obvious is not always right - Page 5 - Backgammon Forum - Discuss Backgammon Strategy, Clubs and Books

 1. 20/13                        Eq.:  +0,096 
       0,522 0,157 0,008 - 0,478 0,137 0,004 CL  +0,068 CF  +0,096 
     
    2. 9/7 8/3*                     Eq.:  -0,040 ( -0,136) 
       0,489 0,161 0,007 - 0,511 0,169 0,006 CL  -0,029 CF  -0,040 
     
    3. 9/4 6/4                      Eq.:  -0,046 ( -0,142) 
       0,484 0,170 0,007 - 0,516 0,167 0,004 CL  -0,026 CF  -0,046 

    4. 21/16 6/4                    Eq.:  -0,624 (-0,720)
       0,408 0,133 0,009 - 0,592 0,359 0,047 CL  -0,449 CF  -0,624
      
        Truncated cubeful rollout (depth 10) with var.redn. 
        147 games, Mersenne Twister dice gen. with seed 765213078 and quasi-random dice 
        Stop when std.errs. are small enough: ratio 0,1 (min. 144 games) 
        Play: world class 2-ply cubeful prune [world class] 
        keep the first 0 0-ply moves and up to 8 more moves within equity 0,16 
        Skip pruning for 1-ply moves. 
        Cube: 2-ply cubeful prune [world class]

The thing is not that slotting and leaving blots is bad. There is no connectivity anymore, and black would like to use that pile on his 6-point to hit loose. In other circumstances he would like to start a new life at white's side of the board by recycling, and most of us are wired to do that. Even that 21/16 6/4 can turn out very well, and it's evaluation shows how bad the bot is in assessing chaotic positions. Though there is a realistic chance that white will hit twice or hits and covers, white's homeboard is so deplorable that black can hardly do anything wrong.

However, black has got a better strategy here. The white checker can be captivated behind black's block, so this should be left intact. Playing 20/13 has several plusses. It's not so bad if white hits, because then black can recycle, and if white hits on the 13-point he loses his control point. In case white doesn't hit, if black is lucky to throw a 6 he can make the 6-prime, and otherwise can attack the checker.

NOTE: My analysis is arrived at after bot experimentation.

Quote

01-18-2015 , 06:39 AM

#102

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

White - Pips 160

Black - Pips 163
Black to Play 5-1
XGID=-b----E-D---dD--ac-d--a-AA:0:0:1:51:0:0:3:0:10

This is the third roll, after 52-split and 43-hit.

Mike

Quote

01-18-2015 , 08:23 AM

#103

yogiman

adept

Join Date: Jun 2012 Posts: 784

No matter what the bot says, I would play bar/20 6/5. As already mentioned in my other topic, if the opponent has no board or a bad board, I like to leave blots (in a not to unsensical way). Particularly when there is a gap in cube handling this can be both exciting and profitable.

Quote

01-18-2015 , 10:18 AM

#104

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

Yeah I guess it's either B/20 6/5 or B/20 24/23.

Quote

01-19-2015 , 08:29 PM

#105

_Z_

adept

Join Date: Jul 2004 Posts: 720

The duplication of 4s looks strong enough for me to play B/20 6/5.

If White's 3 point was locked up, I think slotting would be too dangerous.

Quote

01-20-2015 , 01:03 AM

#106

bleep69

grinder

Join Date: Sep 2008 Posts: 483

Quote:

Originally Posted by _Z_

If White's 3 point was locked up, I think slotting would be too dangerous.

Really? One extra point in white's home board and the play would change? You are still a massive favourite to come out, or is the blitz a real possibility with that extra point?

Quote

01-20-2015 , 02:10 AM

#107

_Z_

adept

Join Date: Jul 2004 Posts: 720

I'm thinking it's a fairly close play and the stronger board is enough to swing the decision.

Quote

01-20-2015 , 04:09 AM

#108

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Problem 24. Black to Play 5-1
XGID=-b----E-D---dD--ac-d--a-AA:0:0:1:51:0:0:3:0:10
White - Pips 160

Black - Pips 163

This is one of the exceptions that makes the rule. Usually it is dangerous to slot and split at the same time. In Problem 24, however, it is correct. The strong duplication of 4s means the downside to slotting is less than it would be otherwise. 4 is a good roll for opponent White whether Black slots or not. The upside, when White does not roll a 4, might be quite large. In addition, there is no strong alternative, and slotting lets Black unstack his 6pt to the optimal destination while White’s inner board is weak.

