I can indeed reproduce it with -t =2000. uct_pass_is_safe() returns false
when there are not enough playouts. So it prefers a sure loss :-( I am
testing a fix.

Jean-loup

2011/8/5 Michael Williams <[email protected]>

> Here is the command line:
>
> ../pachi -t =2000 resign_threshold=0
>
> I thought it was failing at higher strengths as well, but now I can only
> reproduce the problem at that strength.  But it still seems like it should
> be able to solve it.
>
> Here is the SGF:
>
> boardsize 6
> komi -35.5
>
> clear_board
> play b d4
> play w d3
> play b b1
> play w c6
> play b a2
> play w f4
> play b c2
> play w a4
> play b d6
> play w c4
> play b d2
> play w c1
> play b f5
> play w c3
> play b e1
> play w e3
> play b d1
> play w f1
> play b e6
> play w a5
> play b f6
> play w f3
> play b e2
> play w b4
> play b b3
> play w d5
> play b a3
> play w b5
> play b e5
> play w b6
> play b c1
> play w e4
> play b e6
> play w d6
> play b b2
> play w a6
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
> genmove b
> genmove w
>
>
> On Fri, Aug 5, 2011 at 5:33 AM, Jean-loup Gailly <[email protected]> wrote:
>
>> pachi passes for me in this situation. It would help if you gave the .sgf
>> or .gtp file,
>> the pachi version and all the options you have used.
>>
>> Jean-loup
>>
>> 2011/8/5 Michael Williams <[email protected]>
>>
>>> Why does pachi fill it's eye here, when it can simply pass and win the
>>> game (komi = -35.5) ?
>>>
>>>
>>> IN: genmove b
>>> <pre-simulated 816 games>
>>> (UCT tree; root white; extra komi 0.000000; max depth 10)
>>> [F6] 1.000/1002 [prior 0.059/119 amaf 1.000/824 crit 0.000] h=0 c#=3
>>> <fbfb7>
>>>  [pass] 0.995/1004 [prior 0.500/14 amaf 0.000/0 crit -0.005] h=0 c#=5
>>> <fbfbd>
>>> (avg score 0.000000/0 value 0.000000/0)
>>> [186] best 0.000000 | seq                 | can pass(0.995)  A1(0.000)
>>> F1(0.000)
>>> *** WINNER is A1 (1,1) with score 0.0000 (0/1002:186/186 games), extra
>>> komi 0.000000
>>> genmove in 0.10s (1820 games/s, 606 games/s/thread)
>>> playing move A1
>>> Move:  41  Komi: -35.5  Handicap: 0  Captures B: 2 W: 8
>>>       A B C D E F        A B C D E F
>>>     +-------------+    +-------------+
>>>   6 | O O O O . O |  6 | O O O O O O |
>>>   5 | O O . O O . |  5 | O O O O O O |
>>>   4 | O O O . O O |  4 | O O O O O O |
>>>   3 | X X O O O O |  3 | X X O O O O |
>>>   2 | X X X X X X |  2 | X X X X X X |
>>>   1 | X)X X X X . |  1 | X X X X X X |
>>>     +-------------+    +-------------+
>>>
>>>
>>> _______________________________________________
>>> Computer-go mailing list
>>> [email protected]
>>> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
>>>
>>
>>
>> _______________________________________________
>> Computer-go mailing list
>> [email protected]
>> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
>>
>
>
> _______________________________________________
> Computer-go mailing list
> [email protected]
> http://dvandva.org/cgi-bin/mailman/listinfo/computer-go
>
_______________________________________________
Computer-go mailing list
[email protected]
http://dvandva.org/cgi-bin/mailman/listinfo/computer-go

Reply via email to