Author: Albert Silver
Date: 04:30:21 01/13/03
Go up one level in this thread
On January 12, 2003 at 21:18:25, Mike Hood wrote:
>A question out of morbid curiosity: what's the highest incorrect score that
>anyone has seen a top chess program award? ie, what's the highest winning score
>that a chess program gives itself, blind to the fact that it must unavoidably
>lose?
I don't know about losing, but Fritz 6 was known to announce non-existent mates.
Albert
>
>The highest score I have seen is Fritz 8 in the recently posted "Georgian
>Peasant" game.
>
>[D] 8/3P3k/n2K3p/2p3n1/1b4N1/2p1p1P1/8/3B4 w
>
>1. Nf6+ Kg7 2. Nh5+ Kg6 3. Bc2+ Kxh5 4. d8=Q Nf7+ 5. Ke6 Nxd8+ 6. Kf5 e2 7. Be4
>e1=N 8. Bd5 c2 9. Bc4 c1=N 10. Bb5 Nc6 11. Bxc6 Nc7 12. Ba4 Nc2 13. Bxc2 Ne2
>14. Bd1 Nd5 15. Bxe2# 1-0
>
>After 3. Bc2+, Fritz 8's winning score for Black soars ever higher:
>
>Analysis by Fritz 8:
>
>3...Kxh5 4.Bd1+ Kg6 5.Kc6 Nb8+ 6.Kd6 Nxd7 7.Kxd7 Kf5
> -+ (-7.56) Depth: 7/21 00:00:00 71kN
>3...Kxh5!
> -+ (-7.84) Depth: 8/27 00:00:00 182kN
>3...Kxh5!
> -+ (-8.13) Depth: 9/30 00:00:01 470kN
>3...Kxh5--
> -+ (-7.84) Depth: 10/23 00:00:01 489kN
>3...Kxh5 4.Bd1+ Kg6
> -+ (-7.84) Depth: 10/29 00:00:01 678kN
>3...Kxh5!
> -+ (-8.13) Depth: 11/29 00:00:03 1375kN
>3...Kxh5!
> -+ (-8.41) Depth: 12/35 00:00:08 3525kN
>3...Kxh5!
> -+ (-8.69) Depth: 13/38 00:00:22 9391kN
>3...Kxh5!
> -+ (-8.97) Depth: 13/40 00:00:34 14045kN
>3...Kxh5 4.Bd1+ Kg6 5.Kd5 Nf7 6.Kc4 Ne5+ 7.Kb3 Nxd7 8.Be2 Nc7 9.Kc2 Kf6
> -+ (-9.16) Depth: 13/40 00:00:45 18913kN
>3...Kxh5 4.Bd1+ Kg6 5.Kd5 Nf7 6.Kc4 Ne5+ 7.Kb3 Nxd7 8.Kc2 Ne5 9.Be2 Nc7 10.Kd1
>Kf5 11.Kc2
> -+ (-9.34) Depth: 14/40 00:01:27 34783kN
>3...Kxh5!
> -+ (-9.63) Depth: 15/44 00:03:03 72060kN
>3...Kxh5 4.Bd1+ Kg6 5.Kd5 Nf7 6.Kc4 Ne5+ 7.Kb3 Nxd7 8.Kc2 Ne5 9.Be2 c4 10.Bf1
>Nc5 11.Be2 Nd3
> -+ (-9.78) Depth: 16/46 00:06:14 149982kN
>3...Kxh5!
> -+ (-10.06) Depth: 17/47 00:15:11 372788kN
>
>It's not a matter of Fritz choosing the wrong move for Black. Kxh5 is correct,
>but Fritz doesn't expect White to play d8Q!
>
>I'm not just knocking Fritz. None of the chess programs in my possession do any
>better. I even tried setting Fritz's selectivity parameter to zero, to prevent
>null move generation, but found no improvement. So, does anyone have any other
>position that the current top programs totally misjudge?
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.