Game #5
Open Division Final
San Mateo, California
19 January 1997
11 Point Match
David Simpson (Black) vs. Kevin Mansfield (White)
Score: 4 - 6
Analysis by Kit Woolsey
In a match between two previous winners,
the charming Kevin Mansfield makes a third consecutive appearance
in the BGBB monthly finals,
this time to meet David Simpson
("yogi" in FIBSdom).
Interesting checker play and cube decisions abound
in this selected game from their match.
| |
Black
|
White
|
|
1.
|
|
42: 8/4 6/4
|
|
| Play 2a |
|
|
Black to play 64
Candidate Plays Equities
8/2 6/2 -0.203
24/14 -0.212
24/18 24/20 -0.277
13/7 13/9 -0.311
|
| |
|
2.
|
64: 8/2 6/2
|
|
|
|
KW:
|
This response would have been laughed at a few years ago as far too impure.
Today, players are looking at things from a new viewpoint.
Black sees that if he plays 24/18 13/9 or 24/14,
he is just exposing himself to a blot-hitting contest
when his opponent has the stronger board.
Instead,
he prefers to fight on even ground
as far as inner board strength and blots go,
at the cost of going a bit deep.
I prefer his choice.
|
| |
Black
|
White
|
|
2.
|
|
31: 8/5 6/5
|
|
| Play 3a |
|
|
Black to play 52
Candidate Plays Equities
24/17* -0.081
|
| |
|
3.
|
52: 24/17*
|
|
|
|
KW:
|
This hit is a must, despite the obvious dangers.
If Black goes quietly with 13/11 13/8,
he will soon get crushed.
This is an ideal time to spring a back checker,
while White is on the bar.
|
| Play 3b |
|
|
White to play 63
Candidate Plays Equities
Bar/22 13/7 +0.034
Bar/22 24/18 +0.021
Bar/16 -0.002
|
| |
| |
Black
|
White
|
|
3.
|
...
|
63: Bar/22 13/7
|
|
|
KW:
|
White has an interesting choice.
There is nothing wrong with Bar/22 24/18,
spreading his checkers out in an attempt
to catch Black's blot as it comes around.
Instead,
White chooses to unstack his heavy midpoint
and try to hem in Black's back checker.
Since Black does own the 2 point,
there is some danger of being blitzed
if White plays Bar/22 24/18.
I think he makes the right choice.
|
| |
Black
|
White
|
|
4.
|
53: 8/3* 6/3
|
65: Bar/20 13/7
|
|
| Play 5a |
|
|
Black to play 41
Candidate Plays Equities
13/8 -0.344
13/9 6/5* -0.344
17/13 6/5* -0.356
17/13 24/23 -0.397
|
| |
|
5.
|
41: 17/13 24/23
|
|
|
|
KW:
|
This doesn't feel right.
Black is overly concerned about escaping his back checker,
and in order to get that checker to where it can escape
he leaves his overall position a mess.
Six checkers on the midpoint and no structure up front -- ugh.
I think he should simply play 13/8,
making a valuable point in his outer board and unstacking the heavy midpoint.
|
| Play 5b |
|
|
White to play. Double or roll?
Equity (centered)
White 76.0% G 2.8% BG 0.1%
Black 24.0% G 1.1% BG 0.1%
Equity (White): +0.538
Equity (cubeless)
White 61.3% G 24.4% BG 0.9%
Black 38.7% G 10.5% BG 0.7%
Equity (White): +0.367
|
| |
| |
Black
|
White
|
|
5.
|
...
|
Roll
|
|
|
KW:
|
White can certainly consider turning the cube here.
He has a clear advantage with his superior structure,
and that advantage will get huge
if he hits a shot and Black rolls badly from the bar.
He may be being a bit cautious since he is ahead in the match,
but I don't think a two point lead with five points to go
should have much effect on initial cubes.
