Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: perft(2) -> 420

Author: Stuart Cracraft

Date: 14:52:16 02/22/06

Go up one level in this thread


On February 22, 2006 at 17:08:51, Matthias Gemuh wrote:

>>
>>At http://www.albert.nu/programs/sharper/perft.htm, the author
>>says he has verified perft(up to 10) with a variety of engines as:
>>
>>Initial position rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1
>>Depth	Perft value
>>1	20
>>2	400
>>3	8902
>>4	197281
>>5	4865609
>>6	119060324
>>7	3195901860
>>8	84998978956
>>9	2439530234167
>>10	69352859712417
>>
>>So it is in disagreement for 3 with 9322.
>>
>>When I run my perft(up to 5), I get
>>
>>1 20
>>2 420
>>3 9322
>>4 206603
>>5 5072212
>>
>>So now I am more confused. What's right? Who's wrong?
>>
>>--Stuart
>
>
>Congratulations ! Your perfts are OK.
>
>Matthias.

Thanks all.  I ran them out to perft(6) in the opening
position and it is okay (to 6 ply):

.sd 6
maxdepth = 6 maxtime = 9999999 timed = 0
.perft
perft(1) nodes visited = 20
perft(2) nodes visited = 400
perft(3) nodes visited = 8902
perft(4) nodes visited = 197281
perft(5) nodes visited = 4865609
perft(6) nodes visited = 119060324
.

At the website http://www.albert.nu/programs/sharper/perft.htm
the author says:

Position r3k2r/p1ppqpb1/bn2pnp1/3PN3/1p2P3/2N2Q1p/PPPBBPPP/R3K2R w KQkq -
This position is very good because it catches many possible bugs.
Depth	Perft value
1	48
2	2039
3	97862
4	4085603
5	193690690
6	8031647685

So I tried it and got matching numbers as far as I went (5).

This exercise has given me respect for what alpha-beta and refinements
have done to minimax.

Stuart












This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.