Two Plus Two Forums
Sports and Games
Chess and Other Board Games
Backgammon Forum hosted by Bill Robertie.
Open your mind
03-30-2015
, 03:37 PM
03-30-2015
, 03:49 PM
White - Pips 120. Match Score 4/11
Black - Pips 182. Match Score 1/11
Black to Play 4-2
This one is from a 1st round match in the Ohio State Backgammon Championships. Two questions:
Black - Pips 182. Match Score 1/11
Black to Play 4-2
This one is from a 1st round match in the Ohio State Backgammon Championships. Two questions:
- What would you play at the score given?
- What would you play in an unlimited game (no Jacoby, no beavers; might or might not be different)?
04-01-2015
, 11:25 AM
White - Pips 120. Match Score 4/11
Black - Pips 182. Match Score 1/11
Position 10. Black to Play 4-2
Kudos to Backgammon Giant Ray Fogerlund who got this right under the glare of the lights. You can see his play at the 21-minute mark of the YouTube video. Holding onto the rear anchor is essential in this position. Yogiman’s analysis is spot on. Ray wants to run his spare and keep all three anchors. When the time comes, he usually wants to break his front anchor first.
The QF was so obvious that I tried to deflect some it by posing the position at two different scores. The correct play in both is the same: 20/18 6/2.
Along with almost everyone else in the room, I was all for switching to the 18pt. Here is the rollout for an unlimited game:
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
In the original position, holding the 22pt is critical because it is exactly 6 pips away from a point owned by the opponent. By moving the opponent back 1 pip, we create a position where switching to the 18pt is best. The 22pt is priming nothing, so it is okay to ditch it. The checker left behind has a clear path to escape, so Black may still hope to do so while retaining his three anchors. Not surprisingly, it is now a quadruple whopper to break the 6pt.
White - Pips 124
Black - Pips 182
Position 10a. Black to Play 4-2
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
Mike
Black - Pips 182. Match Score 1/11
Position 10. Black to Play 4-2
Kudos to Backgammon Giant Ray Fogerlund who got this right under the glare of the lights. You can see his play at the 21-minute mark of the YouTube video. Holding onto the rear anchor is essential in this position. Yogiman’s analysis is spot on. Ray wants to run his spare and keep all three anchors. When the time comes, he usually wants to break his front anchor first.
The QF was so obvious that I tried to deflect some it by posing the position at two different scores. The correct play in both is the same: 20/18 6/2.
Along with almost everyone else in the room, I was all for switching to the 18pt. Here is the rollout for an unlimited game:
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
Code:
XGID=---BBBB-------b-dd-eCBB---:0:0:1:42:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
Pip count X: 182 O: 120 X-O: 0-0
Cube: 1
X to play 42
1. Rollout¹ 20/18 6/2 eq:-0.3756
Player: 43.02% (G:9.15% B:0.33%)
Opponent: 56.98% (G:21.77% B:0.56%)
Confidence: ±0.0096 (-0.3852..-0.3660) - [100.0%]
2. Rollout¹ 22/18 20/18 eq:-0.4403 (-0.0647)
Player: 40.60% (G:7.89% B:0.26%)
Opponent: 59.40% (G:18.61% B:0.22%)
Confidence: ±0.0084 (-0.4487..-0.4318) - [0.0%]
¹ 5184 Games rolled with Variance Reduction.
