Marijus Kulvietis: The tournament is going faster and faster and we have to be quicker.
Last round we were little sad about long boring draws.
Round n9 had only 2 draws!
Such sudden changes...how,why?
GM B.Annakov: Game number 1. GENIUS-EUGEN7.92 1-0.
Many results depend on pairings. I'm sure this round was such were many draws were impossible.
Many pairings clearly indicated that some program must win,because it was simply stronger or
some sharp playing program was paired vs rather weak and silent opponent....
GM A.Kharlov: And Your first announced result is a good illustration for those words.
GENIUS showed really bad in ODYSSEY. Nevertheless EUGEN showed itself as a clearly weak program making
too many mistakes and if GENIUS had not won that game he would have shown itself totally powerless.
Especially playing in white colors.
But everything was going VERY LOGICALL.
GENIUS wasnt perfoming the best chess. But EUGEN was too weak and making too many mistakes.
I would say this program had lost which had made more mistakes in this game.
One important moment i see in this game-programs still push g pawns bravely for a short attack not thinkink about naked king
and dangerous results in future.
/EUGEN had pushed g5 pawn after a short castling/.
I guess we may remind programs the simple rule-which we use to say to children-beginners:
PLEASE-REMEMBER-PAWNS ARE NEVER MOVING BACK.
GM B.Annakov: Great pity,but in this case- both sides had made mistakes and the winner was not the best chessplayer-but the participant,which had made less mistakes. And maybe was more lucky.
Marijus Kulvietis: I see. Game n2: LITTLE GOLIATH2000-SOCRATES. 1-0
Was it also victory based on mistakes:):)?
GM B.Annakov: We can speak about weak game of SOCRATES. Bad opening and positional game.
GM A.Kharlov: Perhaps we cant speak about positional game while analysing the computer's game. Nothing's bad here.
Machine can't think like a human:
"This position looks good, this may rise some problems in future, this may lead to a bad endgame..."
Computers are counting variations, they have databases-they seek for comfortable schemes which are familiar for them!
IM Oleg Krivonosov: This game went to a rook endgame. And this is a field where computers are really weak.
Yes both sides had made many mistakes as well as in a previous one. But we must congratulate LITTLE GOLIATH.
It had deserved winning. In a typically human ending it had played rather well and had realized it advantage.
The endgame compensates previous mistakes made by LITTLE GOLLIATH in that game.
Marijus Kulvietis: Ok. Lets say deserved win by Little Golliath.
Game number 3: WCHESS2000-COMET B36. 0-1
IM Vaidas Sakalauskas: Long game played till mate:)
GM A.Kharlov: Bad example from those borring
games where everything was clear after a half of a game.
Black had already won. But whites had prefered to play till mate-thus we have nearly 100moves.
But the audience musnt worry. The game went only 40 moves maximum. After this You may stop analysing as everything is too clear even for beginner player.
IM Vaidas Sakalauskas: The game was started as a sicilian. And a variation very familiar
to Lithuanian masters of older generation. Our masters had fashion to play sicilian with Qc7 after our
GM Kveinys had renewed this old variation with some novelties in Manila olympics and other famous events.
Of course their ideas were to have difficult to attack black position. Little similarities in that program duel.
Cant even understand how white have given so easy all the position and victory so fast.
Just after opening went irregular white have given all they could-position, king's weakness,
Black were allowed simply to pick the pawns and win by points as it tols in chessworld.
And after terrible middlegame we can stop analyzing as it is clear black has just to trade dangerous pieces, go to endgame and realize 2 pawns
IM Oleg Krivonosov: And so it happened. Difficult to believe why so fast white simply
had given the win. All black had to do just to take it.
The picture as in some bought games where one side is simply making very clear mistakes.
Sure nothing to analyze -how a program is playing rook endgame with 2 pawns advantage -later even 3.
And WCHESS was playing till mate. Well it has right to play till mate:)
But as see programs dont make such mistakes as a stalemate and etc. Thus i would resign:)
Marijus Kulvietis: Ok .then what about first draw:)Game n4: GROMIT-ChessSystemTal. 1/2-1/2.
GM B.Annakov : Again we have a game with more then 100 moves.
But we cant say it was totally uninteresting. The start was interesting, because programs have run out of theory rather early. Also the game turned into a rather interesting middlegame.
Where the game was difficult to predict and to analyze.
I think black had some advantage and at least more comfortable game.
But again some not correct moves turned the game to the drawish endgame
And finally the programs went to a rook endgame and then even to a endgame with different colors
Bishops. Then it became dead draw. Nevertheless programs continued playing and thus we have 100+ moves while the draw was clear already in some 70+ move.
GM A.Kharlov: Both programs have made mistakes. Not terrible , but important.
I agree-TAL had interesting middlegame with many possible tactical lines. I believe some win was possible here for black. But several bad moves were made showing typical computer's style and no positional understanding. A result- going to a clear draw endgame.
Of course when black finally had exchanged their rook and left in the endgame with different colors bishops-the draw was clear and programs had to agree it at once.
But already earlier trades and turning to the endgame was a way to a draw .
While once more I saw interesting game in a case black would continue tension not trading pieces and searching for more original decisions.
Marijus Kulvietis: Thanks. I understand that ChessSystem Tal had missed some possible chances for a win. Sport is sport:
Game n5: ZARKOV-HIARCS. 1-0.
GM A.Kharlov: One more round I must see how the nice program HIARCS is loosing.
We work with this good program and here it looks terrible.
Seems finally we can clearly confirm that this program we use for analysis-is not prepared for computers
Tournaments or may be not well prepared for Odyssey event. But it really plays very weak .
Now it had lost to Zarkov in a game after 48 moves. I think many GMs who use HIARCS for preparations and analysis would hardly believe to such fact.
