Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Huge Bug in Fritz 9 (Engine)

Author: Uri Blass

Date: 22:25:28 10/18/05

Go up one level in this thread


On October 19, 2005 at 01:22:08, chandler yergin wrote:

>On October 19, 2005 at 01:03:11, chandler yergin wrote:
>
>>On October 18, 2005 at 18:37:24, Robert Hyatt wrote:
>>
>>>On October 18, 2005 at 16:15:01, chandler yergin wrote:
>>>
>>>>On October 18, 2005 at 15:44:02, Uri Blass wrote:
>>>>
>>>>>On October 18, 2005 at 15:09:40, chandler yergin wrote:
>>>>>
>>>>>>On October 18, 2005 at 15:02:20, Uri Blass wrote:
>>>>>>
>>>>>>>On October 18, 2005 at 14:49:41, chandler yergin wrote:
>>>>>>>
>>>>>>>>On October 18, 2005 at 12:11:56, Drexel,Michael wrote:
>>>>>>>>
>>>>>>>>>[D]1n1r3k/1p1n4/2p5/8/3B4/8/4K3/R7 b - - 0 1
>>>>>>>>>
>>>>>>>>>Analysis by Fritz 9:
>>>>>>>>>
>>>>>>>>>5...Se5 6.Lc3 b5 7.Ta7 Kg8 8.Tc7 Te8
>>>>>>>>>  ³  (-0.34)   Depth: 7/19   00:00:00  72kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Ta8 Kh7 8.Ta5 Sd3 9.Ke3
>>>>>>>>>  ³  (-0.37)   Depth: 8/19   00:00:00  169kN
>>>>>>>>>5...Se5 6.Lc5 b5 7.Le7 Te8 8.Lf6+ Kh7 9.Kd2 Sbd7 10.Th1+ Kg6
>>>>>>>>>  ³  (-0.38)   Depth: 9/28   00:00:00  490kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Tb1 Kh7 8.La1 Sa6 9.Lc3 Kg6 10.Th1
>>>>>>>>>  ³  (-0.38)   Depth: 10/26   00:00:00  1234kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Ta7 Sa6 8.Kf1 c5 9.Ke2 Kg8 10.Tb7 Sc7
>>>>>>>>>  ³  (-0.38)   Depth: 11/29   00:00:01  2285kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Th1+ Kg8 8.Tb1 Td3 9.La1 Sbd7 10.Tb4 Kf7 11.Lb2
>>>>>>>>>  ³  (-0.40)   Depth: 12/30   00:00:04  6291kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Th1+ Kg8 8.Tb1 Kf7 9.La5 Td3 10.Lc7 Sbd7 11.Tb4 Kg6
>>>>>>>>>  ³  (-0.41)   Depth: 13/40   00:00:11  15607kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Th1+ Kg8 8.Tb1 Kf7 9.Tf1+ Ke6
>>>>>>>>>  ³  (-0.43)   Depth: 14/38   00:00:30  42044kN
>>>>>>>>>5...Se5 6.Lc3 b5 7.Th1+ Kg8 8.Tb1 Kf7 9.Tf1+ Ke6
>>>>>>>>>  ³  (-0.43)   Depth: 14/46   00:01:27  121127kN
>>>>>>>>>
>>>>>>>>>This is not supposed to be the Giveaway chess engine :)
>>>>>>>>
>>>>>>>>It's not, and it's not a Bug.
>>>>>>>>Right click in your analysis window, you will get a menu
>>>>>>>>One of the selections is 'scroll main line'
>>>>>>>>If this is checkmarked you will get what you posted.
>>>>>>>>Click on 'scroll main line again', this deactivates it, and you will see the
>>>>>>>>normal PV Analysis.
>>>>>>>
>>>>>>>The problem is that Fritz9 choose Ne5 that is a stupid move that gives a piece
>>>>>>>for nothing and I do not see what it has to do with scroll main line.
>>>>>>
>>>>>>Please do as I have noted, and then Post result.
>>>>>>With Scroll activated, it continues to scroll the main PV and is not overwritten
>>>>>>with better lines. If the main PV selected at the time,
>>>>>>was Se5, it will continue to scroll that in depth.
>>>>>
>>>>>I do not have Fritz9 so I cannot do it but I expect every good program to never
>>>>>choose that move(except maybe first iteration when it still did not look at all
>>>>>legal moves) but Fritz does not show first iteration information.
>
>
>New game
>[D]1n1r3k/1p1n4/2p5/8/3B4/8/4K3/R7 b - - 0 1
>
>Analysis by Fritz 5.32:
>
>1. -+ (-4.69): 1...Kg8 2.Rh1 Re8+ 3.Kd3 Ne5+ 4.Bxe5 Rxe5 5.Ra1 Rd5+ 6.Kc2
>2. -+ (-4.63): 1...Kh7 2.Rh1+ Kg6 3.Rh8 Nf8 4.Bb6 Re8+ 5.Kd3 Nbd7 6.Bc7
>3. µ (-0.97): 1...Ne5 2.Bxe5+ Kg8 3.Ra8 Re8 4.Kd3 Rxe5 5.Rxb8+ Kf7 6.Rxb7+
>
>(, MyTown 18.10.2005)

This only proves that Fritz5.32 does not have the bug.
The bug is specific for Fritz9 engine that fails to evaluate Kg8 or Kh7 as -+

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.