Axel Reichert <mail@axel-reichert.de> writes:

GNU Backgammon Position ID: bLcJASJiz8EAOA
Match ID : UYlkARAAIAAE
+13-14-15-16-17-18------19-20-21-22-23-24-+ O: gnubg
| X O O | | O O O O X | 1 point
| X O | | O O O O X |
| | | O X |
| | | |
| | | |
v| |BAR| | 11 point match
| | | |
| | | X |
| X | | X |
| X | | X X | Rolled 11
| O X | | X X O X O | 4 points
+12-11-10--9--8--7-------6--5--4--3--2--1-+ X: axel (Cube: 2)
Pip counts: O 122, X 158

* axel moves 24/23 13/12* 5/4*(2)
Alert: very bad move ( -0.248)

My intention was to move up once and then gain some time by the double
hit to escape. I do not understand why trying to recycle with 2/1* is attractive.

Rolled 11:
1. Rollout 24/23(3) 2/1* Eq.: -0.650
0.281 0.065 0.004 - 0.719 0.279 0.021 CL -0.707 CF -0.650
[0.002 0.001 0.000 - 0.002 0.002 0.000 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
277 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
2. Rollout 24/23(3) 6/5 Eq.: -0.663 (-0.013)
0.277 0.051 0.002 - 0.723 0.277 0.018 CL -0.725 CF -0.663
[0.002 0.001 0.000 - 0.002 0.002 0.000 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
213 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
3. Rollout 24/23(2) 13/12* 2/1* Eq.: -0.720 (-0.070)
0.286 0.073 0.004 - 0.714 0.336 0.040 CL -0.777 CF -0.720
[0.001 0.001 0.000 - 0.001 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
272 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
4. Rollout 24/23(3) 13/12* Eq.: -0.735 (-0.085)
0.289 0.065 0.004 - 0.711 0.356 0.036 CL -0.796 CF -0.735
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
378 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
5. Rollout 24/23(2) 13/12*(2) Eq.: -0.737 (-0.087)
0.280 0.063 0.003 - 0.720 0.343 0.032 CL -0.797 CF -0.737
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
257 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
6. Rollout 24/23(2) 6/5 2/1* Eq.: -0.776 (-0.126)
0.259 0.058 0.003 - 0.741 0.334 0.031 CL -0.831 CF -0.776
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
262 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
7. Rollout 24/23 13/12*(2) 2/1* Eq.: -0.800 (-0.150)
0.253 0.062 0.003 - 0.747 0.332 0.042 CL -0.852 CF -0.800
[0.001 0.001 0.000 - 0.001 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
342 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
8. Rollout 24/23(2) 5/4*(2) Eq.: -0.828 (-0.179)
0.251 0.047 0.002 - 0.749 0.356 0.032 CL -0.886 CF -0.828
[0.002 0.001 0.000 - 0.002 0.003 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
225 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
9. Rollout 24/23 13/12* 6/5 2/1* Eq.: -0.833 (-0.184)
0.258 0.065 0.003 - 0.742 0.356 0.059 CL -0.887 CF -0.833
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
413 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
10. Rollout 24/23(2) 13/12* 6/5 Eq.: -0.840 (-0.190)
0.275 0.063 0.004 - 0.725 0.404 0.052 CL -0.899 CF -0.840
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
369 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune
* 22. Rollout 24/23 13/12* 5/4*(2) Eq.: -0.898 (-0.248)
0.253 0.063 0.004 - 0.747 0.389 0.070 CL -0.951 CF -0.898
[0.002 0.001 0.000 - 0.002 0.002 0.001 CL 0.005 CF 0.005]
Truncated cubeful rollout (depth 7) with variance reduction
381 games, Mersenne Twister dice gen. with seed 446928448 and quasi-random dice
Stop when std.errs. are small enough: limit 0.0050 (min. 36 games)
Play: 1-ply cubeful prune
keep the first 0 0-ply moves and up to 8 more moves within equity 0.16
Cube: 1-ply cubeful prune
Different evaluations after 2 plies:
Play: 0-ply cubeful [expert]
Cube: 1-ply cubeful prune

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On December 14, 2022 at 1:12:30 PM UTC-7, Axel Reichert wrote:

Axel Reichert <ma...@axel-reichert.de> writes:

GNU Backgammon Position ID: bLcJASJiz8EAOA
Match ID : UYlkARAAIAAE
Rolled 11

* axel moves 24/23 13/12* 5/4*(2)
Alert: very bad move ( -0.248)

If I had answered seriously, I had said
24/23 5/4*(2) 2/1* (which is #13 vs.
your move #14 in the default hint list).

My intention was to move up once and then
gain some time by the double hit to escape.

I thought the same. Even after seeing
the rollout, I still would play the same
as I believe rollouts are based on how
the bot would play against itself.

I do not understand why trying to recycle
with 2/1* is attractive.

My reasons for 2/1* vs. your 13/12*
were to not create another blot and
to keep O from anchoring on 1 point.

Misery likes company... :)

MK

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/14/2022 3:12 PM, Axel Reichert wrote:

My intention was to move up once and then gain some time by the double
hit to escape. I do not understand why trying to recycle with 2/1* is attractive.

Well, after 24/23(3), there are only three plausible candidates, and the
bot rates 2/1* and 6/5 about the same. So I think it's not so much that
2/1* is attractive, as that 13/12* is unattractive, since it gives O
extra shots. As for 6/5, it's more constructive strategically than
2/1*, but 2/1* tactically gains a tempo without giving O more shots.

---
Tim Chow

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 12/14/2022 3:12 PM, Axel Reichert wrote:
then gain some time by the double

hit to escape. I do not understand why trying to recycle with 2/1* is attractive.

I don't know that the idea is to recirculate the checker on the deuce
point (and the bots won't tell us).

My take is that a checker prematurely buried on the ace is hardly any
worse than a checker prematurely buried on the deuce; one useful thing
that that checker can be used for is a tempo hit, which always has it's
pluses an minuses.

Of course, if it does recirculate, that's not so bad, and it might well recirculate anyway if left on the deuce. So what's to lose by hitting
loose on the ace, other than the opportunity cost of there being a
better play?

OTOH, the play I dismissed as "passive" is actually quite well respected
by XG, and within the margin of error after 1296 trials. The main
takeaway is to move all three runners up to the edge, and then not leave
blots strewn all over the board. Hitting is not essential.



1. Rollout¹ 24/23(3) 2/1* eq:-0.660
Player: 27.35% (G:5.99% B:0.44%)
Opponent: 72.65% (G:28.04% B:1.88%)
Confidence: ±0.008 (-0.668..-0.652) - [76.3%]
Duration: 4 minutes 29 seconds

2. Rollout¹ 24/23(3) 6/5 eq:-0.664 (-0.004)
Player: 27.17% (G:4.52% B:0.25%)
Opponent: 72.83% (G:27.34% B:1.71%)
Confidence: ±0.008 (-0.672..-0.656) - [23.7%]
Duration: 4 minutes 25 seconds




eXtreme Gammon Version: 2.10, MET: Kazaross XG2



--
Ah....Clem
The future is fun, the future is fair.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)