Poll
Question: Double?
Option 1: Double
votes: 5
Option 2: Play on
votes: 3
Title: Game 1, move 6: Herd to Cube?
Post by: diane on March 25, 2012, 05:32:00 AM
Post by: diane on March 25, 2012, 05:32:00 AM
Time for another look at this cube.
Title: Re: Game 1, move 6: Herd to Cube?
Post by: Julia_H on March 25, 2012, 09:09:49 AM
Post by: Julia_H on March 25, 2012, 09:09:49 AM
Spoiler
I think bar/24 13/8 might have been worth a look, just to keep a spare back on the mid?
Anyway, we will have at least a four-point board with any roll, we have the anchor, we have men to carry on the attack, and there's still a blot on the other side... and we are now level in the race. At worst we end up in a tenable holding game with White awkwardly stripped and stacked. Double.
Anyway, we will have at least a four-point board with any roll, we have the anchor, we have men to carry on the attack, and there's still a blot on the other side... and we are now level in the race. At worst we end up in a tenable holding game with White awkwardly stripped and stacked. Double.
[close]
Title: Re: Game 1, move 6: Herd to Cube?
Post by: ah_clem on March 25, 2012, 01:53:25 PM
Post by: ah_clem on March 25, 2012, 01:53:25 PM
Yesterday, I was inclined towards pushing the cube here. Today, I'm not so sure.
A good strategy in short matches is to get some gammon chances and turn the cube. We've got significant chances here. OTOH, the take seems obvious. Examining PRAT, we have an advantage in position, but are behind in the race and the threats look about even. That says hold.
OTOOH, We've got a fair amount of market losers, and our opponent has terrible structure with those ugly stacks and a stripped midpoint. Finally, I've just run through all 36 numbers and every shake covers at least one of our homeboard blots, so this position is stronger than it looks.
I'm going to say ship it. Every roll gives us a four point board and limits his return shots to no more than 1 in 3, and most of those fail to cover his own blot. PRAT says hold, but there are exceptions and I think this is one.
A good strategy in short matches is to get some gammon chances and turn the cube. We've got significant chances here. OTOH, the take seems obvious. Examining PRAT, we have an advantage in position, but are behind in the race and the threats look about even. That says hold.
OTOOH, We've got a fair amount of market losers, and our opponent has terrible structure with those ugly stacks and a stripped midpoint. Finally, I've just run through all 36 numbers and every shake covers at least one of our homeboard blots, so this position is stronger than it looks.
I'm going to say ship it. Every roll gives us a four point board and limits his return shots to no more than 1 in 3, and most of those fail to cover his own blot. PRAT says hold, but there are exceptions and I think this is one.
Title: Re: Game 1, move 6: Herd to Cube?
Post by: diane on March 26, 2012, 09:48:25 AM
Post by: diane on March 26, 2012, 09:48:25 AM
hmmm..should have moved this this morning when it was decided..now it is tied..and I am not sure I want to break it!
I will extend voting period for 12 hours to allow some discussion and rethinking...
I will extend voting period for 12 hours to allow some discussion and rethinking...
Title: Re: Game 1, move 6: Herd to Cube? ROLLOUT
Post by: ah_clem on March 28, 2012, 02:28:14 PM
Post by: ah_clem on March 28, 2012, 02:28:14 PM
rollout
A fairly close one - holding the cube here is only a .028 error, giving up .22% match winning chances. Massive take, giving up 2.3% MWC.
Well done on both sides.
Cube analysis
Rollout cubeless equity +0.474 (Money: +0.449)
Cubeful equities:
1. Double, take +0.701
2. Double, pass +1.000 ( +0.299)
3. No double +0.672 ( -0.028)
Proper cube action: Double, take
Rollout details:
Centered 1-cube:
0.650 0.213 0.004 - 0.350 0.066 0.002 CL +0.474 CF +0.672
[0.001 0.002 0.000 - 0.001 0.001 0.000 CL 0.004 CF 0.008]
Player O owns 2-cube:
0.657 0.217 0.004 - 0.343 0.065 0.003 CL +1.008 CF +0.701
[0.001 0.003 0.000 - 0.001 0.001 0.000 CL 0.008 CF 0.011]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 910687781 and quasi-random dice
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]
Spoiler
A fairly close one - holding the cube here is only a .028 error, giving up .22% match winning chances. Massive take, giving up 2.3% MWC.
Well done on both sides.
