Two Plus Two Forums
Sports and Games
Chess and Other Board Games
Backgammon Forum hosted by Bill Robertie.
Should Black Hit?
12-14-2014
, 06:58 AM
Here's a situation I've been encountering seemingly more frequently lately. Should black play aggressively and hit, or more conservatively, considering the count is almost even?
White - Pips 94. Match Score 0/5
Black - Pips 93. Match Score 0/5
Black to Play 3-2
Created with www.BGdiagram.com
Black - Pips 93. Match Score 0/5
Black to Play 3-2
Created with www.BGdiagram.com
12-14-2014
, 07:56 AM
Thanks, Yogi. Yep, 6/3*/1 is the correct move, according to GNU. I was torn between making the 12-point, and hoping white doesn't escape (if he does, it's an even race), or hitting. What didn't occur to me was moving to the 1-point. The beauty of that play is now apparent to me. 
The other thing is that black should have doubled, with white taking.
The other thing is that black should have doubled, with white taking.
12-14-2014
, 03:52 PM
When the gap is high in your board, so that escape is easy (and the race is a toss-up after your move), hitting loose is usually better than passive play. Assuming you can safety your blot on the following turn, hitting makes you an instant 70%-to-30% favorite, even when your opponent has a closed board. Letting the opponent escape, of course, leaves the game as a 50%-50% proposition.
Paul Weaver posted a set of these at BgOnline this week. Below are two difficult variants that I constructed. The only difference between them is the location of White’s spares.
Mike
White - Pips 71
Black - Pips 71
Position 1. Black to Play 5-3
White - Pips 72
Black - Pips 71
Position 2. Black to Play 5-3
12-14-2014
, 09:40 PM
Also, moving down from the 3-point to the 1-point, if my math is correct, reduces your chances of getting hit from 36% to 31% - nothing to sneeze at.
I looked at your and Paul Weaver's examples - interesting how a few seemingly inconsequential pips can make a dramatic difference in how you should play a move. I also lol'd at Paul's introduction: I made the mistake of hitting (fourth best play), but Petko then made the much more costly mistake of fanning.
12-15-2014
, 06:26 PM
The main difference between these two positions is not the 1 pip in the race, but how 4-3 plays for White after the hit. In position 2, 4-3 essentially wins right away because White can cash next roll, or at least it's close (though I think it's just double/pass). In position 1, 4-3 just means White is even money or something. So essentially, you can think of the hit in position 2 as 11 shots for White, but the hit in position 1 is more like 10 shots. Not a huge deal, but it is relevant and it could swing the play if things are otherwise pretty close.
12-16-2014
, 04:51 AM
White - Pips 94. Match Score 0/5
Black - Pips 93. Match Score 0/5
Black to Play 3-2
Right. That is one of the reasons why I qualified my statement by prefacing it with the caveat, “When the gap is high in your board, ...” Here is the whole thing:
I believe it was Kent Goulding who coined the phrase, “Four in a row is more than you think.” That certainly is relevant here. This gap is not “high in the board,” and escape will be anything but “easy” when Black eschews the hit. The only reason hitting works here is because you get to continue on to the 1pt. If White hits on the 1pt, he will be trapped behind your prime, and if he does not hit, then he will either be on the bar or else right back where he started, and you will have gained half a roll in the exchange.
The other reason I think you have to be careful about applying my “rule” in this position is because of the gammon dangers represented by Black’s two outside blots. With a centered cube in a money game, where you have the protection of the Jacoby rule, they may not be a concern. In a match or some other circumstance where White might suddenly become too good to double after making a return hit, they are definitely a consideration. If this position does not qualify, imagine a similar position where White had a closed board.
Compared to hitting, making the 12pt is almost a 0.08 error. But that is true only if you continue on to the 1pt after hitting. For the reasons you cited (and those discussed above), making the 12pt actually comes out way ahead of any hitting play that does not continue on to the 1pt. The best of them is 15/13 6/3*, and it trails making the 12pt by more than 0.10!
Here is my 5k XG rollout. It uses my usual unlimited game, no-Jacoby, no-beavers settings. The resulting equities are similar to those generated in Midiboy’s GnuBg rollout.
Mike
Black - Pips 93. Match Score 0/5
Black to Play 3-2
When the gap is high in your board, so that escape is easy (and the race is a toss-up after your move), hitting loose is usually better than passive play. Assuming you can safety your blot on the following turn, hitting makes you an instant 70%-to-30% favorite, even when your opponent has a closed board. Letting the opponent escape, of course, leaves the game as a 50%-50% proposition.70%-30% is an overstatement. In practice, even when you can close out your opponent after he misses the return shot, he will usually still be winning something like 3%-5% of all games. It is probably more accurate, therefore, to say you become a 2-to-1 favorite when you hit. On the other hand, if you have cube access, and can force a pass, then you will have a one-roll position in which you win fully 70% of the games. Covering may never be necessary.
I believe it was Kent Goulding who coined the phrase, “Four in a row is more than you think.” That certainly is relevant here. This gap is not “high in the board,” and escape will be anything but “easy” when Black eschews the hit. The only reason hitting works here is because you get to continue on to the 1pt. If White hits on the 1pt, he will be trapped behind your prime, and if he does not hit, then he will either be on the bar or else right back where he started, and you will have gained half a roll in the exchange.
The other reason I think you have to be careful about applying my “rule” in this position is because of the gammon dangers represented by Black’s two outside blots. With a centered cube in a money game, where you have the protection of the Jacoby rule, they may not be a concern. In a match or some other circumstance where White might suddenly become too good to double after making a return hit, they are definitely a consideration. If this position does not qualify, imagine a similar position where White had a closed board.