I would have spun it.
|
|
|
JF:
|
I would have doubled (level 7 evaluation).
|
| Play 5b |
|
|
White to play 32
Candidate Plays Equities
20/17* 13/11 +0.491
20/17*/15 +0.480
20/17* 6/4 +0.444
|
| |
| |
Black
|
White
|
|
5.
|
...
|
32: 20/17* 13/11
|
|
|
KW:
|
White is willing to leave the extra blot
in order to maximize coverage of his outer board.
I agree with him.
|
| Play 6a |
|
|
Black to play 61
Candidate Plays Equities
Bar/24 23/17 -0.586
Bar/24 13/7 -0.605
|
| |
| |
Black
|
White
|
|
6.
|
61: Bar/24 13/7
|
|
|
|
KW:
|
Black doesn't like throwing another blot
for White's hungry wolves to gobble up,
but Bar/24 23/17,
leaving the double shot and isolating his back checkers isn't very pretty.
Either way, Black is in a lot of trouble.
|
| Play 6b |
|
|
White to play. Double or roll?
Equity (centered)
White 88.4% G 5.3% BG 0.4%
Black 11.6% G 0.6% BG 0.0%
Equity (White): +0.818
Equity (cubeless)
White 68.8% G 29.2% BG 1.8%
Black 31.3% G 10.0% BG 0.7%
Equity (White): +0.579
|
| |
| |
Black
|
White
|
|
6.
|
...
|
Double -> 2
|
|
|
KW:
|
Now the double is very clear.
White has a big advantage,
which is likely to get even bigger.
It is the take which is the question.
|
|
|
JF:
|
I would have doubled (level 7 evaluation).
|
| Play 7a |
|
|
Black to play. Accept or drop?
Equity (centered)
White 88.4% G 5.3% BG 0.4%
Black 11.6% G 0.6% BG 0.0%
Equity (White): +0.818
Equity (Black owns 2 cube)
White 60.6% G 28.0% BG 1.8%
Black 39.4% G 1.9% BG 0.1%
Equity (Black): -0.489
Equity (cubeless)
White 68.8% G 29.2% BG 1.8%
Black 31.3% G 10.0% BG 0.7%
Equity (White): +0.579
|
| |
| |
Black
|
White
|
|
7.
|
Accept
|
|
|
|
KW:
|
This is pretty optimistic.
About all Black has going for him is his equality
in inner board strength,
and with White securely anchored
Black can't even take advantage
of this to attack.
Otherwise, his position is awful.
The match score gives him some incentive to make marginal takes
since his recube vig is pretty hefty,
but I still think he should have let it go.
|
| Play 7b |
|
|
White to play 51
Candidate Plays Equities
7/2*/1* +0.529
17/11 +0.415
|
| |
| |
Black
|
White
|
|
7.
|
...
|
51: 7/2*/1*
|
|
|
KW:
|
This is not the game plan that White had in mind.
He could continue to play purely with 17/11,
but that would give Black a chance to consolidate
and get back in the game.
The blot dangling on Black's bar point
makes White's double hit very attractive.
If Black doesn't roll well fast
White will blow him away,
and even if Black strikes back
White will still have a pretty good position.
I think he makes a good choice.
|
| |
Black
|
White
|
|
8.
|
66: Cannot move
|
|
|
| Play 8b |
|
|
White to play 43
Candidate Plays Equities
7/3 6/3 +0.870
11/7 24/21 +0.790
|
| |
|
8.
|
...
|
43: 11/7 24/21
|
|
|
KW:
|
It is possible that this was a simple oversight.
White was looking
to cover the blot on his ace point with fives or sixes.
When he rolled the 4-3 and couldn't do it,
he may have quickly covered the blot on the bar point
and then looked around for the best three.
I believe he should have made his three point.
Even though that leaves him without any more builders in position,
the solid four-point board will serve him very well in the future.
|
| |
Black
|
White
|
|
9.
|
53: Bar/22 0
|
|
|
| Play 9b |
|
|
White to play 22
Candidate Plays Equities
13/11 7/3*/1 +1.181
7/5 7/3*/1 +1.171
17/15 7/3*/1 +1.166
6/4 7/3*/1 +1.153
7/3(2)* +1.098
|
| |
|
9.
|
...
|
22: 7/3(2)*
|
|
|
KW:
|
This play looks automatic,
but is it?
