Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Here is how Gnubg cheats - take two.
41 views
Skip to first unread message
mu...@compuplus.net
May 1, 2012, 7:11:18 AM5/1/12
to
GNU Backgammon Position ID: 3psDAIBtuwIADA
Match ID : EQErA6AAUAAE
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ O: gnubg
| O O O O O | | O O | 10 points
| O O O O O | | | Rolled 62
| O | | |
| | | |
| | | |
| |BAR| |^ 25 point match
| | | |
| X X | | |
| X X X | | |
X | O X X X X | | |
X | O X X X X | | | 10 points
+24-23-22-21-20-19------18-17-16-15-14-13-+ X: Murat 25x100 (Cube: 2)
First, think about how you would play?
Then, think about how you think gnubg would play?
Here is what you will get if you click on "hint":
===========================================================================
1. Cubeful 3-ply 24/16 MWC: 41.75% 0.214 0.000 0.000 - 0.786 0.139 0.001 3-ply cubeful prune [grandmaster]
2. Cubeful 3-ply 24/18 8/6 MWC: 41.75% ( -0.00%) 0.201 0.000 0.000 - 0.799 0.127 0.001 3-ply cubeful prune [grandmaster]
3. Cubeful 3-ply 24/18 7/5 MWC: 41.75% ( -0.00%) 0.200 0.000 0.000 - 0.800 0.140 0.001 3-ply cubeful prune [grandmaster]
4. Cubeful 3-ply 24/18 6/4 MWC: 41.75% ( -0.00%) 0.208 0.000 0.000 - 0.792 0.168 0.001 3-ply cubeful prune [grandmaster]
5. Cubeful 2-ply 24/18 3/1 MWC: 41.75% ( -0.00%) 0.190 0.000 0.000 - 0.810 0.200 0.005 2-ply cubeful prune [world class]
6. Cubeful 2-ply 8/6 7/1 MWC: 41.75% ( -0.00%) 0.163 0.000 0.000 - 0.837 0.164 0.008 2-ply cubeful prune [world class]
7. Cubeful 2-ply 24/18 5/3 MWC: 41.75% ( -0.00%) 0.180 0.000 0.000 - 0.820 0.202 0.005 2-ply cubeful prune [world class]
8. Cubeful 2-ply 7/1 6/4 MWC: 41.75% ( -0.00%) 0.170 0.000 0.000 - 0.830 0.184 0.008 2-ply cubeful prune [world class]
9. Cubeful 2-ply 24/18 4/2 MWC: 41.75% ( -0.00%) 0.184 0.000 0.000 - 0.816 0.217 0.005 2-ply cubeful prune [world class]
10. Cubeful 2-ply 8/2 7/5 MWC: 41.75% ( -0.00%) 0.156 0.000 0.000 - 0.844 0.169 0.008 2-ply cubeful prune [world class]
11. Cubeful 2-ply 7/1 3/1 MWC: 41.69% ( -0.06%) 0.169 0.000 0.000 - 0.831 0.191 0.009 2-ply cubeful prune [world class]
===========================================================================
In this case, the playing strength and analysis strength are identical.
But, just in case, let's take a look at the rollouts. The first 10 go very quickly. And the 11th best move (i.e. 8/2 6/4) takes a few minutes more to roll out but actually comes on top as below: !!!
