Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: How long before your program avoids RXE8

Author: blass uri

Date: 01:26:37 03/07/99

Go up one level in this thread



On March 07, 1999 at 02:08:34, Dan Kiski wrote:


>"It was a shocker because it looks like it simply loses to 24.Rxe8 Qxe8 25.Bxf6
>leaving White with a material advantage and four active minor pieces. But the
>Ukrainian wizard had set a wicked trap! Play could continue 25...Bb4 26.Qd1 Qe3+
>27.Kh1 Qf4! and the f6 bishop is trapped!! YOW! Or 26.Qf2 Be1 27.Qf1 Qe3+ 28.Kh1
>Bh4! 29.Bxh4 Qxd4 30.Bf2 Qd2 31.Nc5 d4!! 32.Nxb7 Rxc2 winning! Fritz 5.32 takes
>over six minutes on my machine to realize that 24.Rxe8? is a loser and takes
>another 10 before it prefers Kasparov’s 24.Re2. (Junior 5 is a bit quicker.)"
>Quote from Mig.
>
>Most comps choose RxE8 in the game Kasparov played RE2. In takes my Fritz 5.32
>over 20 mins to choose RE2, mig at TWIC claims his takes only 10 mins could be
>he has a much faster computer.
>I am interested in seeing other program times and computers.

Junior5 needs 8 minutes and 33 seconds to find Re2 at iteration 16 after
searching 66123 Knodes

Junior5.3 needs 9 minutes and 9 seconds to find Re2 at iteration 16 after 73089
Knodes.

on the other hand Junior5.3 can finish iteration 15 after 3 minutes and 16
seconds when Junior5 needs 3 minutes and 24 seconds to finish iteration 15.

This is probably the reson that Junior5.3 can find the move if I give it time
control of 1 hour per 10 move+2 hours per 20 moves when Junior5 cannot find Re2
with the same time control.

Junior5 and Junior5.3 do not see that Rxe8 is losing and reject it for
positional reasons.

Junior5's evaluation is 0.05 advantage for white at depth 16  and reject it
because Re2 gives white 0.19 pawn advantage and Junior5.3 has better positional
understanding and evaluates Re2 as 0.13 advantage for black at iteration 16

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.