Discussion:
Lots of tempting possibilities
Add Reply
Tim Chow
2020-07-17 03:41:54 UTC
Reply
Permalink
Unless you're Stick of course. Then you see only one play.

XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22

---
Tim Chow
Peter
2020-07-17 10:23:34 UTC
Reply
Permalink
Post by Tim Chow
Unless you're Stick of course. Then you see only one play.
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
---
Tim Chow
7/3*/1 23/21
Paul Epstein
2020-07-17 12:53:59 UTC
Reply
Permalink
Post by Tim Chow
Unless you're Stick of course. Then you see only one play.
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Actually, I only see two candidates rather than "lots".
However, that just reflects on my sub-expert backgammon
rather than a contradiction of Tim.

(It just occurred to me that Tim/ MIT (where Tim works/used-to-work/
is-an-alum/ or has some connection is a palindrome.)

I would either make my 9 point, or, if all I drink was black coffee
with no sugar, I could make the 4 point and play 23/21.
Would this latter play be as good an idea as subscribing to Harper's magazine?
No, it wouldn't.
Why do I not like the play?
Because it diversifies for Oppy -- six escapes and many non-sixes will
point on me.
I will play 13/9(2). It's of course extremely difficult to escape
a 1-away 5 prime, and we buy time to escape our rear checkers.

How confident am I?
Well, the probability of me being right is the reciprocal of Graham's number.
Why is it so low?
Two reasons -- there's clearly something wrong with my thinking because I
only see two possibilities. Also, Peter never posts the wrong answer.
With his accuracy being what it is, I feel extremely unlucky that Peter is
active on this bg newsgroup rather than the I-predict-the-lottery-numbers
newsgroups.

Paul
b***@gmail.com
2020-07-18 03:35:23 UTC
Reply
Permalink
Post by Tim Chow
Unless you're Stick of course. Then you see only one play.
It's funny cause it's true. (although I'm guessing XG only 'sees' one play also)

Stick
Tim Chow
2020-07-19 20:00:18 UTC
Reply
Permalink
Post by b***@gmail.com
It's funny cause it's true. (although I'm guessing XG only 'sees' one play also)
One definition of XG "seeing" a play is that the play makes it through
XG 3-ply's normal move filter and is subject to a full 3-ply evaluation.
According to this definition, XG "sees" three plays: 23/21 7/3*/1,
24/22 7/3*/1, and 13/9(2).

Two other plays get as far as being subject to a 2-ply evaluation:
7/5 7/3*/1 and 23/21 8/4 6/4.

---
Tim Chow
Tim Chow
2020-07-19 19:29:26 UTC
Reply
Permalink
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22

Attack or prime? XG attacks. But see also the variant.

1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]

2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]

3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]

4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]

5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.207.pre-release

-------
Variant
-------

XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22

1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]

2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]

3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.207.pre-release

---
Tim Chow
b***@gmail.com
2020-07-20 06:05:55 UTC
Reply
Permalink
Post by Tim Chow
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Attack or prime? XG attacks. But see also the variant.
1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]
2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]
3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]
4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]
5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
-------
Variant
-------
XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22
1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]
2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]
3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
---
Tim Chow
Your variant has a checker off the board. My definition for XG seeing a play varies but here it was meant as one playing being clearly superior to the others.

Stick
Paul Epstein
2020-07-20 08:23:05 UTC
Reply
Permalink
Post by b***@gmail.com
Post by Tim Chow
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Attack or prime? XG attacks. But see also the variant.
1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]
2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]
3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]
4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]
5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
-------
Variant
-------
XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22
1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]
2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]
3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
---
Tim Chow
Your variant has a checker off the board. My definition for XG seeing a play varies but here it was meant as one playing being clearly superior to the others.
Stick
But how is the clarity of the superiority measured?
By the equity difference? One good measure of how clear the
superiority is of the best play is whether any other choices
pass XG's filters. Hence the relevance of Tim "Meticulous Analyst" Chow's
comments.

Paul
b***@gmail.com
2020-07-20 08:55:19 UTC
Reply
Permalink
Post by Paul Epstein
Post by b***@gmail.com
Post by Tim Chow
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Attack or prime? XG attacks. But see also the variant.
1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]
2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]
3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]
4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]
5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
-------
Variant
-------
XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22
1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]
2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]
3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
---
Tim Chow
Your variant has a checker off the board. My definition for XG seeing a play varies but here it was meant as one playing being clearly superior to the others.
Stick
But how is the clarity of the superiority measured?
By the equity difference? One good measure of how clear the
superiority is of the best play is whether any other choices
pass XG's filters. Hence the relevance of Tim "Meticulous Analyst" Chow's
comments.
Paul
You're working on the assumption that (m)any know the actual requirements to make it through the move filter. I do. Probably Tim does. Even the average person posting on backgammon forums certainly does not. If three plays made it through the default filter here the other two plays _barely_ made it.

