Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Block or hang back? 2a 4a 1-1 to play

24 views
Skip to first unread message

Walt

unread,
May 10, 2013, 12:21:48 PM5/10/13
to



XGID=-ABABbCBB----B-a--abbbbba-:0:0:1:11:3:1:0:5:3

X:X O:O
Score is X:3 O:1 5 pt.(s) match.
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O O O O O |
| X | | O O O O O |
| | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | X |
| X X | | X O X X |
| X X | | X O X X X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 90 O: 98 X-O: 3-1/5
Cube: 1
X to play 11


--
//Walt

Bradley K. Sherman

unread,
May 10, 2013, 12:25:18 PM5/10/13
to
Walt <walt_...@yahoo.com> wrote:
>Score is X:3 O:1 5 pt.(s) match.
> +13-14-15-16-17-18------19-20-21-22-23-24-+
> | X O O | | O O O O O O |
> | X | | O O O O O |
> | | | |
> | | | |
> | | | |
> | |BAR| |
> | | | |
> | | | |
> | | | X |
> | X X | | X O X X |
> | X X | | X O X X X X |
> +12-11-10--9--8--7-------6--5--4--3--2--1-+
> X:90 O:98, Cube:1, X to play 11

13/12(2),8/7(2)

--bks

crf

unread,
May 10, 2013, 3:15:57 PM5/10/13
to
13/12(2), 3/1

get close, but don't move within direct range, so it's easier to clear the point next time, but leave landing spots available.

badgolferman

unread,
May 10, 2013, 4:03:30 PM5/10/13
to
4/1, 4/3

Freeven

unread,
May 10, 2013, 4:43:30 PM5/10/13
to
When the race is close, it's generally better to make the blocking
point. X will only be 12 pips ahead after this move, so blocking O's 66,
even for a single turn, makes a surprising difference. O rolls 66 about
3% on any given turn, and while this doesn't guarantee him a win, it
increases his chances by a percent or two, especially when he can win
with the cube. If the block is held for more than one turn, the equity
gains add up. Plus, O will often have to choose between wasting pips and
breaking anchor in the face of attackers, either of which improves X's
chances. I'd be surprised if 13/11(2) isn't best here.

Tim Chow

unread,
May 11, 2013, 2:27:41 PM5/11/13
to
On May 10, 4:43 pm, Freeven <no...@nowhere.com> wrote:
> When the race is close, it's generally better to make the blocking
> point. X will only be 12 pips ahead after this move, so blocking O's 66,
> even for a single turn, makes a surprising difference.

I agree with blocking 13/11(2), though I think there are other reasons
for it than just the pip count. X has enough points filled in that
clearing the 11pt when he needs to shouldn't be a big deal, especially
since O's timing is likely to force her to concede something soon. If
X's two checkers on the 7pt were on his 6pt instead then I'd be warier
of playing 13/11(2), not just because of the two extra pips in the
race, but because of the increased risk of leaving a shot on the 11pt
(for example, with an immediate 63).

---
Tim Chow

Walt

unread,
May 11, 2013, 7:00:41 PM5/11/13
to
As we saw in the previous position, filling in homeboard points is not a
priority. 4/3 4/1 is 9th best and a ~.15 error according to XGR++

The real choice seemed to be whether to block 6s with 13/11(2) or to
creep up one point with 13/12(2) and then shuffle the inner board
checkers. There's a good argument to be made for hanging back - don't
make a point you'll just have to clear soon. But here, O is close
enough in the race that blocking is correct.

I and the rest of the chouette voted to creep, settling on 13/12(2) 3/1
as the choice. One our bigger blunders lately.

How important is the race? In the two variants below, I've backed up O
by 5 pips an then by 11 pips. With 5 pips more of a lead, blocking
still comes out on top, but by only .016. With 11 pips more, the
blocking move becomes a .035 error.
1. Rollout� 13/11(2) eq:+0.455
Player: 67.27% (G:1.53% B:0.08%)
Opponent: 32.73% (G:1.14% B:0.03%)
Confidence: �0.007 (+0.448..+0.462) - [100.0%]
Duration: 4 minutes 16 seconds

2. Rollout� 13/12(2) 3/1 eq:+0.375 (-0.080)
Player: 63.67% (G:1.25% B:0.03%)
Opponent: 36.33% (G:1.16% B:0.02%)
Confidence: �0.007 (+0.368..+0.382) - [0.0%]
Duration: 3 minutes 42 seconds

3. Rollout� 13/12(2) 4/3 2/1 eq:+0.371 (-0.084)
Player: 63.64% (G:1.83% B:0.05%)
Opponent: 36.36% (G:1.41% B:0.01%)
Confidence: �0.008 (+0.363..+0.379) - [0.0%]
Duration: 4 minutes 08 seconds

4. XG Roller++ 13/12(2) 8/7(2) eq:+0.370 (-0.084)
Player: 62.96% (G:1.59% B:0.03%)
Opponent: 37.04% (G:1.20% B:0.02%)

5. 3-ply 13/12(2) 7/6(2) eq:+0.335 (-0.119)
Player: 63.10% (G:1.79% B:0.02%)
Opponent: 36.90% (G:1.50% B:0.02%)