Leaving that blot on the ace point could be very costly
if Black rolls an ace.
White has enough power to carry out a straight blitz,
particularly since he will almost certainly be able
to pick up Black's other blot
if he doesn't get stuck on the bar himself.
I strongly believe he should have played impurely by picking and passing,
probably with 17/15 7/3* 3/1.
|
| |
Black
|
White
|
|
10.
|
32: Bar/23 0
|
|
|
| Play 10b |
|
|
White to play 21
Candidate Plays Equities
3/1 3/2* +0.977
4/2*/1 +0.966
21/18* +0.955
|
| |
|
10.
|
...
|
21: 21/18*
|
|
|
KW:
|
Another "automatic" play which might not be as automatic as it looks.
Now Black has good aces and good twos.
Quite possibly White should make a shifting play
of 4/2* 2/1 or 3/2* 3/1,
giving Black fewer good numbers.
White is a favorite to pick up the other blot.
Still if his blitz falls apart he wouldn't like having done that,
while after his actual play even if things go badly
he will be in pretty good shape.
So,
while I would make the same play he made,
I think it is important to notice all the possible alternatives.
|
| |
Black
|
White
|
|
11.
|
32: Bar/23 0
|
|
|
| Play 11b |
|
|
White to play 53
Candidate Plays Equities
6/1 24/21 +0.922
6/1 17/14 +0.890
6/1 18/15 +0.881
6/1 13/10 +0.828
|
| |
|
11.
|
...
|
53: 6/1 17/14
|
|
|
KW:
|
Covering is clear, of course.
I think White chooses the best three,
giving himself maximum coverage of his outer
board as well as duplicating his opponent's twos
(the latter a very minor factor).
|
| |
Black
|
White
|
|
12.
|
61: Cannot move
|
|
|
| Play 12b |
|
|
White to play 62
Candidate Plays Equities
14/6 +0.947
24/18 13/11 +0.933
18/10 +0.922
14/8 18/16 +0.907
14/8 13/11 +0.901
|
| |
|
12.
|
...
|
62: 18/10
|
|
|
KW:
|
Again White shows good technique.
If he makes the anchor with 24/18 13/11,
he would just be faced with the problem of clearing it in the future.
Instead he concentrates on maximum builders for his eight and bar points.
|
| |
Black
|
White
|
|
13.
|
32: Bar/23 13/10
|
|
|
| Play 13b |
|
|
White to play 52
Candidate Plays Equities
14/7 +0.580
13/8 10/8 +0.566
14/9 10/8 +0.554
|
| |
|
13.
|
...
|
52: 14/7
|
|
|
KW:
|
White chooses to leave only one direct shot.
The problem is that the shot is a big one,
since Black would love to hit while escaping a back checker.
I prefer 13/8 10/8.
Locking up the eight point could prove to be very important
if things start to go sour,
and the loss of outfield control and the double shot aren't too serious.
|
| |
Black
|
White
|
|
14.
|
61: 13/6
|
54: 24/15*
|
|
15.
|
53: Cannot move
|
|
|
| Play 15b |
|
|
White to play 42
Candidate Plays Equities
10/6 7/5 +1.223
15/11 7/5 +1.217
13/9 7/5 +1.212
13/7 +1.191
10/6 13/11 +1.153
|
| |
|
15.
|
...
|
42: 15/11 7/5
|
|
|
KW:
|
White sees that there is little reason to make the bar point.
He can no longer plan to win by priming,
since even if Black enters immediately
Black has those three outfield checkers to play with.
Instead White properly clears the bar point
before it becomes a problem later.
|
| |
Black
|
White
|
|
16.
|
32: Bar/23 13/10
|
55: 13/8(2) 11/6 10/5
|
|
17.
|
64: 13/7 13/9
|
|
|
| Play 17b |
|
|
White to play 44
Candidate Plays Equities
8/Off/(2) +1.303
8/4(2) 5/1 4/Off +1.293
|
| |
|
17.
|
...
|
44: 8/4(2) 5/1 4/Off
|
|
|
KW:
|
This looks more flexible than 8/Off(2),
which would potentially leave White in trouble
if he rolled some threes at the wrong time.
