Problem of the Week #25: August 23


Cash game. Center cube. Black on move.




Part (a): Should Black double? Should White take if doubled?




Part (b): Should Black double? Should White take if doubled?




Part (c): Should Black double? Should White take if doubled?

hmm, well they all seem like doubles to me. the only difference seems to be that white's timing is better in b than a and better in c than b

maybe

D/P
D/T
D/T

a) double/drop
b) double/drop
c) double/take

The double seems obvious in all cases. As far as the take, timing is obviously the key issue.

Back games are too complex for a simple formula, but a rule of thumb I sometimes use is whether Villain has enough timing to handle big doubles one time (e.g., 44-66), as a proxy for various other scenarios where Villain starts crunching.

In a and b those rolls would be real ugly timing-wise. In c White wouldn't like such a roll, of course, but he wouldn't be completely cooked.

So given that the other pluses and minuses of this specific position seem to roughly offset, I'd go with that approach.

Obviously, this a problem about timing.

a) Too good to double. White's position is going to crunch most of the time, and the times that it doesn't crunch, black must have had very bad sequences, or white rolled 66 to escape the checkers on the 3 point. I think you're letting white off the hook by doubling here instead of playing on for gammons.

b) Do 6 pips make enough of a difference? Probably, otherwise this wouldn't be a separate problem. I only say it's a double/drop because I think the first one is too good. Over the board, I'd likely keep on rolling and play for the gammon.

c) This one is a double/take. White's timing is now pretty good, so it's going to be a take. Black should be doubling because I don't think he gets gammoned too often playing against a 1-3 backgame. I'd be more wary of doubling against a 1-2 or 2-3 backgame because two anchors next to each other creates many more problems bearing-in.

i agree with both of you now on b) as double/pass, but a) def doesn't seem too good to double for me. white is probably a favorite to crunch, but there are still plenty of sequences where either white rolls some small numbers and/or black rolls a few big ones

Agreed, now looking at Black having to play large doubles one time, he's not gonna be real happy and White will be in very reasonable shape in those lines. It's too easy for Black to fall out of cashing or even double/take territory if he rolls poorly.

Plus in the real world a double is absolutely mandatory IMO because you will get a ton of incorrect takes (assuming our evaluation is correct and taking IS incorrect), and also in position B.

Even if black rolls big doubles, he still needs to leave a WINNING shot for white for white to have a chance. Black still wins the game very often bearing in against a 1-3 anchor. Furthermore, even if white hits from the 3-point anchor, he STILL needs to escape his back two checkers. If the bar point has not been cleared yet, this will require very good rolling.

I think it really depends on the opponent when it comes to backgames. Some players love to play out of a backgame, some just want out so that they don't get gammoned.

Well, now we're getting into how to play a back game properly... White doesn't want to hit "too soon", and depending on the exact position, may well pass up a chance to hit. With adequate timing (e.g., if Black rolls big doubles), he has this sort of flexibility, whereas otherwise he has to hit and hope for the best.

I'm also not sure what you mean by a "winning" shot. White will get a (usually double) shot at some point virtually every time, and often multiple shots, and often the game will still have a bunch of play left if he hits (e.g., can White contain, can he pick up a second checker, etc.), and a full evaluation of all the possibilities is pretty complex.

I'm not claiming White is in great shape by any means, but I think you may be overestimating Black's equity here.


Yeah, but even if only some opponents are taking incorrectly, it's still a pretty big mistake IMO to wait. I mean, if Villain drops, you win the game, and that's pretty nice. It's not like you have him absolutely crushed, even in position A. And winning a gammon in a doubled instead of single game is pretty nice as well, vs. the other type of opponent.

And there's that whole Jacoby Rule thing as well, which I think is still assumed in cash games, although it's been >10 years since I played one. If that rule is in effect, a double is completely obvious, of course.

Perhaps, but we have to wait until next week for Bill to fill in his more expert thoughts on the issue.

A "winning" shot just means that he gets a shot where he's able to hit and then roll his checkers into the outfield to keep pounding you until he can close out his board. And as long as you've got the bar point blocked, he's going to struggle to get his checkers free even if he does hit you.

You bear in safely against a 1-3 backgame more often than you do a 1-2 or 2-3 backgame. The gap is a decent advantage for you, so I'm not sure if you're going to leave as many shots as you think. I think there's a section in Backgammon Boot Camp where (the recently deceased) Walter Trice talks about the three different backgames, and my memory tells me that the 1-3 backgame was the weakest of the three.

Also, leaving a shot or even getting hit doesn't mean that you automatically lose. So even if you get a checker sent back, you've still got equity. It all adds up.

*shrug* I would suggest that in position A, you had probably already waited too long to double and now your best play is to keep going forward.

This whole thing is a weakish line of argument because if you're too good to double and he drops, you're also giving away your equity. Just winning the game when you be winning gammons from those positions isn't "pretty nice." It's like not putting in a value bet on the river when you would have gotten paid off. Yeah, you won the pot and I guess you can feel good about it, but you left some money on the table. (This type of error is magnified in match play.)

But really, we're talking about the difference between strategically correct and technically correct. You can justify all sorts of things as strategically correct if you know that it will induce errors in your opponent's play.

This is not right. You need absolutely tremendous timing to time a 1-2 game because the defender will kill his #s, so much that it virtually never comes up in practice (for the endangered species example of a well timed 1-2 game, though, you can see Robertie's own Advanced Backgammon, vol. 1, #165).