Here is my 5k XG rollout. It uses my usual unlimited game, no-Jacoby, no-beavers settings. The resulting equities are similar to those generated in Midiboy’s GnuBg rollout.
Code:
XGID=--BaCBDB------AAabbbabb-b-:0:0:1:32:0:0:0:0:10
Score is X:0 O:0. Unlimited Game
Pip count X: 93 O: 94 X-O: 0-0
Cube: 1
X to play 32
1. Rollout¹ 6/3* 3/1 eq:+0.670
Player: 74.02% (G:9.98% B:0.23%)
Opponent: 25.98% (G:8.19% B:0.23%)
Confidence: ±0.005 (+0.665..+0.675) - [100.0%]
Duration: 9 minutes 27 seconds
2. Rollout¹ 6/1 eq:+0.644 (-0.026)
Player: 68.43% (G:5.34% B:0.13%)
Opponent: 31.57% (G:4.83% B:0.17%)
Confidence: ±0.005 (+0.639..+0.649) - [0.0%]
Duration: 6 minutes 02 seconds
3. Rollout¹ 15/12 6/4 eq:+0.603 (-0.067)
Player: 66.20% (G:5.44% B:0.17%)
Opponent: 33.80% (G:3.76% B:0.08%)
Confidence: ±0.004 (+0.599..+0.608) - [0.0%]
Duration: 5 minutes 13 seconds
4. Rollout¹ 14/11 6/4 eq:+0.600 (-0.070)
Player: 66.80% (G:4.75% B:0.16%)
Opponent: 33.20% (G:3.29% B:0.07%)
Confidence: ±0.004 (+0.596..+0.605) - [0.0%]
Duration: 6 minutes 26 seconds
5. Rollout¹ 15/12 14/12 eq:+0.592 (-0.078)
Player: 67.48% (G:2.39% B:0.04%)
Opponent: 32.52% (G:1.02% B:0.02%)
Confidence: ±0.005 (+0.588..+0.597) - [0.0%]
Duration: 4 minutes 55 seconds
6. Rollout¹ 15/13 14/11 eq:+0.569 (-0.101)
Player: 65.01% (G:5.62% B:0.17%)
Opponent: 34.99% (G:3.26% B:0.05%)
Confidence: ±0.004 (+0.565..+0.573) - [0.0%]
Duration: 4 minutes 22 seconds
7. Rollout¹ 15/10 eq:+0.566 (-0.104)
Player: 65.54% (G:5.41% B:0.17%)
Opponent: 34.46% (G:3.07% B:0.07%)
Confidence: ±0.004 (+0.562..+0.570) - [0.0%]
Duration: 4 minutes 29 seconds
8. Rollout¹ 15/13 6/3* eq:+0.489 (-0.181)
Player: 72.25% (G:12.93% B:0.34%)
Opponent: 27.75% (G:11.99% B:0.41%)
Confidence: ±0.003 (+0.485..+0.492) - [0.0%]
Duration: 10 minutes 08 seconds
¹ 5184 Games rolled with Variance Reduction.
Dice Seed: 14978697
Moves: 3-ply, cube decisions: XG Roller
Rollout by Taper_Mike
2014-Dec-14
eXtreme Gammon Version: 2.10.199.2658
Flashcard PositionID 000933.xgp
Last edited by Taper_Mike; 12-16-2014 at 05:08 AM.
12-16-2014
, 05:00 AM
12-16-2014
, 12:37 PM
You have the position set to a 5 point match which matters because it skews the take/pass decision after White turns the game around.
There's a way in Gnu to make it the default to start with a money game when you open it up, though off hand I don't remember exactly how.
There's a way in Gnu to make it the default to start with a money game when you open it up, though off hand I don't remember exactly how.
12-17-2014
, 03:12 AM
12-21-2014
, 05:10 PM
Quote:
Right on! I concluded the same thing in my analysis at BgOnline.
It is still amazing that the change causes almost a 0.08 swing between the two positions. In the first position, hitting wins by almost 0.04, while in the second, hitting loses by roughly the same amount.
Mike
It is still amazing that the change causes almost a 0.08 swing between the two positions. In the first position, hitting wins by almost 0.04, while in the second, hitting loses by roughly the same amount.
Mike
thats the sort of position that separates the men from the boys. i'm still in pre-school lol
thanks for all your great posts, i'll be back on the team this week-my computer got a mal-wear thing 8-10 days ago and i can't get onto bgo,
best
eric
12-22-2014
, 06:21 AM
White - Pips 89
Black - Pips 87
Position 1. Black to Play 4-1
White - Pips 78
Black - Pips 75
Position 2. Black to Play 3-1
What are your plays, and more importantly, why?
Both positions are unlimited games, with no Jacoby and no beavers.
Mike
Black - Pips 87
Position 1. Black to Play 4-1
White - Pips 78
Black - Pips 75
Position 2. Black to Play 3-1
What are your plays, and more importantly, why?
Both positions are unlimited games, with no Jacoby and no beavers.
Mike
12-24-2014
, 09:11 AM
This problem is of a different nature, but suits the topic.
White - Pips 97. Match Score 0/64
Black - Pips 118. Match Score 0/64
Black to Play 5-4
Created with www.BGdiagram.com
Black - Pips 118. Match Score 0/64
Black to Play 5-4
Created with www.BGdiagram.com