On the other hand,
that extra checker off could make a difference both
for the gammon and the win (if hit),
and a third checker on the ace point is definitely something to be avoided
if possible.
It looks close,
but I prefer 8/Off(2).
|
| Play 18a |
|
|
Black to play 42
Candidate Plays Equities
9/5 6/4 -1.314
9/5 7/5 -1.321
10/4 -1.333
9/5 10/8 -1.334
7/3 6/4 -1.395
|
| |
| |
Black
|
White
|
|
18.
|
42: 9/5 7/5
|
|
|
|
KW:
|
Another automatic play, right?
Well, maybe not.
Black isn't getting a shot next roll,
so making a point right now isn't vital.
Black's goal is to make both the five and four points,
so he will have a decent reception committee if he hits a shot.
With this in mind,
I think he should have played 9/5 6/4,
starting both points.
Almost anything he then rolls next turn will make one of the points,
and the other will be immediately slotted,
which is what he wants.
After his actual play,
he needs to roll specifically a six or a two to slot the four point,
and he needs his sixes to escape a back checker.
Sometimes two halves are better than one whole.
|
| |
Black
|
White
|
|
18.
|
...
|
66: 6/Off(3) 5/Off
|
|
| Play 19a |
|
|
Black to play 21
Candidate Plays Equities
6/4 6/5 -1.455
6/4 10/9 -1.457
6/4 5/4 -1.471
10/8 6/5 -1.477
10/7 -1.494
|
| |
|
19.
|
21: 10/7
|
|
|
|
KW:
|
Black is concerned about saving pips for the gammon.
I do not agree.
He is unlikely to get off the gammon in any event if he doesn't hit a shot.
If he does hit a shot he simply must have his four point,
and this is the time to start it.
Even if he gets a shot next roll,
hits it,
and doesn't cover,
he still won't mind having the four point slotted.
|
| |
Black
|
White
|
|
19.
|
...
|
64: 5/Off 5/1
|
|
| Play 20a |
|
|
Black to play 44
Candidate Plays Equities
23/15(2) -1.440
23/11 23/19 -1.445
23/7 -1.452
23/15 23/19 7/3 -1.481
23/15 23/19 6/2 -1.489
23/11 7/3 -1.511
|
| |
|
20.
|
44: 23/15(2)
|
|
|
|
KW:
|
There is no reason to run with three checkers.
That wouldn't make life any more awkward for White,
and it would make some of his smaller numbers crushers.
Black runs two checkers into the outfield
because he is concerned about the possible backgammon.
This may be a bit too conservative.
The best play to make his board is clearly 23/7,
but that could cost in the backgammon department.
I prefer a compromise of 23/11 23/19.
This only costs one pip as far as the backgammon goes,
and leaves more flexibility in the outfield.
|
| |
Black
|
White
|
|
20.
|
...
|
22: 3/1(2)
|
|
| Play 21a |
|
|
Black to play 55
Candidate Plays Equities
23/18 15/5 15/10 -1.356
23/13 15/10(2) -1.357
23/13 15/5 -1.373
15/5(2) -1.556
15/5 15/10 6/1 -1.589
|
| |
|
21.
|
55: 23/13 15/10(2)
|
|
|
|
KW:
|
Unlike last turn,
it is now vital for Black to leave only one checker
on the 23 point.
This makes some of his opponent's twos horror rolls,
while if Black holds the anchor
White won't have to play those twos.
|
| |
Black
|
White
|
|
21.
|
...
|
52: 4/Off 4/2*
|
|
| Play 22a |
|
|
Black to play 54
Candidate Plays Equities
Bar/21* 13/8 -0.022
Bar/21* 10/5 -0.050
Bar/21* 21/16 -0.059
Bar/21* 6/1 -0.071
|
| |
|
22.
|
54: Bar/21* 13/8
|
|
|
|
KW:
|
Black properly moves in for the kill.
