Author: Thorsten Czub
Date: 08:21:50 05/30/98
The third game is interesting because of the endgame and the dramatical
situation this endgame had for the 2
operators. Ulf Lorenz operated his parallel machine P.CoNNerS, a program
using Parallel Controlled Conspiracy
Number Search. The program uses certain CONSPIRACY TARGETS and extends
them.
This program was able to search very deep, and if there would be
something to calculate, it would better handle
this. In theory, Conners would be able to see a silent move although it
is 30 plies diep - if this is a target. So in fact
Tiger (on a p2/266 Mhz) had a tough opponent. Ulf said he uses 24
machines. I don't know exactly WHICH kind
of machines.
[Event "7th IPCCC"]
[Site "Paderborn(Germany)"]
[Date "12.2.98"]
[Round "3"]
[White "P.CoNNerS"]
[Black "Chess Tiger"]
[Result "0-1"]
[ECO ""]
[Opening ""]
1.e4 e5 2.Nf3 Nf6 3.d4 Nxe4 4.Bd3 d5 5.Nxe5 Nd7 6.Qe2 Nxe5 {here Ctiger
had to think for the first time} 7.Bxe4
dxe4 8.Qxe4 Bf5
9.Qxe5+ {out of book for Conners} Qe7 10.c3 Qxe5+ 11.dxe5 Bc5 12.Nd2
0-0-0 13.Nb3 Bb6 14.0-0 Rhe8 15.Bf4
h6
16.Rad1 Be6 17.h4 a5 18.Rfe1 a4 19.Nc1 Bc5 20.b3 Be7 21.Bg3 Ba3 {this
was outplayed with a very optimistic
+0.58} 22.c4 Bb4 {+0.57}23.Rf1 Bd2 {+0.66}
24.Ne2 g5 {+0,68} 25.f3 {with this move the game turns arround. Now
black is really a little in advantage} Bf5
{+0,66} 26.Be1 Be3+ {+0.58} 27.Bf2 Rxd1 {+0,62} 28.Rxd1 Bc2 {+0,32}
29.Ra1 Rxe5 {+0,32} 30.Bxe3 axb3
{0.14}
31.axb3 Rxe3 {+0.16}32.Ra8+ Kd7 33.Nd4 Bg6 {+0.16} 34.hxg5 Rd3
{+0.48}35.Nb5 c6 {+0.58}36.gxh6 cxb5
{+1.23} 37.Ra7 Kc7 {+1.50}
38.cxb5 Rxb3 {+2,01} 39.Kf2 {i have to explain that Conners was in time
trouble. I don't remember where exactly
the time problems occured. I explained ulf that i will not allow conners
to lose on time, i would - before his time
runs out, offer him a draw, since i don't think it is fair to misuse
someones trouble into such a way. But on the other
hand i asked the other programmers if they would continue such a game,
and if this would be fair. Since all of them
agreed that they would continue the game if it would be their program, i
did so. But - i want to make sure that I
would have offered ulf a draw if it would have been nessesary. } Rb2+
{+2,12} 40.Kg1 Rxb5 {+2,17} 41.g4 Kd6
{here we joked about databases/tablebases. And ulf said he was to lazy
to implement tablebases before the
tournament. He was so busy in doing more important things. It is obvious
that murphy listened to this...} 42.Kh2 f5
43.Kg3 fxg4 44.fxg4 Ke6
45.Ra4 Kf6 {+1,82} 46.Kh4 Bh7 {+1,76} 47.Rd4 Rb1 {+1,78} 48.Rd7 Rh1+
{+3,47} 49.Kg3 Bb1 {+3,47} 50.g5+
Kxg5 {+4,92}51.Rxb7 Bg6
52.h7 Bxh7 {and there it is, the technical endgame ! Can you imagine
what happened now. Since Stefan Meyer
Kahlen was there, the author of Shredder, I asked him if he would like
to install Shredder2 on a PC next to our
desk. I remembered how Frederic Friedel called ChessBase/Hamburg in a
similar situation between Genius-Mchess
in munich 1993. He called hamburg and asked about the chances to win or
lose or draw this endgame. Now - since
technology has made progress, we only had to install shredder2 on a
machine and it told us exactly and live about
how good the moves are !} 53.Kf2 {it must have been a nightmare. Anybody
counted for the 50th move time
control, since the last capture move was 52nd move, black would have
time until move 102 to get the rook or mate}
Kf4 54.Ke2 {can you imagine how ulf must have felt. Still conners was in
time trouble. Again i was not sure what
to do. Should i continue the game or offer a draw. I tried to find out
what Christophe Theron would say if i would
have told him: i offered a draw since i felt for ulf. So - i continued
the game. But i again told ulf that when we see
that conners would run out of time i would offer a draw a few seconds
before the clock runs out of time. I am sure
this did not made him happier, but ... i thought it would only be fair.
