Author: Eelco de Groot
Date: 18:13:15 10/21/01
Go up one level in this thread
On October 21, 2001 at 19:21:18, Jeremiah Penery wrote:
>On October 21, 2001 at 19:08:02, Eelco de Groot wrote:
>
>>On October 21, 2001 at 17:57:10, Terry McCracken wrote:
>>
>>>On October 21, 2001 at 17:53:03, Andrew Smith wrote:
>>>
>>>>I hope this diagram works...
>>>>
>>>>[D]r1bk3r/1pp2pp1/p2p1n1p/5q2/8/4BN2/PPPQ1PPP/3RR1K1 w - -
>>>>
>>>>How long does it take your program to find the strong move Bb6! ?
>>
>>I think that is a nice challenge for one of Jeremiah Penery's modified Craftys
>>or a Chessmaster with extra high selectivity. But can they beat old Q5T? A
>>properly trained Rebel Q5T "T for Terminator" after having spent a few months in
>>a harsh trainingcamp in the Dutch mountains sees Bb6 in about 0.01 second. This
>>on a Celeron 500 Mhz. Approximately 1/100 of a second,that is if I assume that
>>every ply takes about three times longer than the previous one and seven plies
>>are reached in one second.
>
>>00:00 01.00 -2.32 1.Bxh6 Rxh6
>>00:00 01.03 1.71 1.Bb6
>>00:00 02.00 0.32 1.Bb6 Be6
>>00:00 02.11 0.37 1.Nd4 Qd5 2.Bf4 Qxa2
>>00:00 03.00 0.01 1.Nd4 Qd5 2.Bf4 Re8 3.Rxe8+
>>00:00 03.01 0.01 1.Bb6
>>00:00 03.01 1.75 1.Bb6 Nd5 2.Qxd5 Qxd5
>>00:00 04.00 1.75 1.Bb6 Nd5 2.Qxd5 Qxd5 3.Bxc7+
>>11:13 12.00 2.36 1.Bb6 Nd5 2.Ba5 b6
>>
>>Positions : 517.422.149
>>Nodes per second : 210.505
>
>I've been doing some strange things to Crafty 18.12, and I just finished one,
>and came here to find a couple positions to test it. :) It seems I've found a
>good one, but the search seems unstable here. :(
>
>White(1): analyze
>Analyze Mode: type "exit" to terminate.
> time surplus 0.00 time limit 30.00 (3:30)
> nss depth time score variation (1)
> 1 0.01 2.07 1. Bb6
> 1-> 0.04 2.07 1. Bb6
> 2 0.04 1.81 1. Bb6 Ne4
> 2-> 0.06 1.81 1. Bb6 Ne4
> 3 0.08 2.00 1. Bb6 Nd5 2. Nd4
> 3-> 0.11 2.00 1. Bb6 Nd5 2. Nd4
> 4 0.14 1.88 1. Bb6 Nd5 2. Re5 dxe5 3. Qxd5+
> 4-> 0.22 1.88 1. Bb6 Nd5 2. Re5 dxe5 3. Qxd5+
> 5 0.44 2.08 1. Bb6 Ne8 2. Qe3 Qf6 3. Bd4
> 5-> 0.68 2.08 1. Bb6 Ne8 2. Qe3 Qf6 3. Bd4
> 6 1.38 1.92 1. Bb6 Nd5 2. Bxc7+ Kxc7 3. c4 Qf4
> 4. Qxd5
> 6 2.63 1.98 1. Qc3 Ne8 2. Bd4 Rg8 3. Qc4 Qf4
> 6-> 3.06 1.98 1. Qc3 Ne8 2. Bd4 Rg8 3. Qc4 Qf4
> 7 5.60 -- 1. Qc3
> 7 6.37 1.57 1. Qc3 Bd7 2. Bb6 cxb6 3. Rxd6 Ne4
> 4. Qxg7 Nxd6 5. Qxh8+
> 7 9.44 ++ 1. Bb6!!
