Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Why can't Crafty 18.01 solve these, but 17.14 can!?

Author: Robert Hyatt

Date: 14:30:37 02/12/01

Go up one level in this thread


On February 12, 2001 at 01:26:24, Jouni Uski wrote:

>I use Crafty mostly to endgame analysis, where it's very good specially in
>pawn/rook endings. After upgrading to 18.01 (Chessbase) I was a little
>disappointed to find some positions, which it can't solve, but 17.14 can.
>Here's 3 examples:
>
>[D]5k2/p1p4R/1pr5/3p1pP1/P2P1P2/2P2K2/8/8 w - -
>[D]k7/8/2p5/4P3/4B3/6p1/8/2n3K1 w - -
>[D]8/2p5/p1P1p3/P3kb2/1P4pp/8/5R2/7K w - -
>
>            17.14    18.01
>1. Kg3      84s      >600s
>2. Bd3      86s      >600s
>3. Rxf5      1s      >600s
>
>I tested in AMD450Mhz and 116MB hash.
>
>Jouni



Do you have any position.bin files lying around?  If so delete them.  I
just ran this and got the following on my quad:

         (2)   10->   0.72   0.47   1. Ke2 Rxc3 2. Rd7 Rc2+ 3. Kd3 Rf2
                                    4. Ke3 Ra2 5. Rxd5 Ra3+ 6. Kd2 Ra2+
                                    7. Kc3 Rxa4 8. Rxf5+ Ke7 9. Rf6 <HT>
               11     0.92   0.56   1. Ke2 Rxc3 2. Rd7 Rc2+ 3. Kd3 Rf2
                                    4. Ke3 Ra2 5. Rxd5 Ra3+ 6. Ke2 <HT>
               11->   1.40   0.56   1. Ke2 Rxc3 2. Rd7 Rc2+ 3. Kd3 Rf2
                                    4. Ke3 Ra2 5. Rxd5 Ra3+ 6. Ke2 <HT>
               12     2.10   0.47   1. Ke2 Rxc3 2. Rd7 Rc2+ 3. Kd3 Rf2
                                    4. Ke3 Ra2 5. Rxd5 Ra3+ 6. Ke2 Ra2+
                                    7. Kf3 Ra3+ 8. Kf2 Ke7 9. Rxf5 Rxa4
               12     4.84   0.82   1. Rd7 Rxc3+ 2. Ke2 Ra3 3. Rxd5 a5
                                    4. Rxf5+ Kg7 5. d5 Rxa4 6. Ke3 b5 7.
                                    Re5 Kg6
         (2)   12->   4.84   0.82   1. Rd7 Rxc3+ 2. Ke2 Ra3 3. Rxd5 a5
                                    4. Rxf5+ Kg7 5. d5 Rxa4 6. Ke3 b5 7.
                                    Re5 Kg6
               13     7.72   1.05   1. Rd7 Rxc3+ 2. Ke2 Ra3 3. Rxd5 Kg7
                                    4. Rxf5 Rxa4 5. Kd3 Rb4 6. Re5 c5 7.
                                    Re7+ Kg6 8. Rd7
               13->  10.82   1.05   1. Rd7 Rxc3+ 2. Ke2 Ra3 3. Rxd5 Kg7
                                    4. Rxf5 Rxa4 5. Kd3 Rb4 6. Re5 c5 7.
                                    Re7+ Kg6 8. Rd7
               14    18.48   0.81   1. Rd7 Rxc3+ 2. Ke2 c6 3. Rxa7 Rc4
                                    4. Kd3 c5 5. dxc5 bxc5 6. g6 Rxf4 7.
                                    Rf7+ Ke8 8. a5 Rf3+ 9. Ke2
                                    Rf7+ Ke8 8. a5 Rf3+ 9. Ke2
               14    38.21   0.94   1. Kg3 Rxc3+ 2. Kh4 Rd3 3. g6 Rxd4
                                    4. Kg5 Rc4 5. Kf6 Rc6+ 6. Kxf5 Rc4
                                    7. Rf7+ Kg8 8. Rd7 d4 9. Rd8+ Kg7 10.
                                    Rd7+ Kh6
         (2)   14->  38.21   0.94   1. Kg3 Rxc3+ 2. Kh4 Rd3 3. g6 Rxd4
                                    4. Kg5 Rc4 5. Kf6 Rc6+ 6. Kxf5 Rc4
                                    7. Rf7+ Kg8 8. Rd7 d4 9. Rd8+ Kg7 10.
                                    Rd7+ Kh6
               15    47.05   1.02   1. Kg3 Rxc3+ 2. Kh4 Rd3 3. g6 Rxd4
                                    4. Kg5 Rc4 5. Kf6 Rc6+ 6. Kxf5 d4 7.
                                    Rd7 Rc4 <HT>
         (4)   15->   1:17   1.02   1. Kg3 Rxc3+ 2. Kh4 Rd3 3. g6 Rxd4
                                    4. Kg5 Rc4 5. Kf6 Rc6+ 6. Kxf5 d4 7.
                                    Rd7 Rc4 <HT>

IE 38 seconds to go back to Kg3 and it sticks there.

Position 2:

         (2)   16    43.63   1.43   1. e6 Ne2+ 2. Kh1 g2+ 3. Bxg2 Ng3+
                                    4. Kh2 Nf5 5. Kh3 Ne7 6. Kh4 Kb7 7.
                                    Kg5 Kc7 8. Kf6 Kd6 9. Bf3 Ng8+ 10.
                                    Kf5 c5
               16    52.34     ++   1. Bd3!!

I stopped it there...

Position 3:

Thru one minute it likes b5, which probably won't change on your machine
as it is significantly slower.  at 1:11 b5 fails low and it moves on to
the right move.  This might be a result of the removed recapture extension,
looking at the PV.  I am sure it plays better with this off, overall.  I
tested for thousands of games to compare.



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.