. Since Black rolls first, and has a 33% chance to really hurt his opponent, I would cube. In practice, it’s a great cube—even if it is wrong—because there will be many who pass.

Is it a take? Heck if I know! But wait a second. What about Woolsey’s Law? If I am not sure whether to take or pass, then this must be a cube!

And as for the take, this is where I came in. If I can’t figure out one of these cube problems, probably it is a take.

Mike

Position ID: fwcAAKTtNgIwAA
Match ID: AQEAAAAAAAAA

In order to think about this correctly, you need to know that if Black had 5 off and White closed out one checker, that would be a very close take/pass decision if White has perfect spares.

Here, we have 19 rolls (including 66) that make a 6 prime which is almost as good as closing out a checker. So those will result in a solid pass. So the 19 good numbers result in a pass and the 17 bad numbers result in a take on average. Is the pass after the good numbers bigger than the take after the bad numbers? I think it is, considering that there are lots of 5s that cover the 2 point. Like if White rolls 4-1, Black is maybe happy he took, but it has to pretty close. There are only 6 numbers that don't cover the 2 point, and those still leave Black a big underdog.

So double/pass, but the pass should be fairly close, I would think.

too many market losers, so double.

drop only 4 off.

new problem...

not that i'm all that good, lol

Aargh!, I went through this chapter in 501 a couple of months ago and can't quite remember the details. More revision needed.

I think it's a double/drop, but I must confess to reading the answers given which has no doubt swayed me.

I believe the take-point is at 5 checkers born off, but that's with no cube. Is the equity difference multiplied by the cube, so that the decision changes with the cube?

Beginner's question with a very simple answer. No, the cube decision remains the same irrespective of the amount of money which is played for. So 75% for double and 25% for take. I must have had some brain freeze. The question was motivated by the fact that the double-point in a race is slightly higher when the cube has been turned.

Position ID: mOeGAQZwt00AIQ Match ID: cAkAAAAAAAAE