A 3-ply, unlimited-game rollout in XG has the splot (bar/20 6/5) beating two-up (bar/20 24/23) by 30 millipoints (of equity). That’s 0.03, big enough to make the splot a clear winner.

Inspired by the messages from Z and Bleep, I ran XGR++ evals for two variants. If you let White make his 3pt by moving 9/3, then there is a 65 millipoint swing. Now, playing two-up (bar/20 24/23) beats the splot (bar/20 6/5) by about 35 millipoints. On the other hand, when you give White his 3pt by moving 6/3 instead, then the swing is only about 5 millipoints. Splotting is the clear winner by 25 millipoints.

Explanations?

Variant 1. Two-up is best.
XGID=-b----E-D---dD---c-d--b-AA:0:0:1:51:0:0:3:0:10
White - Pips 154

Black - Pips 163

Variant 2. The splot is best.
XGID=-b----E-D---dD--ac-c--b-AA:0:0:1:51:0:0:3:0:10
White - Pips 157

Black - Pips 163

Mike

Last edited by Taper_Mike; 01-20-2015 at 04:19 AM.

Quote

01-20-2015 , 05:23 AM

#109

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

To clarify:

In Variant 1, 4 is no longer duped. That explains why playing two up is best. Slotting against a stronger board is a clear error.

My question has to do with Variant 2. Like Z, I expected that giving White a stronger board would weaken the splot. In fact, it does, but not by very much. Why is that?

Interestingly, if you switch White's made point in Variant 2 from the 3pt to the 4pt, giving him an even stronger board, then the splot wins by a larger margin (about 33 millipoints).

Mike

Quote

01-20-2015 , 09:49 AM

#110

yogiman

adept

Join Date: Jun 2012 Posts: 784

Position ID: O5wzAiDsbgYAQw Match ID: QYkWAAAAAAAE

White - Pips 107

Black - Pips 128
Black to Play 5-5
Created with www.BGdiagram.com

Quote

01-20-2015 , 12:06 PM

#111

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

Hesitating between B/20 21/6 and B/20 21/16 6/1*(2) but my QF nose tells me it could be the latter.

Quote

01-20-2015 , 12:08 PM

#112

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

Somehow I don't like B/20 8/3 6/1*(2).

Quote

01-20-2015 , 01:34 PM

#113

_Z_

adept

Join Date: Jul 2004 Posts: 720

Switching looks right by a ton to me. B/20 21/16 6/1*(2). If White dances, he's in big gammon danger, and if he enters Black will still have plenty of attacking/hitting chances.

If instead Black plays B/20 21/16(2) 8/3 (second best, I think), it's White that's now attacking.

Switching is ugly, but backgammon isn't a beauty contest!

Quote

01-20-2015 , 01:39 PM

#114

_Z_

adept

Join Date: Jul 2004 Posts: 720

Quote:

Originally Posted by Taper_Mike

Variant 2. The splot is best.

That is the position I had in mind when I figured the stronger board would swing the play against the splot. I don't get it either. White's stronger board should dictate a safer play, all things being equal. Obviously something isn't equal, but I don't see what.

Quote

01-20-2015 , 06:52 PM

#115

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Quote:

Originally Posted by _Z_

[Variant 2] is the position I had in mind when I figured the stronger board would swing the play against the splot. I don't get it either. White's stronger board should dictate a safer play, all things being equal. Obviously something isn't equal, but I don't see what.

Variant 3. Black to Play 5-1
XGID=-b----E-D---dD--ac-c-b--AA:0:0:1:51:0:0:3:0:10
White - Pips 159

Black - Pips 163

Perhaps the answer is revealed in Variant 3 (above).

Here, the splot (bar/20 6/5) is stronger, by a small amount, than in the OP (Problem 24). Slotting wins the XGR++ eval by 33 millipoints.

I suspect there are two forces at play. One is the need to avoid risky slots when your opponent has the stronger board. As we saw in Variants 1 and 2, that works against the splot.

