TalkChess.com

If your iteration depth is 10 and you want to store a value into the TT at ply 9, you should store it with depth = 1 because you searched that position only one ply deep. If you would really store depth = 10 in this scenario, it is no wonder that your engine does not find the win earlier, because a...

So I think I might have a bug in search... or maybe some nasty bug in repetition detection... hard to say, there can be 1000 reasons for that... Don't forget that the repetition draw curiously ends up as the root PV. This could point to some error with the way your hash store and hash probe handles...

Ok here's my hash probing code: const BitTransEntry *te = trans.entry( h ); if ( te ) { if ( (te->type & 3) && te->depth >= depth ) { // potential trans_table hit! i32 val = te->getValue( ply ); switch( te->type & 3 ) { case bttFull: se.move.packed = te->move.packe...

There is a simple way to find out if you have a TT bug. 1. Write a second hash calculator that computes the entire hash from scratch. Then when you incrementally update your Zobrist hash in debug mode, also update the full hash. Compare both hash values. If they do not agree, then you have a hash c...

Hi all, I have a strange problem with my chess engine. I'm pretty sure I've a bug somewhere, either in the search or in the transposition table code. I've no clue so I wanted to ask if someone ever ran into a similar problem. I've a position (b3r1k1/2P4p/p2K2p1/8/p2N4/R3Pp2/8/8 w - - 0 52). When I ...

No I do the draw test before I even probe the TT. When I detect a draw I simply return 0. But of course it eventually gets stored in the parent node anyway. I wonder if other engines use some magic to handle draws by repetition (esp. when using TTs)? The strange thing is that even if I disable TT c...

Thanks for reply the PV prior to depth 14 is d6d7 g8f7 d4f3 a8f3 a3a4 e8e7 d7d8 e7e8 d8d7, score cp 0 (no change since depth 9) depth 14: d6d7 e8f8 c7c8q f8c8 d4f3 c8c5 f3d4 c5c4 d7e6 a8b7 e6e5 c4c5 e5d6 c5a5 a3c3, score cp -255 1-meg version: depth 9-11 d6d7 g8f7 d4f3 a8f3 a3a4 e8e7 d7d8 e7e8 d8d7...

Thanks for reply the PV prior to depth 14 is d6d7 g8f7 d4f3 a8f3 a3a4 e8e7 d7d8 e7e8 d8d7, score cp 0 (no change since depth 9) depth 14: d6d7 e8f8 c7c8q f8c8 d4f3 c8c5 f3d4 c5c4 d7e6 a8b7 e6e5 c4c5 e5d6 c5a5 a3c3, score cp -255 1-meg version: depth 9-11 d6d7 g8f7 d4f3 a8f3 a3a4 e8e7 d7d8 e7e8 d8d7,...

Hi all, I have a strange problem with my chess engine. I'm pretty sure I've a bug somewhere, either in the search or in the transposition table code. I've no clue so I wanted to ask if someone ever ran into a similar problem. I've a position (b3r1k1/2P4p/p2K2p1/8/p2N4/R3Pp2/8/8 w - - 0 52). When I l...