===========================================================================
1. Rollout 8/2 6/4 MWC: 41.89%
0.167 0.000 0.000 - 0.833 0.182 0.005 CL 42.90% CF 41.89%
[0.002 0.000 0.000 - 0.002 0.002 0.000 CL 0.03% CF 0.03%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
2. Rollout 24/16 MWC: 41.75% ( -0.14%)
0.214 0.000 0.000 - 0.786 0.139 0.001 CL 44.08% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
3. Rollout 24/18 8/6 MWC: 41.75% ( -0.14%)
0.201 0.000 0.000 - 0.799 0.127 0.001 CL 43.96% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
4. Rollout 24/18 7/5 MWC: 41.75% ( -0.14%)
0.200 0.000 0.000 - 0.800 0.140 0.001 CL 43.85% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
5. Rollout 24/18 6/4 MWC: 41.75% ( -0.14%)
0.208 0.000 0.000 - 0.792 0.168 0.001 CL 43.73% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
6. Rollout 24/18 3/1 MWC: 41.75% ( -0.14%)
0.191 0.000 0.000 - 0.809 0.185 0.002 CL 43.29% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
7. Rollout 24/18 5/3 MWC: 41.75% ( -0.14%)
0.181 0.000 0.000 - 0.819 0.192 0.002 CL 43.08% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
8. Rollout 8/6 7/1 MWC: 41.75% ( -0.14%)
0.168 0.000 0.000 - 0.832 0.164 0.006 CL 43.07% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
9. Rollout 24/18 4/2 MWC: 41.75% ( -0.14%)
0.189 0.000 0.000 - 0.811 0.210 0.003 CL 43.05% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
10. Rollout 7/1 3/1 MWC: 41.75% ( -0.14%)
0.176 0.000 0.000 - 0.824 0.189 0.006 CL 42.99% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
11. Rollout 7/1 6/4 MWC: 41.75% ( -0.14%)
0.173 0.000 0.000 - 0.827 0.187 0.006 CL 42.95% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
12. Rollout 8/2 7/5 MWC: 41.75% ( -0.14%)
0.162 0.000 0.000 - 0.838 0.174 0.006 CL 42.89% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
13. Cubeful 2-ply 8/2 3/1 MWC: 41.50% ( -0.39%)
0.152 0.000 0.000 - 0.848 0.197 0.009
2-ply cubeful prune [world class]
===========================================================================
There you have it. Plain and simple. In broad daylight. Actually a better proof than the one I posted the last time, since in this case the playing and analysis strengths are identical.
So, why doesn't gnubg make the best move 8/2 6/4 but instead moves 24/16...?
What kind of cocksucking motherfuckers are involved in this "gnubg project"...??
MK
Match ID : EQErA6AAUAAE
+-1--2--3--4--5--6-------7--8--9-10-11-12-+ O: gnubg
| O O O O O | | O O | 10 points
| O O O O O | | | Rolled 62
| O | | |
| | | |
| | | |
| |BAR| |^ 25 point match
| | | |
| X X | | |
| X X X | | |
X | O X X X X | | |
X | O X X X X | | | 10 points
+24-23-22-21-20-19------18-17-16-15-14-13-+ X: Murat 25x100 (Cube: 2)
First, think about how you would play?
Then, think about how you think gnubg would play?
Here is what you will get if you click on "hint":
===========================================================================
1. Cubeful 3-ply 24/16 MWC: 41.75% 0.214 0.000 0.000 - 0.786 0.139 0.001 3-ply cubeful prune [grandmaster]
2. Cubeful 3-ply 24/18 8/6 MWC: 41.75% ( -0.00%) 0.201 0.000 0.000 - 0.799 0.127 0.001 3-ply cubeful prune [grandmaster]
3. Cubeful 3-ply 24/18 7/5 MWC: 41.75% ( -0.00%) 0.200 0.000 0.000 - 0.800 0.140 0.001 3-ply cubeful prune [grandmaster]
4. Cubeful 3-ply 24/18 6/4 MWC: 41.75% ( -0.00%) 0.208 0.000 0.000 - 0.792 0.168 0.001 3-ply cubeful prune [grandmaster]
5. Cubeful 2-ply 24/18 3/1 MWC: 41.75% ( -0.00%) 0.190 0.000 0.000 - 0.810 0.200 0.005 2-ply cubeful prune [world class]
6. Cubeful 2-ply 8/6 7/1 MWC: 41.75% ( -0.00%) 0.163 0.000 0.000 - 0.837 0.164 0.008 2-ply cubeful prune [world class]
7. Cubeful 2-ply 24/18 5/3 MWC: 41.75% ( -0.00%) 0.180 0.000 0.000 - 0.820 0.202 0.005 2-ply cubeful prune [world class]
8. Cubeful 2-ply 7/1 6/4 MWC: 41.75% ( -0.00%) 0.170 0.000 0.000 - 0.830 0.184 0.008 2-ply cubeful prune [world class]
9. Cubeful 2-ply 24/18 4/2 MWC: 41.75% ( -0.00%) 0.184 0.000 0.000 - 0.816 0.217 0.005 2-ply cubeful prune [world class]
10. Cubeful 2-ply 8/2 7/5 MWC: 41.75% ( -0.00%) 0.156 0.000 0.000 - 0.844 0.169 0.008 2-ply cubeful prune [world class]
11. Cubeful 2-ply 7/1 3/1 MWC: 41.69% ( -0.06%) 0.169 0.000 0.000 - 0.831 0.191 0.009 2-ply cubeful prune [world class]
===========================================================================
In this case, the playing strength and analysis strength are identical.