On the other side is the need to challenge a budding prime immediately. In Variant 3, the opponent has started to build a threatening prime. If White makes his 5pt on the next turn, Black is going to be in deep trouble. It does not matter whether Black gets hit in the process or not. That being the case, Black should try the big play that might swing the game back in his direction. Splitting to the 20pt gives Black a chance to advance his anchor, and thereby defuse White’s prime. When Black’s slot on the 5pt is hit, it might actually help him to make the 20pt anchor. Otherwise, he can start working on a prime of his own.

Both of these forces are also at work in Variant 2. In that position, the threat of the prime is less. Evidently, however, it is still enough to almost balance the danger represented by White’s improved board.

Mike

Last edited by Taper_Mike; 01-20-2015 at 07:07 PM.

Quote

01-20-2015 , 07:01 PM

#116

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Quote:

Originally Posted by _Z_

Switching looks right by a ton to me. B/20 21/16 6/1*(2). If White dances, he's in big gammon danger, and if he enters Black will still have plenty of attacking/hitting chances.

This sounds right to me.

I confess, I read this before I worked it out for myself. Usually, I try to avoid that.

Mike

Quote

01-20-2015 , 08:17 PM

#117

_Z_

adept

Join Date: Jul 2004 Posts: 720

Quote:

Originally Posted by Taper_Mike

On the other side is the need to challenge a budding prime immediately. In Variant 3, the opponent has started to build a threatening prime. If White makes his 5pt on the next turn, Black is going to be in deep trouble. It does not matter whether Black gets hit in the process or not. That being the case, Black should try the big play that might swing the game back in his direction. Splitting to the 20pt gives Black a chance to advance his anchor, and thereby defuse White’s prime. When Black’s slot on the 5pt is hit, it might actually help him to make the 20pt anchor. Otherwise, he can start working on a prime of his own.

I buy that Black is more likely to get an advanced anchor when he slots, and that it is more important to get an advanced anchor after White has put his position together a bit.

But your argument is essentially saying that Black should be less averse to getting hit on his side of the board by virtue of the fact that White has a stronger front position. Maybe that is what's going on in these specific positions, but it runs counter to all accepted backgammon theory.

Interesting stuff in any case.

Quote

01-20-2015 , 09:23 PM

#118

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Quote:

Originally Posted by _Z_

But your argument is essentially saying that Black should be less averse to getting hit on his side of the board by virtue of the fact that White has a stronger front position. Maybe that is what's going on in these specific positions, but it runs counter to all accepted backgammon theory.

Don't forget the powerful duplication of 4s. That is present here, but not in the more general case you cite.

Black is pretty much stuck entering on the 20pt. Once he does that, my guess is that White hits with any 4. All the slot does is move the hit from one side of the board to the other. Of course, as is often the case with duplication, Black will be toast when White tosses a 44. Otherwise, he is not risking a double hit when White rolls 41 or 43. White should be pointing with those rolls. In any event, those rolls dish out significant hurt even when Black plays two-up.

Look at it another way. The biggest problem with the splot is a hit and dance. Isn't that almost just as bad when Black plays two-up?

Mike

Last edited by Taper_Mike; 01-20-2015 at 09:38 PM.

Quote

01-21-2015 , 04:14 AM

#119

bleep69

grinder

Join Date: Sep 2008 Posts: 483

I'm pleased to say that I got the B/20 21/16 6/1*(2) move right, but the only reason I got it right is because of the "Obvious is not always right" thread title, which forces you to look at alternative moves. I'm pretty sure that OTB, where there is no distinction between obvious and not-so-obvious, I would have played the obvious move.

Thanks for running the various positions Mike. I need to go through the responses a few more times to try and make sense of them properly.

Quote

01-21-2015 , 06:28 AM

#120

yogiman

adept

Join Date: Jun 2012 Posts: 784

Quote:

Originally Posted by bleep69

I appreciate your response, but it is because of this reason that you are asked to come up with arguments for your move. And peeking is forbidden, except for me!

Quote

01-21-2015 , 07:19 AM

#121

yogiman

adept

Join Date: Jun 2012 Posts: 784

If that blot would not be present the two plays would be about even money, but it has a great gammon potential for black. By hitting, that blot is momentarily frozen, and that makes 6/1*(2) stand out.