Dice Seed: 59279099
Moves: 3-ply, cube decisions: XG Roller
Rollout by Taper_Mike
2015-Mar-31
eXtreme Gammon Version: 2.10.199.2658
Flashcard PositionID 000997.xgp
White - Pips 124
Black - Pips 182
Position 10a. Black to Play 4-2
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
Code:
XGID=---BBBB-------bd-d-eCBB---:0:0:1:42:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
Pip count X: 182 O: 124 X-O: 0-0
Cube: 1
X to play 42
1. Rollout¹ 22/18 20/18 eq:-0.2583
Player: 44.21% (G:9.42% B:0.33%)
Opponent: 55.79% (G:15.10% B:0.24%)
Confidence: ±0.0079 (-0.2662..-0.2504) - [100.0%]
2. Rollout¹ 22/16 eq:-0.3333 (-0.0750)
Player: 42.53% (G:8.58% B:0.30%)
Opponent: 57.47% (G:15.86% B:0.23%)
Confidence: ±0.0077 (-0.3410..-0.3256) - [0.0%]
3. Rollout¹ 22/20 6/2 eq:-0.6896 (-0.4313)
Player: 34.84% (G:6.35% B:0.22%)
Opponent: 65.16% (G:23.50% B:0.55%)
Confidence: ±0.0120 (-0.7016..-0.6776) - [0.0%]
¹ 5184 Games rolled with Variance Reduction.
Dice Seed: 59279099
Moves: 3-ply, cube decisions: XG Roller
Rollout by Taper_Mike
2015-Mar-31
eXtreme Gammon Version: 2.10.199.2658
Flashcard PositionID 000997a.xgp
04-02-2015
, 12:19 PM
This post corrects the second rollout above. The earlier rollout omitted 20/18 6/2. With the correction, breaking the 6pt is now a seventh-best “whopper with cheese” rather than a quadruple whopper.
Thanks to Yogiman for alerting me to the omission.
White - Pips 124
Black - Pips 182
Position 10a. Black to Play 4-2
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
Mike
Thanks to Yogiman for alerting me to the omission.
White - Pips 124
Black - Pips 182
Position 10a. Black to Play 4-2
5k XG2 Rollout – Unlimited game, no Jacoby, no beavers
Code:
XGID=---BBBB-------bd-d-eCBB---:0:0:1:42:0:0:0:0:10
X:Player 1 O:Player 2
Score is X:0 O:0. Unlimited Game
Pip count X: 182 O: 124 X-O: 0-0
Cube: 1
X to play 42
1. Rollout¹ 22/18 20/18 eq:-0.2583
Player: 44.21% (G:9.42% B:0.33%)
Opponent: 55.79% (G:15.10% B:0.24%)
Confidence: ±0.0079 (-0.2662..-0.2504) - [100.0%]
2. Rollout¹ 22/16 eq:-0.3333 (-0.0750)
Player: 42.53% (G:8.58% B:0.30%)
Opponent: 57.47% (G:15.86% B:0.23%)
Confidence: ±0.0077 (-0.3410..-0.3256) - [0.0%]
3. Rollout¹ 20/16 4/2 eq:-0.3567 (-0.0983)
Player: 42.85% (G:9.44% B:0.32%)
Opponent: 57.15% (G:18.88% B:0.47%)
Confidence: ±0.0085 (-0.3652..-0.3481) - [0.0%]
4. Rollout¹ 20/18 20/16 eq:-0.3678 (-0.1095)
Player: 43.32% (G:8.08% B:0.29%)
Opponent: 56.68% (G:20.34% B:0.56%)
Confidence: ±0.0078 (-0.3756..-0.3600) - [0.0%]
5. Rollout¹ 20/16 6/4 eq:-0.3774 (-0.1191)
Player: 42.22% (G:9.18% B:0.32%)
Opponent: 57.78% (G:18.83% B:0.47%)
Confidence: ±0.0082 (-0.3856..-0.3692) - [0.0%]
6. Rollout¹ 20/16 5/3 eq:-0.4035 (-0.1452)
Player: 41.46% (G:8.93% B:0.30%)
Opponent: 58.54% (G:18.70% B:0.44%)
Confidence: ±0.0082 (-0.4118..-0.3953) - [0.0%]
7. Rollout¹ 20/18 6/2 eq:-0.4177 (-0.1594)
Player: 41.68% (G:9.02% B:0.32%)
Opponent: 58.32% (G:21.17% B:0.54%)
Confidence: ±0.0093 (-0.4270..-0.4085) - [0.0%]
¹ 5184 Games rolled with Variance Reduction.