But, just in case, let's take a look at the rollouts. The first 10 go very quickly. And the 11th best move (i.e. 8/2 6/4) takes a few minutes more to roll out but actually comes on top as below: !!!
===========================================================================
1. Rollout 8/2 6/4 MWC: 41.89%
0.167 0.000 0.000 - 0.833 0.182 0.005 CL 42.90% CF 41.89%
[0.002 0.000 0.000 - 0.002 0.002 0.000 CL 0.03% CF 0.03%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
2. Rollout 24/16 MWC: 41.75% ( -0.14%)
0.214 0.000 0.000 - 0.786 0.139 0.001 CL 44.08% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
3. Rollout 24/18 8/6 MWC: 41.75% ( -0.14%)
0.201 0.000 0.000 - 0.799 0.127 0.001 CL 43.96% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
4. Rollout 24/18 7/5 MWC: 41.75% ( -0.14%)
0.200 0.000 0.000 - 0.800 0.140 0.001 CL 43.85% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
5. Rollout 24/18 6/4 MWC: 41.75% ( -0.14%)
0.208 0.000 0.000 - 0.792 0.168 0.001 CL 43.73% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
6. Rollout 24/18 3/1 MWC: 41.75% ( -0.14%)
0.191 0.000 0.000 - 0.809 0.185 0.002 CL 43.29% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
7. Rollout 24/18 5/3 MWC: 41.75% ( -0.14%)
0.181 0.000 0.000 - 0.819 0.192 0.002 CL 43.08% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
8. Rollout 8/6 7/1 MWC: 41.75% ( -0.14%)
0.168 0.000 0.000 - 0.832 0.164 0.006 CL 43.07% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
9. Rollout 24/18 4/2 MWC: 41.75% ( -0.14%)
0.189 0.000 0.000 - 0.811 0.210 0.003 CL 43.05% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
10. Rollout 7/1 3/1 MWC: 41.75% ( -0.14%)
0.176 0.000 0.000 - 0.824 0.189 0.006 CL 42.99% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
11. Rollout 7/1 6/4 MWC: 41.75% ( -0.14%)
0.173 0.000 0.000 - 0.827 0.187 0.006 CL 42.95% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
12. Rollout 8/2 7/5 MWC: 41.75% ( -0.14%)
0.162 0.000 0.000 - 0.838 0.174 0.006 CL 42.89% CF 41.75%
[0.000 0.000 0.000 - 0.000 0.000 0.000 CL 0.00% CF 0.00%]
Truncated cubeful rollout (depth 10) with var.redn.
145 games, Mersenne Twister dice gen. with seed 917523411 and quasi-random dice
Stop when std.errs. are small enough: ratio 0.01 (min. 144 games)
Play: grandmaster 3-ply cubeful prune [grandmaster]
keep the first 0 0-ply moves and up to 16 more moves within equity 0.32
Skip pruning for 1-ply moves.
keep the first 0 2-ply moves and up to 4 more moves within equity 0.08
Cube: 3-ply cubeful prune [grandmaster]
13. Cubeful 2-ply 8/2 3/1 MWC: 41.50% ( -0.39%)
0.152 0.000 0.000 - 0.848 0.197 0.009
2-ply cubeful prune [world class]
===========================================================================
There you have it. Plain and simple. In broad daylight. Actually a better proof than the one I posted the last time, since in this case the playing and analysis strengths are identical.
So, why doesn't gnubg make the best move 8/2 6/4 but instead moves 24/16...?
What kind of cocksucking motherfuckers are involved in this "gnubg project"...??
MK
mu...@compuplus.net
May 1, 2012, 7:14:46 AM5/1/12
to
This was the 33rd match of 25-points I have been playing against gnubg, using mersenne twister dice rolls with the seed corresponding to 25+match number, in this case "25033", and I was to roll 54 next... :))
MK
badgolferman
May 1, 2012, 7:29:49 AM5/1/12
to
mu...@compuplus.net wrote:
>This was the 33rd match of 25-points I have been playing against
>gnubg, using mersenne twister dice rolls with the seed corresponding
>to 25+match number, in this case "25033", and I was to roll 54
>next... :))
I too have noticed hints directing me to play a certain way which I
>This was the 33rd match of 25-points I have been playing against
>gnubg, using mersenne twister dice rolls with the seed corresponding
>to 25+match number, in this case "25033", and I was to roll 54
>next... :))
normally wouldn't, immediately followed by a fantastic roll for gnubg.