Code:

    1. bar/20 21/16 6/1*(2)         Eq.:  +0,292
       0,623 0,270 0,008 - 0,377 0,094 0,003 CL  +0,426 CF  +0,292
     
    2. bar/20 21/16(2) 8/3          Eq.:  +0,174 ( -0,118)
       0,620 0,170 0,005 - 0,380 0,089 0,002 CL  +0,324 CF  +0,174
      
        Truncated cubeful rollout (depth 10) with var.redn.
        147 games, Mersenne Twister dice gen. with seed 764685121 and quasi-random dice
        Stop when std.errs. are small enough: ratio 0,1 (min. 144 games)
        Play: world class 2-ply cubeful prune [world class]
        keep the first 0 0-ply moves and up to 8 more moves within equity 0,16
        Skip pruning for 1-ply moves.
        Cube: 2-ply cubeful prune [world class]

Do you think things would change very much if the two checkers on the 8-point are moved to the 10 point?

White - Pips 107

Black - Pips 132
Black to Play 5-5
Created with www.BGdiagram.com

Quote

01-21-2015 , 09:57 AM

#122

bleep69

grinder

Join Date: Sep 2008 Posts: 483

Quote:

Originally Posted by yogiman

I appreciate your response, but it is because of this reason that you are asked to come up with arguments for your move. And peeking is forbidden, except for me!

I find this very difficult - the rationale, not the peeking. For very obvious things like making a point in the home base it is simple, but the more esoteric reasoning seems to be beyond me at this point.

What makes it very frustrating is that when I then read what Taper Mike, or anyone actually says about the situation, it isn't esoteric at all, but plain common sense.

I get the same feeling when going through 501 Problems. Sometimes I stumble on the correct move, and then when the text explains why it is so, it could be nothing but!

Practice, practice and more practice, both at the game, and the formulating of the reasons behind the moves...

Quote

01-21-2015 , 11:12 AM

#123

yogiman

adept

Join Date: Jun 2012 Posts: 784

Ofcourse you are not alone. At least you understand when it's told. Here you have a good training ground to learn to verbalize your analysis.

Back to the solution of the point shift. 6/1(2) is still best, though not so much anymore, because black's equity has dropped significantly.

Code:

 
1. bar/20 21/16 6/1*(2)         Eq.:  +0,176
       0,586 0,243 0,007 - 0,414 0,102 0,004 CL  +0,316 CF  +0,176
    
2. bar/20 21/16(2) 10/5         Eq.:  +0,138 ( -0,039)
       0,610 0,160 0,005 - 0,390 0,096 0,002 CL  +0,285 CF  +0,138

        Truncated cubeful rollout (depth 10) with var.redn.
        144 games, Mersenne Twister dice gen. with seed 764552881 and quasi-random dice
        Stop when std.errs. are small enough: ratio 0,1 (min. 144 games)
        Play: world class 2-ply cubeful prune [world class]
        keep the first 0 0-ply moves and up to 8 more moves within equity 0,16
        Skip pruning for 1-ply moves.
        Cube: 2-ply cubeful prune [world class]

The reason is presumably that in the original problem those two black checkers point on the 2-point, together with the builders on the 4&5 points. If white enters with a two dice, he can only get away from his confinement by 24 or 25. However, with the 4 he is going to hit and with the 5 he will safety his other blot. So there are many more ways to hit loose or make the 2-point compared to the latter problem.

Quote

01-22-2015 , 06:53 AM

#124

yogiman

adept

Join Date: Jun 2012 Posts: 784

Position ID: bXYYAwiwrw0BBg Match ID: cAkFAAAAAAAE

White - Pips 107

Black - Pips 136
Black to Play 2-1
Created with www.BGdiagram.com

Quote

01-22-2015 , 01:02 PM

#125

_Z_

adept

Join Date: Jul 2004 Posts: 720

The safe play of 13/11 7/6 looks like it's playing not to lose as opposed to playing to win. The game plan is to contain the back checker. 6/3* is the usual tactic here, but the problem is it's something like 20 shots. You can play 13/11 6/5 and leave fewer shots and one less blot for arguably a better containment situation. So that's better than hitting.

But actually there's this other play -- 6/5 6/4 which 20 shots and 3 blots, but it's really great for containment when missed. I think you'll be cubing a lot after getting missed. It's a little hard to see over-the-board, but I think it might be best.

Quote

Page 5 of 13

First

1 2 3 4 5 6 7 8 9 10

Last

Post Reply Subscribe

...

Page 5 of 13

First

1 2 3 4 5 6 7 8 9 10

Last