obvious is not always right - Page 3 - Backgammon Forum - Discuss Backgammon Strategy, Clubs and Books

Problem 12
There is a danger that Black could get trapped behind White’s block. He needs a 5 or 41 to escape, and since he has a 5 now, perhaps he should run. To my eye, 23/12 seems the best way to do that. Black should have an easy time picking up the third blot.
Mike

Right, when you play it hasty you would be tempted to hit, however, the third blot can wait and is also not really important for the gammon. A bad double and/or not being able to throw that 5 in time could easily convert a sure gammon into a loss.

Code:

    1. 23/12                        Eq.:  +1,522 
       0,894 0,748 0,039 - 0,106 0,016 0,001 CL  +1,557 CF  +1,522 
     
    2. 18/12 14/9*                  Eq.:  +1,411 ( -0,112) 
       0,862 0,715 0,037 - 0,138 0,018 0,001 CL  +1,458 CF  +1,411 
       
        Truncated cubeful rollout (depth 10) with var.redn. 
        145 games, Mersenne Twister dice gen. with seed 763539624 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]

Quote

01-03-2015 , 09:11 AM

#52

yogiman

adept

Join Date: Jun 2012 Posts: 784

Position ID: m/UGADCY25QBDA
Match ID: QYkJAAAAAAAE

White - Pips 105

Black - Pips 147
Black to Play 3-2
Created with www.BGdiagram.com

Quote

01-03-2015 , 07:06 PM

#53

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Keep ’em coming. These are great positions.

Black holds the hammer in Problem 13, but the game could turn around in a hurry if he breaks his anchor. White holds the stronger board, so Black does not want to get hit and dance. I would make the 10pt and 11pt.

Mike

Quote

01-03-2015 , 07:14 PM

#54

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

I'd go with 22/20*/17 and hope I don't get struck by lightning.

Quote

01-04-2015 , 08:09 AM

#55

yogiman

adept

Join Date: Jun 2012 Posts: 784

The good thing about 22/20*/17 is that it is a winning move

. However, making an extra blocking point with 13/10 turns out to be more important than having an extra blot within white's homeboard. Though we could do better (understatement).

Hitting does several good things for white:
minor:
-he can hit back in his own homeboard
-55, 56 hit a blot from the bar
major:
-it increases his chances to make that 5-point anchor
-most importantly it gives white the opportunity to recycle a checker and as a consequence to avert the breakdown of his homeboard.

13/11 13/10 directs his strategy to making his 5-point, at the same time creating a broken 6-prime.

Code:

    1. 13/11 13/10                  Eq.:  +0,570 
       0,748 0,221 0,010 - 0,252 0,071 0,001 CL  +0,656 CF  +0,570 
      
    2. 22/20* 13/10                 Eq.:  +0,286 ( -0,285) 
       0,676 0,218 0,013 - 0,324 0,170 0,010 CL  +0,403 CF  +0,286 
     
    3. 22/20*/17                    Eq.:  +0,224 ( -0,347) 
       0,664 0,207 0,012 - 0,336 0,197 0,013 CL  +0,337 CF  +0,224 
      
        Truncated cubeful rollout (depth 10) with var.redn. 
        146 games, Mersenne Twister dice gen. with seed 763710237 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]

Quote

01-04-2015 , 12:23 PM

#56

ChanningW

stranger

Join Date: Jan 2015 Posts: 9

It wouldn't surprise me if 13/6 is what GNU says is the best move. I've noticed it actually favors leaving a shot on the midpoint rather than to run out of the backfield when far behind. If white rolls a 1 it might be better to get hit on the midpoint than get hit 3/2*

Quote

01-04-2015 , 12:25 PM

#57

yogiman

adept

Join Date: Jun 2012 Posts: 784

White - Pips 162

Black - Pips 143
Black to Play 5-5
Created with www.BGdiagram.com

Quote

01-05-2015 , 08:14 AM

#58

yogiman

adept

Join Date: Jun 2012 Posts: 784

Oops, I forgot the code. It is not so complicated, though, but easy to overlook. Black is very well ahead and needs to move from the 18-point. Unless he is lucky, he has to dump checkers behind white's anchor, but the blot on the 1-point is in his way, so he likes to move it further up.

Code:

1. 18/13(2) 6/1*(2)             Eq.:  +0,662
       0,680 0,115 0,002 - 0,320 0,063 0,001 CL  +0,413 CF  +0,662
      
    2. 8/3(2) 6/1*(2)               Eq.:  +0,435 ( -0,227)
       0,610 0,156 0,003 - 0,390 0,093 0,002 CL  +0,284 CF  +0,435
      
    3. 18/13(2) 8/3(2)              Eq.:  +0,337 ( -0,325)
       0,595 0,121 0,004 - 0,405 0,107 0,002 CL  +0,206 CF  +0,337
     
    4. 18/8(2)                      Eq.:  +0,295 ( -0,367)
       0,591 0,103 0,002 - 0,409 0,111 0,002 CL  +0,175 CF  +0,295
      