} Rh2+ 55.Ke1 {would tiger be able to get the
rook ? Fast enough ? the cheerleaders arround the board told us which
moves have to be played, and how many
moves until the rook is lost. It went up and down. There was only a
chance for tiger to win by chance - avoiding the
3time repetition or by blunder of Conners} Ke3 {the white king is so
near at the edge, isn't there a solution for
black ?!?} 56.Rf7 Bd3 57.Re7+ Be4 58.Rf7 Ra2
59.Rd7 Rg2 60.Rf7 Rb2 61.Rd7 Re2+ 62.Kd1 Rg2 63.Kc1 Rc2+ 64.Kd1 Rc6
65.Ke1 Rc1+
66.Rd1 Rc2 67.Rd7 Rb2 68.Rd8 Rg2 69.Rf8 Rc2 70.Rd8 Rh2 71.Rf8 Re2+
72.Kd1 Rd2+
73.Ke1 Ra2 74.Rd8 Re2+ 75.Kd1 Rg2 76.Kc1 Rc2+ 77.Kd1 Rc3 78.Ke1 Bg6
79.Kf1 Rc2
80.Kg1 Be4 81.Rf8 Rg2+ 82.Kf1 Rb2 83.Kg1 Rc2 84.Rf7 Rg2+ 85.Kf1 Rg5
86.Ke1 Rc5
87.Rd7 Rh5 88.Rf7 Ra5 89.Rd7 Ra1+ 90.Rd1 Ra3 91.Kf1 Ra5 92.Re1+ Kf3
93.Rc1 Bd3+ {mate in 9}
94.Kg1 Rg5+ {mate6} 95.Kh2 Kf2 {mate4}96.Kh3 0-1
In the end P.Conners made a mistake that made it possible for black to
win shortly before the 50 move rule.
Unbelievable. It was horrible. I left so much sweat, and Ulf also, that
we both lost 4 kg or more weight.
But there was enough coke and in the evenings we ate roast potatoes with
eggs , called bratkartoffeln in a famous
pub called "lötlampe". I am sure we compensated the weigth-loss to
anybodies confidence.
So Tiger had 2.5 from 3 and my promise to win this tournament for
Christophe was still possible.
Nimzo98 was maybe the strongest program / opponent in this competition.
It made w whitewash in the netherlands
champioships. And I was sure somebody would have to stop nimzo,
otherwise nimzo will win almost any game.
But it was still to early. And maybe nobody really seriously thought
that Tiger would have chances to win the
tournament, despite myself. A reason was that I played manually
tournament games before Paderborn. And my
results were quite ok-ish !
My old chart said the following bean-counted results:
Chess Tiger 11.2 against
Opponent ELO Time Control Results
win draw loss
Hiarcs6 2532 40/120 2 0 1
Nimzo98 2530 40/120 2 2 0
Junior 4.6 ? 40/120 2 1 1
Rebel9 2533 40/120 + 60/60 1 1 1
Mchess7 2516 40/120 + 60/60 2 1 1
Shredder2 2503 40/120 2 1 1
Fritz5 ? 40/120 0 0 3
CM5000 >2387 40/120+60/60 1 1 1
Genius5 2498 40/120 0 2 1
Virtual2 ? 40/120 2 2 0
-----------------------------------------------------------------
35 games: 19.5-15.5 against ELO 2522 average 14 11 10
Ok - obviously Tiger had a problem with Fritz5 ! I don't think this had
something to do with playing strengh since
it has had no problems with nimzo and not with other strong programs.
But fritz5 did not participated in paderborn.
BTW: the elo-numbers came from ssdf-list
from november or at this time I played the games. I was very sure that
Tiger had chances to win the tournament.
Not only because of the results, but even more because the games were
very good. It is still a question to me how
Christophe was able to program that many progress into his engine from
the waterloo-result of tiger in paris to
december, january. Strange. If he would have participated with tiger
11.2 in PARIS, he would have had chances to
win there too. What a pity the progress came in the engine AFTER paris.
In the next round I had to play against Nimzo. I was in no way afraid.
My games before paderborn with the
commercial version of Nimzo98 were ok. Of course Chrilly came with a new
monster, the paderborn version,
another progress for nimzo. But - i was not afraid.
So - to be continued with the fourth round, ChessTiger 11.2 vs.
Nimzo-Paderborn...
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.