Problem of the Week #11: May 18




Cash game, center cube. Black to play 6-1.

8/1*. Partly because it is against one man back and hence we attack; partly because other moves all look bad to me.

If we play 6 with the back men, we have to leave 3 blots which is too dangerous against white's 3p board and 3 (or 5) spare attackers. We can also to 13/7 for the 6, but then the 1 sucks: 24/23 is too dangerous, 7/6 make the position awkward, 13/7 6/5 leaves double direct shot and only one spare attacker while after 8/1*, only 1s and 26,53 hit and we have 4 attackers.

Messing up your gapped 6 prime to attack deep seems really bad to me. You hold the 7 point, so there's not much value in making the 1 point (which is the logical followup after hitting there). You don't have enough checkers down there to complete a closeout since you're still back at the midpoint. I think that this play is a pretty big blunder.

How do you win from this position? You can't win a race and you can't blitz. You also can't hang out in the outfield to try to pick off that last checker when it makes a break for home. That leaves you with your prime as the only path to victory.

I'd play 13/7 6/5. Even though you're down by 15 in the race, your timing is terrible because too many pips are tied up in your ace point anchor. This means that getting hit isn't as bad as it otherwise might be because you preserve your timing a bit.

White is desperate for a 4 to run that back checker, and if he fails to get it, you'll have two spares for a 6 prime, plus you would break the 8 point to make the 5 point (which makes a connected 5 prime with the back of the prime slotted -- a 5.5 prime?), giving you what amounts to a triple shot to cover. Also, if white fails to hit your blot at the midpoint, you've got a few extra indirect shots to make the 5 point as well.

Making your 5 point is huge in terms of becoming a real threat to white. If I rolled a 21 here, I would break the bar point to make the 5 point and give him the 6 to escape. If he fails in the next roll, you're looking really good with a double direct to cover and force white to roll exactly 1-6 to escape.

I'm with Aaron on 13/7, 6/5. To win this game you need to find a way to contain white's back checker over the long term. Making the five point is huge in this regard and is well worth a little risk. As Aaron said, getting hit isn't really a problem given that you're already well behind in the race and white's board is nothing special. Getting hit may give you the extra timing you need to evolve into some sort of holding game. As it stands your timing is hopeless.

Does 24/18, 6/5 just leave way too many blots? The idea is to leave the best chance of making the 5 or 18 points next roll should nothing too disastrous happen.

I'm still learning and I think I know why this move isn't good, but it would really help if I could get 1 or 2 really solid reasons.

1) Breaking that anchor gives white all sorts of opportunities to pound on you. You leave white with a double shot (4x/6x) as well as a pointing roll (53). These sequences lead to a lot of gammons against you, but it's not as if your play increases your gammons very much.

2) If you're slotting the 5 point, you want to maximize your chances of actually making that point. After your play, you only have indirect shots to cover.

Thanks for the explanation! I completely understand point number 2, however, I have to ask, how relevant is the part in no.1 about "These sequences lead to a lot of gammons against you, but it's not as if your play increases your gammons very much" ?

Is it because this game is "gammonish"? As in, either one or the other side stands a significant chance to win a gammon depending on how the luck goes? In such a game, it does seem logical that you would be attaching more weight to increasing your gammon chances in any given play, but I want to be sure I understand the principle.

Black isn't going to win many gammons. Even if black primes the back checker, white will have plenty of time to get most of his checkers around before black can close him out, so there aren't too many gammons there (white might leave a shot or something, but it's not as if black has all day to wait around for one. The alternate route this game takes is that black goes into an ace point game, but that doesn't win many gammons there, either.

White shouldn't be getting too many gammons here because black holds an anchor (white cannot close black out). But if you lose the anchor, white can immediately go into a blitz and potentially close out, preventing black from advancing his outfield checkers any further.

Yeah I looked very briefly at 24/18 just because 13/7 breaking the midpoint is not particularly appealing, but it is just way too loose, plus the impact of the extra builder with 13/7 is significant.

24/18
24/23
nothing else better than that.

8/1* for me here too. It's not great, but it's the best I can come up with. Splitting the back checkers is a no-no, I think. Whites hitting numbers are well diversified, so we could easily find ourselves with two men on the bar next roll.

13/7 6/5 is the only other option to consider, but I feel that leaving a double shot is too much and even if we do make the sixprime (which is by no means guaranteed even if white misses) escaping the back checkers before our prime collapses again is gonna be a problem.

13/7, 6/5. I'm in agreement with the analysis Aaron W. offered. No other play lets black turn the cube correctly if white misses hitting next roll.

Thats my choice to.

I think 8/1* is really, really bad btw, hitting loose on the ace point is basically never good for starters. Plus you're breaking up your prime and any number white rolls to roll on is bad for you. The 1 hits and the 2 and 5 position him to escape your gapped prime with a single good roll, at a time when you're still loose in your home board and have no builders in support. My second choice is 24/18 24/23, but I think 13/7 6/5 is definitely better.

I wouldn't play 24/23 24/18 in a million years. But if you really like pain...

I just did the GNU rollouts. Here they are with a bit of commentary. I'll put the whole thing in spoil tags in case there are still some people reading this thread who don't want the solution.