        Truncated cubeful rollout (depth 10) with var.redn.
        147 games, Mersenne Twister dice gen. with seed 763581857 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]

Quote

01-05-2015 , 08:18 AM

#59

yogiman

adept

Join Date: Jun 2012 Posts: 784

Problem 15

Position ID: 7rMNACDW7lAAMA Match ID: UQkOAAAAAAAE
(So copy the above line, go to the Gnu-screen, and do ctrl+v)

White - Pips 81

Black - Pips 117
Black to Play 4-3
Created with www.BGdiagram.com

Quote

01-08-2015 , 06:29 AM

#60

yogiman

adept

Join Date: Jun 2012 Posts: 784

Code:

    1. 11/7 10/7                    Eq.:  +0,660 
       0,696 0,151 0,005 - 0,304 0,084 0,005 CL  +0,459 CF  +0,660 
     
    2. 11/7 6/3                     Eq.:  +0,543 ( -0,117) 
       0,651 0,099 0,003 - 0,349 0,096 0,008 CL  +0,301 CF  +0,543 
      
    3. 10/3                         Eq.:  +0,500 ( -0,160) 
       0,615 0,091 0,003 - 0,385 0,060 0,003 CL  +0,262 CF  +0,500 
     
    4. 6/3 5/1*                     Eq.:  +0,481 ( -0,179) 
       0,621 0,116 0,005 - 0,379 0,105 0,008 CL  +0,249 CF  +0,481 
       
        Truncated cubeful rollout (depth 10) with var.redn. 
        145 games, Mersenne Twister dice gen. with seed 763581857 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]

6/3 5/1* looks attractive to the blind eye. But if black doesn't throw a 3 as a first step to escape in the next few turns, his board will crumble. White should be allowed to throw, so that those back checkers will get some more space.

10/3 at least keeps white throwing, though 11/7 6/3 is better, because if white throws that 6, he has to send the checker to the bar, and black likely doesn't have to break up his board. If black throws a 3 next turn he will complete his prime instead of moving his back checker, and is favorite to get another white checker.

11/7 10/7 is best because if white hits he needs to throw 25 or 26, otherwise in most cases he will leave a blot + space at the other side of the board. A bonus is that black has now two numbers to complete his prime.

Quote

01-08-2015 , 03:50 PM

#61

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

This has been a great series of problems!

Let's get a wider variety of responses, please. It can't be just the Yogiman and me. Besides, I keep getting them wrong.

Mike

Quote

01-09-2015 , 05:36 AM

#62

yogiman

adept

Join Date: Jun 2012 Posts: 784

Thanks, Mike, makes me almost blush. And you know, I have not created a database, but get them on a day-to-day basis, so those positions are no rarities. Playing a quick 0-ply I make a selection of the ones that make me think or stumble. So don't expect every time a great problem like the last one.

The problems that have a lot of QF are also interesting, because what counts is the analysis. And ofcourse, you can try and play them.

Position ID: l9nAHAC2bYIBAw Match ID: cAkPAAAAAAAE

White - Pips 110

Black - Pips 114
Black to Play 6-3
Created with www.BGdiagram.com

Quote

01-09-2015 , 07:28 AM

#63

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

In this position, White would love for Black to disengage. That way, the game could be decided by a race, the only measure whereby White is competitive with Black. White also has an inside blot at the moment, something Black should take note of.

The only play I am sure is wrong is the board-busting safe play, 8/2 4/1. Black's board is his winning ticket, and he should not let go of it for any reason.

I am torn between running one checker off the anchor, a common tactic when the action is in the outfield, and playing one checker down from midpoint. The blots don't scare me either way.

I guess the best way to keep contact with White's checkers in the outfield is to retain the midpoint. That leaves breaking the anchor. There are three ways to to do that:

20/17, 8/2
20/14, 8/5
20/11

Of the three I like 20/11. It is the safest, and if left unmolested, Black could easily safety the checker. He would have 17 pips he could play before needing to dismantle any other part of his structure.

Mike

Quote

01-09-2015 , 06:02 PM

#64

yogiman

adept

Join Date: Jun 2012 Posts: 784

Code:

1. 20/11                        Eq.:  +0,376
       0,587 0,136 0,002 - 0,413 0,036 0,000 CL  +0,276 CF  +0,376
   
2. 13/7 8/5                     Eq.:  +0,255 ( -0,120)
       0,559 0,123 0,003 - 0,441 0,026 0,000 CL  +0,217 CF  +0,255
      [0,003 0,004 0,000 - 0,003 0,002 0,000 CL   0,007 CF   0,014]
        Truncated cubeful rollout (depth 10) with var.redn.
        146 games, Mersenne Twister dice gen. with seed 765006646 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]

Quote

01-09-2015 , 06:07 PM

#65

yogiman

adept

Join Date: Jun 2012 Posts: 784

Position ID: bzsHAAKY3VEBAw Match ID: cIkSAAAAAAAE

White - Pips 64

Black - Pips 141
Black to Play 5-4
Created with www.BGdiagram.com

Quote

01-10-2015 , 12:46 AM

#66

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

Assuming Jacoby on, I think I like 20/11, blocking 6s and keeping a few blots scattered around to patrol against White's escape. If we get hit, we get hit, but we are way behind anyway.