Stick
Paul Epstein
2020-07-20 17:00:03 UTC
Reply
Permalink
Post by b***@gmail.com
Post by Paul Epstein
Post by b***@gmail.com
Post by Tim Chow
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Attack or prime? XG attacks. But see also the variant.
1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]
2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]
3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]
4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]
5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
-------
Variant
-------
XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22
1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]
2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]
3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
---
Tim Chow
Your variant has a checker off the board. My definition for XG seeing a play varies but here it was meant as one playing being clearly superior to the others.
Stick
But how is the clarity of the superiority measured?
By the equity difference? One good measure of how clear the
superiority is of the best play is whether any other choices
pass XG's filters. Hence the relevance of Tim "Meticulous Analyst" Chow's
comments.
Paul
You're working on the assumption that (m)any know the actual requirements to make it through the move filter. I do. Probably Tim does. Even the average person posting on backgammon forums certainly does not. If three plays made it through the default filter here the other two plays _barely_ made it.
Stick
No, I'm not using that assumption.
By analogy, suppose there exists a recent "Encyclopedia of Great
Mathematicians" and that this book/website has a great reputation
among the cognoscenti and that Tim is a fan of this encyclopedia.

Then because I don't know many areas of mathematics (and who does?),
I might say "I define a great mathematician as someone with an entry
in The Encyclopedia of Great Mathematicians."
If I say that, that doesn't mean that I need to know the full details of
the inclusion criteria.

Paul
b***@gmail.com
2020-07-21 02:33:57 UTC
Reply
Permalink
Post by Paul Epstein
Post by b***@gmail.com
Post by Paul Epstein
Post by b***@gmail.com
Post by Tim Chow
XGID=-A-a-BCBC---cB---cbdb--AA-:0:0:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O X X |
| X O O | | O O |
| O | | O |
| | | O |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X |
| O X X | | X X |
| O X X | | X X O X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 133 X-O: 0-0
Cube: 1
X to play 22
Attack or prime? XG attacks. But see also the variant.
1. Rollout¹ 23/21 7/3* 3/1 eq:-0.323
Player: 40.16% (G:13.86% B:0.14%)
Opponent: 59.84% (G:14.59% B:0.83%)
Confidence: ±0.012 (-0.335..-0.311) - [100.0%]
2. Rollout¹ 24/22 7/3* 3/1 eq:-0.397 (-0.074)
Player: 38.35% (G:13.00% B:0.14%)
Opponent: 61.65% (G:14.31% B:0.80%)
Confidence: ±0.013 (-0.410..-0.384) - [0.0%]
3. Rollout¹ 13/9(2) eq:-0.425 (-0.102)
Player: 39.55% (G:11.57% B:0.32%)
Opponent: 60.45% (G:18.11% B:0.92%)
Confidence: ±0.012 (-0.438..-0.413) - [0.0%]
4. Rollout¹ 23/21 8/4 6/4 eq:-0.440 (-0.117)
Player: 38.52% (G:12.67% B:0.35%)
Opponent: 61.48% (G:18.30% B:1.12%)
Confidence: ±0.013 (-0.453..-0.427) - [0.0%]
5. Rollout¹ 7/5 7/3* 3/1 eq:-0.461 (-0.137)
Player: 36.72% (G:12.69% B:0.14%)
Opponent: 63.28% (G:15.03% B:0.90%)
Confidence: ±0.012 (-0.473..-0.449) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
-------
Variant
-------
XGID=-A-a-BCBC---bB--bcbbb--AA-:1:1:1:22:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O O | | O O X X |
| X O O O | | O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 126 X-O: 0-0
Cube: 2, X own cube
X to play 22
1. Rollout¹ 23/21 8/4 6/4 eq:-0.195
Player: 39.84% (G:0.00% B:0.00%)
Opponent: 60.16% (G:14.18% B:0.99%)
Confidence: ±0.012 (-0.207..-0.184) - [100.0%]
2. Rollout¹ 13/9(2) eq:-0.245 (-0.050)
Player: 37.91% (G:0.00% B:0.00%)
Opponent: 62.09% (G:15.24% B:0.87%)
Confidence: ±0.012 (-0.257..-0.233) - [0.0%]
3. Rollout¹ 23/21 7/3* 3/1 eq:-0.269 (-0.074)
Player: 35.79% (G:0.00% B:0.00%)
Opponent: 64.21% (G:11.41% B:0.65%)
Confidence: ±0.012 (-0.281..-0.257) - [0.0%]
¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller
eXtreme Gammon Version: 2.19.207.pre-release
---
Tim Chow
Your variant has a checker off the board. My definition for XG seeing a play varies but here it was meant as one playing being clearly superior to the others.
Stick
But how is the clarity of the superiority measured?
By the equity difference? One good measure of how clear the
superiority is of the best play is whether any other choices
pass XG's filters. Hence the relevance of Tim "Meticulous Analyst" Chow's
comments.
Paul
You're working on the assumption that (m)any know the actual requirements to make it through the move filter. I do. Probably Tim does. Even the average person posting on backgammon forums certainly does not. If three plays made it through the default filter here the other two plays _barely_ made it.
Stick
No, I'm not using that assumption.
By analogy, suppose there exists a recent "Encyclopedia of Great
Mathematicians" and that this book/website has a great reputation
among the cognoscenti and that Tim is a fan of this encyclopedia.
Then because I don't know many areas of mathematics (and who does?),
I might say "I define a great mathematician as someone with an entry
in The Encyclopedia of Great Mathematicians."
If I say that, that doesn't mean that I need to know the full details of
the inclusion criteria.
Paul
If you don't know how the move filter filters, it is near meaningless to know a move made it through the filter. Your example is biased. It says right in the title if you're in the book you're a great mathematician. Nowhere in XG's move filter is it even implied if a play makes it through the filter it should be accepted as a play one should have considered.