> 7 12.84 2.63 1. Bb6 Qd7 2. Ne5 dxe5 3. Qa5 cxb6
> 4. Qxb6+ Ke7 5. Rxe5+
> 7-> 13.94 2.63 1. Bb6 Qd7 2. Ne5 dxe5 3. Qa5 cxb6
> 4. Qxb6+ Ke7 5. Rxe5+
> 8 21.14 2.48 1. Bb6 Qd7 2. Qb4 Ne8 3. Ne5 Qe6 4.
> Bc5 Qf5
> 8-> 31.06 2.48 1. Bb6 Qd7 2. Qb4 Ne8 3. Ne5 Qe6 4.
> Bc5 Qf5
> 9 40.74 -- 1. Bb6
> 9 1:37 0.00 1. Bb6 cxb6 2. Ne5 Be6 3. Nc6+ bxc6
> 4. Qxd6+ Kc8 5. Qxc6+ Kb8 6. Qxb6+
> Kc8 7. Rd8+ Rxd8 8. Qc6+ Kb8 9. Qb6+
> Kc8
>
>It takes a while searching for another move, so I moved ahead with Bb6 cxb6:
>
> 6-> 2.83 4.50 2. Qxd6+ Bd7 3. Re5 Kc8 4. Rxf5 Bxf5
> 5. Qxb6
> 7 4.81 -- 2. Qxd6+
> 7 10.73 -0.05 2. Qxd6+ Bd7 3. Re5 Qh7 4. Re3 Kc8
> 5. Rc3+ Bc6 6. Rxc6+ bxc6 7. Qxc6+
> Kb8 8. Qxb6+ Kc8
> 7 14.07 0.00 2. Ne5 Be6 3. Nc6+ bxc6 4. Qxd6+ Kc8
> 5. Qxc6+ Kb8 6. Qxb6+ Kc8 7. Rd8+ Rxd8
> 8. Qc6+ Kb8 9. Qb6+ Kc8
> 7-> 21.38 0.00 2. Ne5 Be6 3. Nc6+ bxc6 4. Qxd6+ Kc8
> 5. Qxc6+ Kb8 6. Qxb6+ Kc8 7. Rd8+ Rxd8
> 8. Qc6+ Kb8 9. Qb6+ Kc8
> 8 26.22 -0.34 2. Ne5 Ne8 3. Nc4 Qg6 4. Nxb6 Bh3 5.
> f4 Rc8
> 8 26.72 ++ 2. Qxd6+!!
> 8 51.42 1.31 2. Qxd6+ Bd7 3. Re5 Qxc2 4. Qxb6+ Qc7
> 5. Qd4 Kc8 6. Rc5 Bc6
> 8-> 1:13 1.31 2. Qxd6+ Bd7 3. Re5 Qxc2 4. Qxb6+ Qc7
> 5. Qd4 Kc8 6. Rc5 Bc6
> 9 1:53 1.66 2. Qxd6+ Bd7 3. Re5 Qxc2 4. Qxb6+ Qc7
> 5. Qd4 Kc8 6. Rc5 Bc6 7. Rdc1
>
>I will have to work out some search instabilities. :(
Hi Jeremiah,
Just decrease the value of Bishops, that should do it, or the value of queen
near enemy King a bit higher, that's a bit less crude.
When I am looking at that position it seems that the pawn on c7 is the whole
basis of Black's king safety against that battery of rooks and Queen. It is
defended only by the King so to human eyes the position seems pretty dangerous
for Black.
That new Crafty 18.12 is interesting, I hear Prof. Hyatt has implemented some
new eval code and some deeper extensions in the way Deep Blue did things.
I hope I can quote some text from the Crafty-email list, dangerous business I
know quoting e-mails, but maybe they throw me in a cell with Tina and Terry and
Steven and... Then I trust Michael will bail us out then so he can continue
arguing how we don't understand anything he's been saying, which is not so far
from the truth, but eh about Crafty 18.12 Robert says:
"I have a new version ready to test. It is a bit experimental, with a
change in the way search extensions are limited.
The previous versions simply clamped the extensions to a max of one
ply of extension per ply of search.