Since this is a money game with centered cube the Jacoby rule is usually in effect meaning that you cannot win a gammon unless the cube has been turned. If you owned the cube I would agree that this is marginally TG.

My guess is that #2 is a marginal drop and #3 is a clear take.

I lose the memory game.

Bearing in (best to worst): 2-3, 1-3, 1-2
Bearing off (best to worst): 1-2, 1-3, 2-3

Looks like I have another reading project.

I would say that the modern way of looking at this is that there is no "best" or "worst" backgame. The problem is timing. The majority of the time you will not be able to achieve a well timed back game, so a 2-3 is "best" in the sense that it can overcome the problem of inadequate timing since it gets shots earlier.

Once the board collapses pretty much any game is very bad. If you can manage a well timed back game you almost never have a choice between them since your opponent is so far forward if you actually have timing. I could make a wild conjecture that in the well timed case you prefer the deepest points you can have. But in practice the 1-2 just almost always crashes (not that other back games don't, but you get the idea).

Of course, but the point is that if you are too good in A, it is very close IMO (it's not like you have 8 checkers trapped or anything, he can just give up on the back game and run off the gammon in many scenarios) and you are gaining just a little bit of equity by continuing on vs cashing, whereas if Villain takes incorrectly it's a large mistake, and in this type of position you see loose takes all the time unless you are playing an absolutely world-class player. So the risk-reward is WAY in favor of spinning the cube.

And now is the time to do it, because it's the type of deceptive position where even a very good player may space out and take, whereas waiting even one roll that doesn't go White's way, and now the drop becomes very clear.

a) No double, take.
It doesn't look like black is ahead enough in the race and he'll most likely have to start taking apart the prime pretty soon. White definitely has some problems in this position, but I don't think black is ahead by enough here.
b) double, take
c) double, take

In c, I think the double is clear and in b, it is rather borderline.

Dude, Black has trapped four of White's checkers behind a 5-prime. He doesn't WANT to be ahead in the race at the moment.

"The best backgame is the one when your opponent is playing it"-Unknown, probably Arkaidy Tsinis though.

All of these are doubles against anyone and unless you are an expert you should probably pass all of them, including the last one with the better timing. Snowie can probably take the last one as it will know exactly how to maximize his timing, when to abandon the backgame if necessary and when to reship the cube if things go well.

These are not trivial decisions, and if you are thinking of accepting cubes in backgame situations you need to remember that you have the more difficult game to play to achieve all of your equity.

Yeah, I hate these positions, cause usually when you reach them the cube has already been turrned earlier on. I will just await the answers, and hopefully learn some good guidelines.

Btw, it's a cash game so we can't be too good to double with the cube in the center.

"I didn't get really get good at hold'em until I started playing more hands than most." -- Schneids (paraphrased)

This mentality is like folding all "marginal" hands all the time because you're afraid of getting yourself in a hard spot. The illusion of playing safe collides head on with the reality of making poor decisions, and you often lose in the long run.

At some point, you should want learn how to play a backgame well, and you'll never learn if you never play them. Try to make "technically correct" decisions and learn what you can from the mess you've made.

Bravo, Aaron!

The way to learn any game is to keep making the plays you think are correct. Afterwards, you study and analyze to refine your notion of "correct". Do that over and over again and, while you'll take some lumps, you'll make steady progress. Otherwise you just get stuck at some lower level and can't improve.

You should definitely learn to play backgames as that is part of the game etc...

But when analyzing whether something is an apporpriate take or pass, one's ability to play the resulting position is very relevant imo.

Take this line of reasoning to its logical conclusion: If you're bad at backgammon, you should never play.

C'mon dudes, sheets is making a valid point.

If a take is very marginal in theory, and you are likely to face many tough decisions, and your opponent's moves will virtually all be forced, you are temporarily the worse player (regardless of the actual skill levels) and may want to drop.

He didn't say "never play back games". There will be plenty of other ones in your future. It's pretty damn unlikely that you are going to learn something tremendous from one game.

Plus, I think it's more than a bit of a stretch to generalize from what sheets said to "never play backgammon".

I'm not saying that the point is invalid. The underlying (philosophical) question is this: On what basis you should make a decision?

My perspective:
* Aim to make the technically correct decision

Sheet:
If you believe that the position is a take, but then you choose not to play because you're not sure whether you'll play it well, you're giving up equity now out of fear that you *MIGHT* give up equity in the future. Taken to its logical conclusion, a player who isn't very good (who has an high probability of giving up future equity) probably shouldn't even play the game because he's bound to give up a ton of equity.

Giving up on spots because you think you might screw them up in the future (especially coming out of a quiz problem!) is a poor decision-making heuristic. You don't learn how to play well by practicing playing poorly.

Ok how about...."If the difference between 2 plays or decisions is very small, then the less experienced player should opt for the position which is easier to play, while if the difference is large, than he should go with the "correct" play."

official disclaimer: This in no way suggests to quit backgammon if one cannot play perfectly.

Hopefully we can all live with that.

FYP.

Fear has nothing to do with it. It's a straightforward EV calculation (at least in theory) given that one recognizes one is not perfect, which is apparently verboten to acknowledge for some reason around here.

Add a correction factor for "cool stuff I might have learned if I had played this game out" if you like. Then subtract another correction for time equity if you DO have an edge on your opponent in non-forced-move positions, because back games take forever to play.

This statement is both incorrect and irrelevant, and if you don't see that, I'm afraid it's beyond my ability to make you see it, imperfect being that I am.

So I'll stop there.