Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: What is the reason that crafty needs 10 plies to find Nf5

Author: blass uri

Date: 14:53:19 04/29/00

Go up one level in this thread


On April 29, 2000 at 14:05:23, Robert Hyatt wrote:

>On April 29, 2000 at 13:29:11, blass uri wrote:
>
>>This position is from the game insomniac0.63-chessmaster6000
>>
>>[D]6k1/pp2nppp/7r/8/3q4/1P1p1P2/P2P1RP1/RQB3K1 b - - 0 1
>>
>>Crafty17.10 needs 10 plies to find Nf5 and likes Qh4 with a draw evaluation at
>>depth 9 when after Nf5 it needs only 5 plies to see that white is losing.
>>
>>I know that crafty is a null mover but I thought that null movers do not prune
>>more than 3 plies.
>>
>>Uri
>
>
>I use R=3 and R=2 depending on depth.  But I do it recursively so it is
>possible to lose 5-6 plies in wild sacrificial lines.
>
>In this position, it only takes me 8 seconds to find Nf5 even though the depth
>is 10. Score is +13.  By 40 seconds and depth=12, score=+20.
>
>                9->   1.69   0.00   1. ... Qh4 2. Rf1 Qd4+ 3. Rf2
>               10     2.32   0.00   1. ... Qh4 2. Rf1 Qd4+ 3. Rf2
>               10     8.24     ++   1. ... Nf5!!
>               10    36.03  13.18   1. ... Nf5 2. Qc2 dxc2 3. d3 Ng3 4.
>                                    Bxh6 Ne2+ 5. Kf1 Qxa1+ 6. Kxe2 Qd1+
>                                    7. Ke3 c1=Q+ 8. Kd4 Qxh6 9. Kd5 Qxd3+
>                                    10. Kc5 <HT>
>               10->  36.04  13.18   1. ... Nf5 2. Qc2 dxc2 3. d3 Ng3 4.
>                                    Bxh6 Ne2+ 5. Kf1 Qxa1+ 6. Kxe2 Qd1+
>                                    7. Ke3 c1=Q+ 8. Kd4 Qxh6 9. Kd5 Qxd3+
>                                    10. Kc5 <HT>
>               11    36.20     ++   1. ... Nf5!!
>         (2)   11->  41.58  13.57   1. ... Nf5 2. Qc2 dxc2 3. d3 Ng3 4.
>                                    Bxh6 Ne2+ 5. Kf1 Qxa1+ 6. Kxe2 Qd1+
>                                    7. Ke3 c1=Q+ 8. Kd4 Qxh6 9. Kd5 Qxd3+
>                                    10. Kc5 <HT>
>               12    41.85     ++   1. ... Nf5!!
>               12    49.16  20.07   1. ... Nf5 2. Qc2 dxc2 3. d3 Ng3 4.
>                                    Bxh6 Ne2+ 5. Kf1 Qxa1+ 6. Kxe2 Qd1+
>                                    7. Ke3 c1=Q+ 8. Kd4 Qxh6 9. Ke5 Qe3+
>                                    10. Kd6 Qdxd3+ 11. Kc7 Qxf2 12. Kxb7
>                                    <HT>
>
>
>In this case, I don't mind taking 8 seconds to find this...

The problem is that crafty needs longer time to avoid 26.exd4 from the same
game:

[D]6k1/pp2nppp/1q5r/8/3r4/1P1pP3/P2P1PP1/RQB2RK1 w - - 0 1

Crafty needs even longer time to see the previous good sacrifice of chessmaster
from the same game.


I think that crafty prunes Nf5 because of null move pruning and Nf5 should be
extended because it threats mate in 2.

I think that crafty can call chest to recognize mate threats(you can analyze
mate threats to bigger depth then regular threats and every time you use crafty
to see if there is a threat of d plies to use chest to discover if there is a
mate threat of d+2 plies and to extend the move if chest discovers a mate
threat)

I guess that you do not lose a lot of time by this idea because chest is a fast
mate solver.

Uri



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.