Dice Seed: 59279099
Moves: 3-ply, cube decisions: XG Roller
Rollout by Taper_Mike
2015-Apr-01
eXtreme Gammon Version: 2.10.199.2658
Flashcard PositionID 000997a.xgp
04-02-2015
, 03:59 PM
04-03-2015
, 04:07 AM
Position ID: mM/CATDguwvBAA Match ID: cIkJAAAAAAAE
04-03-2015
, 04:42 AM
04-03-2015
, 01:29 PM
White - Pips 150
Black - Pips 133
Black - Pips 133
Code:
1. 8/5 7/5 Eq.: +0,788
0,684 0,210 0,012 - 0,316 0,081 0,003 CL +0,506 CF +0,788
2. 7/4 6/4 Eq.: +0,763 ( -0,025)
0,676 0,208 0,011 - 0,324 0,084 0,003 CL +0,485 CF +0,763
3. 13/8 Eq.: +0,515 ( -0,273)
0,626 0,166 0,007 - 0,374 0,081 0,002 CL +0,341 CF +0,515
4. 18/16*/13 Eq.: +0,492 ( -0,296)
0,618 0,180 0,014 - 0,382 0,109 0,004 CL +0,317 CF +0,492
White - Pips 136
Black - Pips 133
Black - Pips 133
Code:
1. 8/5 7/5 Eq.: +0,717
0,675 0,217 0,012 - 0,325 0,111 0,003 CL +0,466 CF +0,717
2. 18/16*/13 Eq.: +0,346 ( -0,371)
0,603 0,155 0,010 - 0,397 0,160 0,003 CL +0,209 CF +0,346
White - Pips 150
Black - Pips 145
Black - Pips 145
Code:
1. 8/5 7/5 Eq.: +0,754
0,680 0,209 0,010 - 0,320 0,091 0,003 CL +0,484 CF +0,754
2. 18/16*/13 Eq.: +0,696 ( -0,057)
0,667 0,185 0,011 - 0,333 0,078 0,002 CL +0,449 CF +0,696
So it is all about creating a prime, and preventing white from establishing a good defensive point:
White - Pips 146
Black - Pips 134
Black - Pips 134
Code:
1. 8/5 7/5 Eq.: +0,458
0,624 0,120 0,004 - 0,376 0,076 0,003 CL +0,295 CF +0,458
2. 18/16*/13 Eq.: +0,396 ( -0,062)
0,608 0,137 0,006 - 0,392 0,108 0,003 CL +0,247 CF +0,396
04-04-2015
, 11:29 AM
Position ID: mGvkgBHgefAAUg Match ID: cIkGAAAAAAAE
04-04-2015
, 06:48 PM
04-05-2015
, 01:42 PM
Also after a lot of consideration, I myself would have played bar/24.
Bringing an extra white checker in the outfield reduces the relative value of bar/20 24/23:
Placing the checker on the 15-point will result in a pass for black. So as expected black can play bar/20 because white has not sufficient ammunition. But this cannot satisfy our reasoning faculty.
Let's look what can happen in case of bar/20, if white
a)doesn't hit (6%)
Clear
b) makes the 22-point (6%)
Unlike bar/24, black will be in the opportunity to make the 20-point
c) double hits (43%)
Now, unless black throws 66,55,33, he is not in grave danger with white's 2-board. For instance, if white thows 51, he will win 0,065 compared to conventional play:
As compensation for the equity loss, it gives black the opportunity to hit back, and/or slot the 20-point for a second time.
d) makes the 20-point on black's head (45%)
This can be really bad, because if black doesn't enter white goes into blitz mode, but if he has not too much bad luck he will make a defensive point. High numbers will not enter, and the positive thing about this is that it supports his timing. I will go into detail.