Stick
Tim Chow
2020-07-21 03:02:28 UTC
Reply
Permalink
Post by b***@gmail.com
You're working on the assumption that (m)any know the actual requirements to
make it through the move filter. I do. Probably Tim does.
In fact, I don't know exactly what XG does. I don't think I've seen it
documented precisely anywhere. GNU Backgammon is more transparent. If
you go to the "Move filter setup" then you can see that, for example,
a 2-ply analysis will analyze everything at 0-ply, then take up to 8
moves with 0.160 of the top, and subject these to 2-ply analysis.
Presumably XG does something similar, but I haven't seen it spelled out.

---
Tim Chow
b***@gmail.com
2020-07-21 09:25:57 UTC
Reply
Permalink
Post by Tim Chow
Post by b***@gmail.com
You're working on the assumption that (m)any know the actual requirements to
make it through the move filter. I do. Probably Tim does.
In fact, I don't know exactly what XG does. I don't think I've seen it
documented precisely anywhere. GNU Backgammon is more transparent. If
you go to the "Move filter setup" then you can see that, for example,
a 2-ply analysis will analyze everything at 0-ply, then take up to 8
moves with 0.160 of the top, and subject these to 2-ply analysis.
Presumably XG does something similar, but I haven't seen it spelled out.
---
Tim Chow
Options >> Settings >> Analysis >> Search Interval.

Stick
Tim Chow
2020-07-21 17:04:52 UTC
Reply
Permalink
Post by b***@gmail.com
Options >> Settings >> Analysis >> Search Interval.
Yes, I knew about that, but it doesn't (for example) explain how XG decides,
when doing a 3-ply evaluation, which plays to analyze at 2-ply only, and
which ones to subject to a full 3-ply analysis.

---
Tim Chow

Tim Chow
2020-07-21 02:42:02 UTC
Reply
Permalink
Post by b***@gmail.com
Your variant has a checker off the board.
Ha! Good catch. Let me give here the variant I was originally going to
post, before I found the "better" (but unfortunately incorrect!) variant.

XGID=-A-a-BCBC---cB---cbbbb-AA-:1:1:1:22:0:0:0:0:10

X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O O X X |
| X O O | | O O O |
| O | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| O X | | X | +---+
| O X X | | X X | | 2 |
| O X X | | X X O X | +---+
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 140 O: 129 X-O: 0-0
Cube: 2, X own cube
X to play 22

1. Rollout¹ 24/22 8/4 6/4 eq:-0.497
Player: 28.53% (G:7.56% B:0.22%)
Opponent: 71.47% (G:23.57% B:1.28%)
Confidence: ±0.009 (-0.506..-0.488) - [66.9%]

2. Rollout¹ 24/22 7/3* 3/1 eq:-0.499 (-0.003)
Player: 26.91% (G:7.61% B:0.09%)
Opponent: 73.09% (G:19.41% B:1.06%)
Confidence: ±0.007 (-0.507..-0.492) - [33.1%]

3. Rollout¹ 13/9(2) eq:-0.530 (-0.034)
Player: 26.90% (G:6.48% B:0.18%)
Opponent: 73.10% (G:21.75% B:1.21%)
Confidence: ±0.008 (-0.538..-0.522) - [0.0%]

¹ 1296 Games rolled with Variance Reduction.
Dice Seed: 271828
Moves: 3-ply, cube decisions: XG Roller

eXtreme Gammon Version: 2.19.207.pre-release

---
Tim Chow
Loading...