He doesn't mind leaving the blot in 4-4 range -- in fact,
if White rolls 4-4 Black would rather be hit
so he can go back and pick up the other blot.
|
| |
Black
|
White
|
|
22.
|
...
|
61: Bar/18*
|
|
23.
|
52: Bar/23* 21/16
|
44: Bar/17*/9*
|
|
24.
|
53: Bar/20 Bar/22
|
|
|
| Play 24b |
|
|
White to play 31
Candidate Plays Equities
9/5* -0.116
18/14 -0.224
9/6 18/17 -0.303
|
| |
|
24.
|
...
|
31: 18/14
|
|
|
KW:
|
Is this right?
The two relevant considerations are number of shots and ease of getting home.
Clearly White will be better off being on the five point
if he survives,
since he will be a direct four away from safety.
What about shots?
Staying on the nine point leaves 27 shots just from the 6 away and 4 away
(it helps to have taken the time to memorize these).
Add in the 5-2, and it comes to 29 shots.
If he plays 9/5*,
several of the sixes, fours, and aces miss
(such as 6-4, 4-4, 6-1, 3-1, 2-1, 1-1).
So 9/5* is safer immediately as well as being closer to home.
Therefore it must be superior.
|
| Play 25a |
|
|
Black to play. Double or roll?
Equity (Black owns 2 cube)
Black 84.5% G 0.0% BG 0.0%
White 15.5% G 9.4% BG 0.6%
Equity (Black): +0.589
Equity (White owns 4 cube)
Black 65.9% G 0.0% BG 0.0%
White 34.1% G 5.6% BG 0.4%
Equity (White): -0.259
Equity (cubeless)
Black 73.1% G 0.0% BG 0.0%
White 26.9% G 10.3% BG 0.7%
Equity (Black): +0.352
|
| |
| |
Black
|
White
|
|
25.
|
Double -> 4
|
|
|
|
KW:
|
Black sees his chance.
If he hits the shot and White flunks,
Black will be on his way
toward closing out two checkers and having a big advantage.
Since as we have seen the cube is of little value to White,
this is an excellent time to double.
|
|
JF:
|
I would not have doubled (level 7 evaluation).
|
| Play 25b |
|
|
White to play. Accept or drop?
Equity (Black owns 2 cube)
Black 84.5% G 0.0% BG 0.0%
White 15.5% G 9.4% BG 0.6%
Equity (Black): +0.589
Equity (White owns 4 cube)
Black 65.9% G 0.0% BG 0.0%
White 34.1% G 5.6% BG 0.4%
Equity (White): -0.259
Equity (cubeless)
Black 73.1% G 0.0% BG 0.0%
White 26.9% G 10.3% BG 0.7%
Equity (Black): +0.352
|
| |
| |
Black
|
White
|
|
25.
|
...
|
Pass
|
|
|
KW:
|
No way!
True,
White should be more conservative with his takes
than for money,
since cube ownership and gammons are worth little to him at this score.
Still,
he can use all four points,
and he is far from lost.
Black might miss this shot,
or White might enter quickly
and get the other checker to safety.
Black still has to build his four point
before he really has anything,
and that may not be so easy.
I believe this is a clear take.
|
|
JF:
|
I would have accepted (level 7 evaluation).
|
|
|
Black went on to win the match 11-7.
|
The game was recorded on tape and transcribed by Richard McIntosh.
Rollouts were made by Richard McIntosh,
using
JellyFish Analyzer 2.02.
Rollout results show equities
for the player on move.
Candidate plays were better than or
within 0.100 equity of the actual plays,
evaluated at level 7.
Parameter
values for rollouts on moves were:
- level 5
- 7776 games (36x216)
- horizon 7
- seed 119
Standard deviations of equity estimates were between 0.005 and 0.010.
Parameter values for rollouts on cube decisions were:
- level 5
- 23328 games (36x648)
- full game
- seed 119
- settlement limit 0.550
Copyright © 1996-2010
BackGammon By the Bay