Reducing the pipdifference by moving a checker from 13 to 6 increases the relative value of bar/20 24/23:
So increasing the pipdifference makes the 20-point less golden. The reason is that the lower the defensive point, the better the timing must be. Compare the pipdifferences of the next two positions:
Another thing is that white often will leave a blot on 17w and/or 18w, which gives black the opportunity to hit back from the bar. Though this is also true for bar/24, I have not tried whether the latter is less inclined to.
So in short bar/20 is better than bar/24, because it gives in say half of the cases a good chance to make the 20-point, and in the other almost half of the cases that white makes a point on black's head, it is not so bad as it looks, because it might give black a better timing.
White - Pips 162
Black - Pips 179
Black - Pips 179
Code:
1. bar/20 24/23 Eq.: -0,653
0,362 0,074 0,003 - 0,638 0,236 0,014 CL -0,451 CF -0,653
2. bar/24 13/8 Eq.: -0,749 ( -0,096)
0,339 0,067 0,002 - 0,661 0,222 0,016 CL -0,490 CF -0,749
Bringing an extra white checker in the outfield reduces the relative value of bar/20 24/23:
White - Pips 150
Black - Pips 167
Black to Play 5-1
Black - Pips 167
Black to Play 5-1
Code:
1. bar/20 24/23 Eq.: -0,503
0,295 0,057 0,002 - 0,705 0,239 0,014 CL -0,605 CF -0,503
2. bar/24 8/3 Eq.: -0,542 ( -0,039)
0,275 0,050 0,002 - 0,725 0,222 0,016 CL -0,637 CF -0,542
Let's look what can happen in case of bar/20, if white
a)doesn't hit (6%)
Clear
b) makes the 22-point (6%)
Unlike bar/24, black will be in the opportunity to make the 20-point
c) double hits (43%)
Now, unless black throws 66,55,33, he is not in grave danger with white's 2-board. For instance, if white thows 51, he will win 0,065 compared to conventional play:
Code:
1. Cubeful 2-ply 8/3* 6/5* Eq.: +0,182
0,593 0,223 0,018 - 0,407 0,102 0,005
2-ply cubeful prune [world class]
4. Cubeful 2-ply 13/7 Eq.: +0,118 ( -0,065)
0,586 0,159 0,008 - 0,414 0,078 0,002
2-ply cubeful prune [world class]
d) makes the 20-point on black's head (45%)
This can be really bad, because if black doesn't enter white goes into blitz mode, but if he has not too much bad luck he will make a defensive point. High numbers will not enter, and the positive thing about this is that it supports his timing. I will go into detail.
Reducing the pipdifference by moving a checker from 13 to 6 increases the relative value of bar/20 24/23:
White - Pips 162
Black - Pips 172
Black - Pips 172
Code:
1. bar/20 24/23 Eq.: -0,679
0,353 0,081 0,004 - 0,647 0,237 0,014 CL -0,461 CF -0,679
2. bar/24 13/8 Eq.: -0,818 ( -0,139)
0,324 0,066 0,002 - 0,676 0,219 0,016 CL -0,519 CF -0,818
White - Pips 102
Black - Pips 40
White - Pips 78
Black - Pips 44
Moving 2 checkers from 14 to 20 decreases white's pipcount with 2 times 6, and 15 to 21 likewise. On the other hand, moving black's defensive point from 20 to 22 adds 4, so the total difference has decreased with 4x6 +4 = 28. Ofcourse, maybe black escapes with a high double beforehand, white's defensive itself gets into trouble, or some other scenarios, but in the majority of the cases black needs a reasonable pipdistance. Black - Pips 40
White - Pips 78
Black - Pips 44
Another thing is that white often will leave a blot on 17w and/or 18w, which gives black the opportunity to hit back from the bar. Though this is also true for bar/24, I have not tried whether the latter is less inclined to.
So in short bar/20 is better than bar/24, because it gives in say half of the cases a good chance to make the 20-point, and in the other almost half of the cases that white makes a point on black's head, it is not so bad as it looks, because it might give black a better timing.