I also briefly thought about the fishy-looking 11/6 12/8 to secure a few blots and hope that we can still pick up White if he jumps over our modest blockade. I know we're miles behind in the race, but we're still against a 5-pt board so getting hit now would really hurt our winning chances.

Quote

01-11-2015 , 05:42 AM

#67

yogiman

adept

Join Date: Jun 2012 Posts: 784

Well, this one is to get you on a wrong QF-track.

No, I will be honest, 0-ply can be f*cking terrible. Playing it safe is way way better than moving 20/11. A 11, 12, 13, 14, 15 for white is the pathway for black to disaster, and 16 will also be no success-story in most cases. And on top of that there are the indirect hits of the other two black blots.

Quote

01-11-2015 , 06:28 AM

#68

bleep69

grinder

Join Date: Sep 2008 Posts: 483

20/11 was my move as well. Afraid to say I didn't even look at the other option.

Quote

01-11-2015 , 06:32 AM

#69

yogiman

adept

Join Date: Jun 2012 Posts: 784

Position ID: 2HbgASGY22ABJg Match ID: cAkXAAAAAAAE

White - Pips 147

Black - Pips 164
Black to Play 6-5
Created with www.BGdiagram.com

Quote

01-11-2015 , 06:46 AM

#70

bleep69

grinder

Join Date: Sep 2008 Posts: 483

I hit and cover on the 1 point. Nothing else looks appealing.

Quote

01-11-2015 , 10:45 AM

#71

Robertie

old hand

Join Date: May 2004 Posts: 1,954

Quote:

Originally Posted by bleep69

I hit and cover on the 1 point. Nothing else looks appealing.

I think that's supposed to be the 'obvious' play, although this is clearly a difficult position.

Black is getting squeezed and running out of time, so I think he has to make his move now. I'd try 22/16 14/9, with massive duplication of 4s. If Black gets away with this, he should be able to put his game back together on the next turn. And even if he's hit, he's 75% to reenter.

Anchor-breaking positions are inherently very tough. You don't want to break too soon when there are still constructive and less risky alternatives. If you wait too long, however, it's too risky amd you just need to sit on the position and hope for something good. You're looking for a sweet spot in the middle where your game really improves if you get away with the play, while the downside when you get whacked is still manageable. This position looks like a good candidate for anchor-breaking.

Quote

01-11-2015 , 10:52 AM

#72

Taper_Mike

adept

Join Date: Sep 2009 Posts: 1,030

Quote:

Originally Posted by bleep69

I hit and cover on the 1 point. Nothing else looks appealing.

I am completely clueless.

If you were captain of my chouette team, I would support your choice. It dupes 2, and leaves all 7s plus 21, 22, and 32 for White to make return hits. That makes 11 return shots. Hitting loose on the 1pt, 14/8, 6/1* is another possibility. That would also leave 11 returns. Furthermore, it would risk only one checker, and a let you keep the block. What is does not do, however, is give 9 dances. A third candidate is to slot the 3pt, risking 12 shots.

Late addition: I did not see Robertie's post until now. I think maybe he hit it on the nose.

Mike

Quote

01-11-2015 , 07:17 PM

#73

uberkuber

veteran

Join Date: Jul 2009 Posts: 3,032

Quote:

Originally Posted by Robertie

He might not have time to get away. Wouldn't that move (22/16 14/9) lead to a double? Is it an easy take or not so easy (assuming Jacoby)?

Quote

01-12-2015 , 06:14 AM

#74

yogiman

adept

Join Date: Jun 2012 Posts: 784

You didn't peak, well done!

Yes, all moves lead to a double, take.
White has only one direct shot at the blot on the 22-point. The bot shows that moving 18 to 19 will result in a close decision with 14/3. And removing the duplication by moving 24 to 23 also will result in a close decision. Those two factors make the leap to freedom very favorite.

Code:

    1. 22/16 14/9                   Eq.:  -0,568
       0,384 0,103 0,005 - 0,616 0,273 0,027 CL  -0,424 CF  -0,568

    2. 22/17 14/8                   Eq.:  -0,572  (-0,004)
       0,372 0,091 0,004 - 0,628 0,244 0,015 CL  -0,420 CF  -0,572 
      
    3. 14/3                         Eq.:  -0,708 ( -0,140) 
       0,342 0,080 0,003 - 0,658 0,215 0,012 CL  -0,462 CF  -0,708 
      
    4. 7/1* 6/1                     Eq.:  -0,720 ( -0,152)
       0,326 0,110 0,002 - 0,674 0,229 0,013 CL  -0,478 CF  -0,720
     
        Truncated cubeful rollout (depth 10) with var.redn. 
        147 games, Mersenne Twister dice gen. with seed 765220346 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]