However I am hesitent to suggest gnubg set me up for the fall, knowing
what roll was coming. Too many times have I seen the impossible roll
from opponents I play on Yahoo! Backgammon that defeated me.
Having said that, what possible outcome are you hoping for by railing
against the program and trying to prove that it cheats? Are you trying
to discredit it to the point that it becomes abandoned? It's a
fantastic piece of software -- all for free. I appreciate the time
developers put into creating and refining the product. If you want to
discredit any program pick on the ones that take your money and cheat.
That's a far worse offense.
Philippe Michel
May 1, 2012, 2:23:40 PM5/1/12
to
On 2012-05-01, mu...@compuplus.net <mu...@compuplus.net> wrote:
> On Tuesday, May 1, 2012 5:11:18 AM UTC-6, mu...@compuplus.net wrote:
>> GNU Backgammon Position ID: 3psDAIBtuwIADA
>> Match ID : EQErA6AAUAAE
>> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ O: gnubg
>> | O O O O O | | O O | 10 points
>> | O O O O O | | | Rolled 62
>> | O | | |
>> | | | |
>> | | | |
>> | |BAR| |^ 25 point match
>> | | | |
>> | X X | | |
>> | X X X | | |
>> X | O X X X X | | |
>> X | O X X X X | | | 10 points
>> +24-23-22-21-20-19------18-17-16-15-14-13-+ X: Murat 25x100 (Cube: 2)
>>
>> First, think about how you would play?
>>
>> Then, think about how you think gnubg would play?
> On Tuesday, May 1, 2012 5:11:18 AM UTC-6, mu...@compuplus.net wrote:
>> GNU Backgammon Position ID: 3psDAIBtuwIADA
>> Match ID : EQErA6AAUAAE
>> +-1--2--3--4--5--6-------7--8--9-10-11-12-+ O: gnubg
>> | O O O O O | | O O | 10 points
>> | O O O O O | | | Rolled 62
>> | O | | |
>> | | | |
>> | | | |
>> | |BAR| |^ 25 point match
>> | | | |
>> | X X | | |
>> | X X X | | |
>> X | O X X X X | | |
>> X | O X X X X | | | 10 points
>> +24-23-22-21-20-19------18-17-16-15-14-13-+ X: Murat 25x100 (Cube: 2)
>>
>> First, think about how you would play?
>>
>> Then, think about how you think gnubg would play?
> Ooops, I forgot to share the secret.
>
>
> I was to roll 54 next... :)
It plays 24/16 or whatever it wishes, you double, it passes, you have a 54
opening to play. What's the problem ?
mu...@compuplus.net
May 3, 2012, 6:19:37 AM5/3/12
to
On Tuesday, May 1, 2012 5:29:49 AM UTC-6, badgolferman wrote:
> Having said that, what possible outcome are you hoping for by railing
> against the program and trying to prove that it cheats? Are you trying
> to discredit it to the point that it becomes abandoned? It's a
> fantastic piece of software -- all for free. I appreciate the time
> developers put into creating and refining the product. If you want to
> discredit any program pick on the ones that take your money and cheat.
> That's a far worse offense.
I agree on your last statement but as Patti Beadles had said a century ago, "they are a small incestuous circle"... Many of the same people who contribute to gnubg also contribute to the commercial product to some degree or another...
> Having said that, what possible outcome are you hoping for by railing
> against the program and trying to prove that it cheats? Are you trying
> to discredit it to the point that it becomes abandoned? It's a
> fantastic piece of software -- all for free. I appreciate the time
> developers put into creating and refining the product. If you want to
> discredit any program pick on the ones that take your money and cheat.
> That's a far worse offense.
The problem with the backgammon bots, unlike some music composing bot for example, is that they are all in-bred, mutated variations of the same "metal tasting", unfounded, unproven garbage...!
Freeware doesn't mean anything. We don't have to credit bullshit just because it's for free. On the contrary, we don't need to give credit to any other bullshit just because it's published as a book or a web site and costs $19.95.
In order to be useful, bg bots need to take a very sharp turn and proceed in a totally different direction.
MK
mu...@compuplus.net
May 3, 2012, 6:24:39 AM5/3/12
to
MK
0 new messages