� 1296 Games rolled with Variance Reduction.
Moves: 3-ply, cube decisions: XG Roller


eXtreme Gammon Version: 2.10, MET: Kazaross XG2


////////////variant 1//////////////////////////

XGID=-ABABbCBB----Bb----bbbbba-:0:0:1:11:3:1:0:5:3

X:Player 1 O:Player 2
Score is X:3 O:1 5 pt.(s) match.
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O | | O O O O O O |
| X O | | O O O O O |
| | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | X |
| X X | | X O X X |
| X X | | X O X X X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 90 O: 103 X-O: 3-1/5
Cube: 1
X to play 11

1. 4-ply 13/11(2) eq:+0.582
Player: 70.34% (G:1.17% B:0.01%)
Opponent: 29.66% (G:0.86% B:0.00%)

2. 4-ply 13/12(2) 3/1 eq:+0.566 (-0.016)
Player: 69.41% (G:1.27% B:0.02%)
Opponent: 30.59% (G:0.99% B:0.01%)

3. 4-ply 13/12(2) 4/3 2/1 eq:+0.555 (-0.027)
Player: 69.40% (G:1.40% B:0.02%)
Opponent: 30.60% (G:1.04% B:0.01%)

4. 4-ply 13/12(2) 8/7(2) eq:+0.543 (-0.040)
Player: 69.35% (G:1.31% B:0.01%)
Opponent: 30.65% (G:0.89% B:0.01%)

5. 3-ply 13/12(2) 4/2 eq:+0.498 (-0.085)
Player: 68.51% (G:1.49% B:0.01%)
Opponent: 31.49% (G:1.19% B:0.03%)


eXtreme Gammon Version: 2.10, MET: Kazaross XG2


/////////////variant 2//////////////

XGID=-ABABbCBB----Bb---abbbbb--:0:0:1:11:3:1:0:5:3

X:Player 1 O:Player 2
Score is X:3 O:1 5 pt.(s) match.
+13-14-15-16-17-18------19-20-21-22-23-24-+
| X O O | | O O O O O |
| X O | | O O O O O |
| | | |
| | | |
| | | |
| |BAR| |
| | | |
| | | |
| | | X |
| X X | | X O X X |
| X X | | X O X X X X |
+12-11-10--9--8--7-------6--5--4--3--2--1-+
Pip count X: 90 O: 109 X-O: 3-1/5
Cube: 1
X to play 11

1. XG Roller+ 13/12(2) 8/7(2) eq:+0.699
Player: 74.85% (G:3.18% B:0.21%)
Opponent: 25.15% (G:0.91% B:0.01%)

2. XG Roller+ 13/12(2) 4/3 2/1 eq:+0.698 (-0.001)
Player: 74.71% (G:3.33% B:0.23%)
Opponent: 25.29% (G:1.31% B:0.03%)

3. XG Roller+ 13/12(2) 3/1 eq:+0.681 (-0.019)
Player: 74.19% (G:1.47% B:0.03%)
Opponent: 25.81% (G:0.90% B:0.00%)

4. XG Roller+ 13/11(2) eq:+0.664 (-0.035)
Player: 74.38% (G:1.36% B:0.01%)
Opponent: 25.62% (G:0.93% B:0.01%)

5. 4-ply 13/12(2) 4/2 eq:+0.667 (-0.033)
Player: 73.41% (G:2.24% B:0.06%)
Opponent: 26.59% (G:1.52% B:0.02%)


eXtreme Gammon Version: 2.10, MET: Kazaross XG2



Freeven

unread,
May 11, 2013, 11:15:20 PM5/11/13
to
On 5/11/2013 2:27 PM, Tim Chow wrote:
> On May 10, 4:43 pm, Freeven <no...@nowhere.com> wrote:
>> When the race is close, it's generally better to make the blocking
>> point. X will only be 12 pips ahead after this move, so blocking O's 66,
>> even for a single turn, makes a surprising difference.
>
> I agree with blocking 13/11(2), though I think there are other reasons
> for it than just the pip count.

I also think there are other reasons than just the pip count, and said
so in my original post. But the race is generally the most important
consideration in these positions, so that's the one I emphasized.

> X has enough points filled in that
> clearing the 11pt when he needs to shouldn't be a big deal, especially
> since O's timing is likely to force her to concede something soon. If
> X's two checkers on the 7pt were on his 6pt instead then I'd be warier
> of playing 13/11(2), not just because of the two extra pips in the
> race, but because of the increased risk of leaving a shot on the 11pt
> (for example, with an immediate 63).

Well, yeah. If you change several features of the position, one of which
alters its very nature, it's not surprising that things change. Moving
two of X's checkers inside not only changes the race, and takes away an
important landing spot, it starves X of most of his timing. You've now
made that the dominant feature of the position, and the best move will
reflect that.

I think the variations Walt posted do a good job of confirming my
observations about the race. If you want to test the value of the extra
landing spot, I'd suggest finding a variation that tries to isolate that
variable without introducing timing issues into the mix. Maybe move X's
checkers from his 8pt to his 7pt (or vice versa), and compensate for the
change in the race by shifting O's outfield checkers.
0 new messages