Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: TB's & Castling (Opps, I did post the wrong FEN)

Author: Slater Wold

Date: 20:16:54 03/06/02

Go up one level in this thread


On March 06, 2002 at 23:13:43, Jeremiah Penery wrote:

>On March 06, 2002 at 22:49:38, Slater Wold wrote:
>
>>On March 06, 2002 at 22:27:17, Uri Blass wrote:
>>
>>>On March 06, 2002 at 19:18:16, Slater Wold wrote:
>>>
>>>>I am 100% sure TB's do indeed take castling into consideration.
>>>
>>>
>>>No
>>>You do not understand how tablebases work.
>>>There are no moves in tablebases.
>>>
>>>The engine generates all the legal moves and looks in the tablebases after these
>>>moves to see distance to mate.
>>
>>
>>Um, well, according to Hyatt, it would tell the TB "o-o" and it wouldn't return
>>anything.  I am very well aware how TB's work.
>
>No, that's not how they work.  The engine would feed it the position after the
>move O-O was made, and the TB would return MATE in 2.  But if you give the
>position 2 plies before the castling, the TB will return MATE in 4, because it
>didn't calculate the distance to mate with castling rights in consideration.  If
>you feed Crafty the position where castling is possible, it will return the
>correct Mate in 2 score, from tablebases.
>
>White(1): new
>White(1): 8/8/8/8/8/6P1/6k1/4KR1R w K -
>White(1): analyze
>Analyze Mode: type "exit" to terminate.
>end-game phase
>              clearing hash tables
>              time surplus   0.00  time limit 30.00 (3:30)
>         nss  depth   time  score   variation (1)
>                1     0.00  21.61   1. g4
>                1->   0.00  21.61   1. g43
>                2     0.00  21.31   1. g4 Kg3
>                2->   0.00  21.31   1. g4 Kg3
>                3     0.00     ++   1. g4!!
>                3     0.00  21.84   1. g4 Kg3 2. g5
>                3     0.04  Mat06   1. Rf2+ Kxh1 <HT>
>                3     0.04  Mat05   1. Rh2+ Kxh2 <HT>
>                3     0.04  Mat04   1. Rh8 Kxg3 <HT>
>                3->   0.04  Mat04   1. Rh8 Kxg3 <HT>
>               <snip>
>              endgame tablebase-> probes done=277  successful=153
>analyze.White(1): Rf4
>White(1): Rf4
>              clearing hash tables
>              time surplus   0.00  time limit 30.00 (3:30) [easy move]
>         nss  depth   time  score   variation (1)
>                1     0.00  12.45   1. ... Kxg3
>                1->   0.00  12.45   1. ... Kxg3
>                2     0.00     --   1. ... Kxg3
>                2     0.00  Mat03   1. ... Kxg3 <HT>
>                2->   0.00  Mat03   1. ... Kxg3 <HT>
>                3     0.00  Mat03   1. ... Kxg3 <HT>
>                3->   0.00  Mat03   1. ... Kxg3 <HT>
>               <SNIP>
>              endgame tablebase-> probes done=2  successful=2
>analyze.Black(1): Kxg3
>Black(1): Kxg3
>              clearing hash tables
>              time surplus   0.00  time limit 30.00 (3:30)
>         nss  depth   time  score   variation (1)
>                1     0.00  12.45   2. O-O
>                1->   0.00  12.45   2. O-O4
>                2     0.00     ++   2. O-O!!
>                2     0.00  Mat02   2. O-O <HT>
>                2->   0.00  Mat02   2. O-O <HT>
>                3     0.00  Mat02   2. O-O <HT>
>                3->   0.00  Mat02   2. O-O <HT>
>                4     0.00  Mat02   2. O-O <HT>
>                4->   0.00  Mat02   2. O-O <HT>
>              time=0.04  cpu=0%  mat=10  n=19  fh=100%  nps=10k
>              ext-> chk=0 cap=0 pp=0 1rep=0 mate=0
>              predicted=0  nodes=19  evals=0
>              endgame tablebase-> probes done=3  successful=3
>              hashing-> trans/ref=82%  pawn=0%  used=0%
>analyze.White(2):
>
>
>>>If castling is legal then the engine looks at the tablebases to see the distance
>>>to mate after castling in order to see the mate in 2 score.
>>
>>
>>According to Hyatt, no it doesn't.
>
>Uri was right, and that's exactly what they do.  It's just that they don't take
>castling rights into consideration in the mate scores they return.  If castling
>is possible at the root, the engine will generate that move and feed the
>resulting position to the TB, which will return the now-correct mate score.


Crafty is actually putting out an eval though.  No other engine I have does.  It
just returns in 0.00 the TB results.  Crafty took .5 seconds to find the correct
move, while say Shredder, finds it in 0.00 with a 100% TB return.  There is
_zero_ evaluation from the engine.


>>The correct position is:
>>
>>[D]8/8/8/8/8/6P1/6k1/4KR1R w K -



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.