The current version uses the deep blue approach of allowing two plies
of extension for two consecutive plies of search. On average, it would
be about the same of course, but in some cases the new way is a bit more
relaxed and will search deeper along some lines. I have tested it and it
is working fine. The question is, is this better than the old version.
I am going to release the source for version 18.12, to let others play
with this to see what they think. It may "stay" or I may revert to the
pre 18.12 extension limit depending on the comments/results...
Let me know how this looks..."
Just by accident Q5T also plays Kxb3 in the Arasan-Postmodernist endgame Jon
Dart posted. What's that? It doesn't give a piece away on move one? Surely then
on move two, yes, of course, 2.Be4! Kxe4, there it is! +1.83 at ply 17, very
convincing it is not.
Thinking....
+---+---+---+---+---+---+---+---+ -
8 | | | | | | | | |
+---+---+---+---+---+---+---+---+ 00:00:00 00:00:00
7 | | | | | | | p | |
+---+---+---+---+---+---+---+---+
6 | | | B | | | *N| | | Time
+---+---+---+---+---+---+---+---+ Hash 13 Mb
5 | *p| | | | | p | | | Engine Rebel Century 2.0
+---+---+---+---+---+---+---+---+ Personality "Q5T"
4 | p | | | *K| | | | |
+---+---+---+---+---+---+---+---+
3 | | *p| *p| | | | | |
+---+---+---+---+---+---+---+---+
2 | | K | | | | | | |
+---+---+---+---+---+---+---+---+
1 | | | | | | | | |
+---+---+---+---+---+---+---+---+
a b c d e f g h
[D]8/6P1/2B2n2/p4P2/P2k4/1pp5/1K6/8 w - -
[Event ""]
[Site ""]
[Date "-9928.-54.-20"]
[Round ""]
[White ""]
[Black ""]
[Result "*"]
[BlackElo ""]
[WhiteElo ""]
[FEN "8/6P1/2B2n2/p4P2/P2k4/1pp5/1K6/8 w - - 0 1"]
1.*
00:00 01.05 1.77 1.Kxb3
00:00 02.00 -2.05 1.Kxb3 Kd3
00:00 03.00 1.12 1.Kxb3 Kd3 2.Bd5 Nxd5 3.g8=Q
00:00 04.00 1.12 1.Kxb3 Kd3 2.Bd5 Nxd5 3.g8=Q
00:00 05.00 1.12 1.Kxb3 Kd3 2.Bd5 Nxd5 3.g8=Q
00:00 06.00 -0.18 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kd4 (0)
00:00 07.00 -0.14 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kd4 Ke6 (0)
00:00 08.00 -0.27 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kc4 Ke5 5.Kd3 (0)
00:00 09.00 -0.80 1.Kxb3 Kd3 2.Bd5 c2 3.g8=Q Nxg8
4.Bc4+ (0)
00:00 10.00 1.99 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kc4+ Kg6 5.Kb5 Kxg7 6.Kxa5 (0)
00:01 11.00 1.68 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kc4 Kg6 5.Kb5 Kxg7 6.Kxa5 Nd5 (1)
00:02 12.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kd5
4.Kc2 Ke4 5.Kc3 Kd5 6.Kc2 Ke4 (2)
00:06 13.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kd5
4.Kc2+ Ke4 5.Kb3 Kd4 6.Kc2 Ke4 7.Kb3
Kd4 8.Kc2 (5)
00:12 14.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kd5
4.Kd3 Ng8 5.Kc3 Nf6 6.Kd3 Ng8 7.Kc3 (8)
00:23 15.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kd5+
4.Kd3 Ng8 5.Kc3 Nf6 6.Kd3 Ng8 7.Kc3 (16)
00:51 16.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4
02:38 17.00 1.83 1.Kxb3 Kd3 2.Be4+ Kxe4
07:09 18.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Ng8
4.Kc4 Nh6 5.Kb5 Nxf5 (286)
24:28 19.00 0.00 1.Kxb3 Kd3 2.Be4+ Kxe4 3.Kxc3 Kxf5
4.Kd4 (888)
Positions : 581.927.875
Nodes per second : 197.665
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.