Author: Robert Hyatt
Date: 16:14:24 10/13/99
Go up one level in this thread
On October 13, 1999 at 18:11:10, KarinsDad wrote:
>On October 13, 1999 at 17:50:30, Robert Hyatt wrote:
>
>>On October 13, 1999 at 17:23:14, Howard Exner wrote:
>>
>>>Here's an example of pawn majority knowledge. A program
>>>with pawn majority knowledge of the queen side configuration in this
>>>example may well find the winning move Qg6+.
>>>
>>>4Q3/6pk/2pq4/3p4/1p1P3p/1P1K1P2/1PP3P1/8 b - - bm Qg6;
>>>
>>>After the exchange of queens will follow the the advance of the black king to
>>>the kingside, then exchanges of the kingside pawns, and finally the king can
>>>march over to pick up the queenside pawns. Is knowledge that white cannot break
>>>through on the queenside essential for solving this? Would any program find Qg6?
>>
>>
>>not a good example it seems. black can win tactically also... Crafty
>>finds Qg6 instantly, even on my notebook... And says that black is +2
>>from winning material after a lot of checks. the majority code I have is
>>fairly good at saying 'black has good winning chances here' but it can't
>>override winning 3 pawns or so, which Qg6+ does.
>>
>> 8 0.91 ++ 1. ... Qg6+!!
>> 8 1.47 1.86 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qxc2 5. Kg1 Qxb2 6. Kh2 Qxb3
>> 7. Qxc6
>> 8-> 1.70 1.86 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qxc2 5. Kg1 Qxb2 6. Kh2 Qxb3
>> 7. Qxc6
>> 9 3.07 1.83 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qd1+ 5. Kf2 Qxc2+ 6. Kg1 Qxb2
>> 7. Kh2 Qxb3 8. Qd8 Kg6 9. Qxh4
>> 9-> 3.22 1.83 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qd1+ 5. Kf2 Qxc2+ 6. Kg1 Qxb2
>> 7. Kh2 Qxb3 8. Qd8 Kg6 9. Qxh4
>> 10 5.81 2.00 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qd1+ 5. Kf2 Qxc2+ 6. Kg1 Qxb2
>> 7. Kh2 Qxd4 8. Qxc6 Qf4+ 9. Kh1 d4
>> 10-> 6.00 2.00 1. ... Qg6+ 2. Ke3 Qg5+ 3. Kf2 Qd2+
>> 4. Kf1 Qd1+ 5. Kf2 Qxc2+ 6. Kg1 Qxb2
>> 7. Kh2 Qxd4 8. Qxc6 Qf4+ 9. Kh1 d4
>>
>>and the score continues to slowly rise...
>
>Are you sure you supplied the proper position?
>
>1. Qg3+
>2. Ke3 Qxe8+
>
>losing the queen.
>
>KarinsDad :)
I cut and pasted the FEN. Netscape (on Linux) very lightly shows the text I
have highlighted for the 'paste' operation. I may have missed the first
character of the FEN string which would account for this...
Running the right position I get this:
11 14.84 -1.03 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kg5
4. Ke3 Kf5 5. cxb4 Kg5 6. Ke2 Kf4 7.
Kf2 g5 8. Ke2 Kg3 9. Kf1 h3 <HT>
11-> 15.71 -1.03 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kg5
4. Ke3 Kf5 5. cxb4 Kg5 6. Ke2 Kf4 7.
Kf2 g5 8. Ke2 Kg3 9. Kf1 h3 <HT>
12 19.71 -1.02 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke1 Kg3 9. Kf1
12-> 36.79 -1.02 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke1 Kg3 9. Kf1
13 43.81 -0.89 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke3 Kg3 9. b5 cxb5
which seems to be what Howard was looking for.
My error.
I _still_ hate netscape. :)
Bob
btw the score keeps getting better for black...
13-> 55.54 -0.89 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke3 Kg3 9. b5 cxb5
14 1:11 -0.79 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke1 Kf4 9. Kd2 Kg3 10.
b5 cxb5
14-> 1:31 -0.79 1. ... Qg6+ 2. Qxg6+ Kxg6 3. c3 Kf5
4. Ke3 g5 5. cxb4 g4 6. Kf2 Kf4 7.
fxg4 Kxg4 8. Ke1 Kf4 9. Kd2 Kg3 10.
b5 cxb5
but it likes it early because of the potential outside passer that the new
code recognizes... But don't think this new code solves everything yet, it
still has work to be done...
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.