Cube analysis
Rollout cubeless equity +0.474 (Money: +0.449)
Cubeful equities:
1. Double, take +0.701
2. Double, pass +1.000 ( +0.299)
3. No double +0.672 ( -0.028)
Proper cube action: Double, take
Rollout details:
Centered 1-cube:
0.650 0.213 0.004 - 0.350 0.066 0.002 CL +0.474 CF +0.672
[0.001 0.002 0.000 - 0.001 0.001 0.000 CL 0.004 CF 0.008]
Player O owns 2-cube:
0.657 0.217 0.004 - 0.343 0.065 0.003 CL +1.008 CF +0.701
[0.001 0.003 0.000 - 0.001 0.001 0.000 CL 0.008 CF 0.011]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 910687781 and quasi-random dice
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]
[close]
Title: Re: Game 1, move 6: Herd to Cube?
Post by: ah_clem on April 12, 2012, 02:39:58 PM
Post by: ah_clem on April 12, 2012, 02:39:58 PM
revised rollout
UPDATE: The previous rollout actually was correct, but there's a glitch in the sgf file. The rollout below is for the following position
GNU Backgammon Position ID: LM+DA0Bqm4kBAw
Match ID : QYm2AAAAAAAE
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: Stiefnu (Cube: 2)
| X O | O | O X O O | 0 points
| X O | | O X O |
| O | | O |
| O | | O |
| | | |
v| |BAR| | 5 point match
| | | |
| | | |
| O | | |
| O X | | X X X | Rolled 55
| O X X | | X X X X X | 0 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: Herd
Note that the checker is still on the bar. Not wanting to waste a perfectly good rollout, I've left the post below intact, but feel free to ignore it as it's irrelevant to this match.
Now that I have the sgf file I see that I screwed up entering the position when I rolled this one out. I placed steifnu's runner on the ace point instead of on the bar. One might not think that it makes that much difference, but it's actually quite huge turning a close cube/easy take into a big cube/massive pass.
Apologies to the Herd, and especially to steifnu for bringing the bad news.
Cube analysis
Rollout cubeless equity +0.707 (Money: +0.704)
Cubeful equities:
1. Double, pass +1.000
2. Double, take +1.334 ( +0.334)
3. No double +0.896 ( -0.104)
Proper cube action: Double, pass
Rollout details:
Centered 1-cube:
0.709 0.337 0.005 - 0.291 0.055 0.002 CL +0.707 CF +0.896
[0.001 0.002 0.000 - 0.001 0.001 0.000 CL 0.004 CF 0.008]
Player Stiefnu owns 2-cube:
0.721 0.345 0.004 - 0.279 0.054 0.003 CL +1.583 CF +1.334
[0.002 0.003 0.001 - 0.002 0.001 0.001 CL 0.009 CF 0.012]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 917086324 and quasi-random dice
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]
Spoiler
UPDATE: The previous rollout actually was correct, but there's a glitch in the sgf file. The rollout below is for the following position
GNU Backgammon Position ID: LM+DA0Bqm4kBAw
Match ID : QYm2AAAAAAAE
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: Stiefnu (Cube: 2)
| X O | O | O X O O | 0 points
| X O | | O X O |
| O | | O |
| O | | O |
| | | |
v| |BAR| | 5 point match
| | | |
| | | |
| O | | |
| O X | | X X X | Rolled 55
| O X X | | X X X X X | 0 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: Herd
Note that the checker is still on the bar. Not wanting to waste a perfectly good rollout, I've left the post below intact, but feel free to ignore it as it's irrelevant to this match.
Now that I have the sgf file I see that I screwed up entering the position when I rolled this one out. I placed steifnu's runner on the ace point instead of on the bar. One might not think that it makes that much difference, but it's actually quite huge turning a close cube/easy take into a big cube/massive pass.
Apologies to the Herd, and especially to steifnu for bringing the bad news.
Cube analysis
Rollout cubeless equity +0.707 (Money: +0.704)
Cubeful equities:
1. Double, pass +1.000
2. Double, take +1.334 ( +0.334)
3. No double +0.896 ( -0.104)
Proper cube action: Double, pass
Rollout details:
Centered 1-cube:
0.709 0.337 0.005 - 0.291 0.055 0.002 CL +0.707 CF +0.896
[0.001 0.002 0.000 - 0.001 0.001 0.000 CL 0.004 CF 0.008]
Player Stiefnu owns 2-cube:
0.721 0.345 0.004 - 0.279 0.054 0.003 CL +1.583 CF +1.334
[0.002 0.003 0.001 - 0.002 0.001 0.001 CL 0.009 CF 0.012]
Full cubeful rollout with var.redn.
1296 games, Mersenne Twister dice gen. with seed 917086324 and quasi-random dice
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]
[close]