IM Vaidas Sakalauskas: It was a battle till the end. Zarkov didn't show a miracle-but it had used
Properly the mistakes of Hiarcs. After slavian defence was played Zarkov played more creative and
At least escaping too clear mistakes. Some moves and novelties were doubtful, but no clear bad moves.
While HIARCS had missed a tactical strike at 19 move . Strange as it wasn't some positional mistake. But the situation where computer's seem to be strong. Just simple tactical counting.
After 20.Nxd7 Kxd7 21.axb cxb 22.Bf3 Bc8- Hiarcs was lost already.23Re1-tied move -last killing move.
Difficult to say how such solid program as HIARCS can miss such fast loosing episodes.
Rest of the game was simply the show how Zarkov had realized clearly win position.
Of course such perfomance clearly shows that HIARCS is not prepared for the tournament.
Marijus Kulvietis: Great pity-I also use HIARCS program for my correspondence game analysis.
Ok. Game n6: VIRTUAL CHESS2-REBEL-TIGER 14,6. 1/2-1/2.
IM Oleg Krivonosov: One of those rare draw results . While REBEL-TIGER with it's ambitions must fight for every half point trying to show how it can win!
The game was rather original with not traditional opening and strange novelties. I would say it was seen from TIGER's game some efforts to be active in an early middlegame. White were also trying some activity ,but after greater trades after a 32 move the position seemed to me draw. Where both human players would agree fast as position was really drawish and some forced complications were dangerous to both sides.
IM Vaidas Sakalauskas: But programs are programs. They had taken risk to play.
And they had played till the nearly 100 moves trading pieces, gong to an endgame ,but finally
Both sides had only proved that Oleg is right:It was really a draw already after 32 move and no exchanges had given no chances for some change. Both sides haven't made any superterrible mistake and the draw result was confirmed. We cant accuse TIGER very much-it was hopeless very early to risk for any win . The game was strange-but Virtual Chess have run into an early draw.
Marijus Kulvietis: Game n7: GANDALF 4.32-CRAFTY 18.12. 1-0.
IM Vaidas Sakalauskas : An important moment for us. As our team prepares to play
A match vs GANDALF program. And as we see it had killed a powerful CRAFTY.
GM A.Kharlov:Lets see HOW it had killed that "powerful CRAFTY".
You may remember that GANDALF played e4-e5-Nf3-Nc6-Bb4.
It means GANDALF denies Ruy Lopez and playes Italian game. So You guys may prepare
Two knights defence or Italian game if you play e4-e5. But please don't see the game GANDALF-CRAFTY game as an example: It was terrible game by CRAFTY.
GM B.Annakov: Really we cant recognize CRAFTY here. It had played so poor!
We cant believe CRAFTY doesn't know how to answer Italian game. It had played so passive playing such opening decisions as h6 and 0-0, giving all initiative to WHITES from the first opening moves and then GANDALF had overplayed CRAFTY so easy. It was even a surprise for me.
IM Oleg Krivonosov: Really. GANDALF just calculated better and having better position from the first stages it had overplayed CRAFTY simply as a much better program.
Really a surprise having in mind that CRAFTY is a well known program.
IM Vaidas Sakalauskas: All the game had ended practically the 20 moves.
And Gandalf played much better. The rest after 20 moves was simply killing the wounded opponent.
But our supporters may not worry-GANDALF calculated good and had overplayed CRAFTY clearly in all moments. But we are not going to play in a manner CRAFTY did:
Marijus Kulvietis: I hope so: Game n8: PATZER-SHREDDER4. 0-1
IM Oleg Krivonosov: Very nice and original Sicilian game. Patzer was solid enough.
And Shredder tried to show it's tactician's skills.
Perhaps the most interesting moment was when SHREDDER has made a positional sacrifice of a pawn.
IM Kenneth Frey pays attention at this moment and we must.
Program had made a positional sacrifice!
GM B.Annakov: The most magnificent moment in that game. And really very important in discussing about the computer's chess. But Annotator had written this sac with signs "?!" .
As this sacrifice was rather doubtful. We may guess maybe simply SHREDDER was calculating bad some tactics….?
IM Oleg Krivonosov: Nevertheless SHREDDER tried it's best and played interesting.
The sac was doubtful, but the ambitions of SHREDDER to fight for a win were succesfull.
Black were active and played all the game searching for a win lines and situations.
Of course SHREDDER is not so powerful as it's new brother SHREDDER5 and it has made some not the best decisions in a game vs PATZER. Nevertheless it's ambitions had ended well.
Of course the victory was possible to reach in some faster and better ways.
Of course not the best game by PATZER.
A deserved win by SHREDDER4.
Marijus Kulvietis: Thanks. A game n9: MCHESS8-YACE BERLIN. 0-1.
IM Oleg Krivonosov: We have a series of games now where black had won.
This time we have a case when white had helped black.
Playing the opening not the best turning from strong normal position to rather doubtful and not correct.
Black had got a good chances for an equal and difficult to predict fight after an opening.
IM Vaidas Sakalauskas : White has played as iit is played against a much weaker opponent .
They were playing too bravely, too risky and sure not correct. One of those episodes when we can say
That white had lost rather then black had won. Of course we can say compliments to YACE as it managed to use the mistakes and not correct play of WHITES.
Sure the game was decided at the early stadium after the opening already-when white have lost position in an early middlegame. Nevertheless Yace had to be solid till the very end in order to use the received advantage.
Marijus Kulvietis: Game n10: DEEP FRITZ-NIMZO8. 0-1
GM A.Kharlov: A pairing with a great intrigue. Two titans-both solid and famous and both possible to show a nice tactical chess with the best computer chess features. And the game I think was really GMs game.
GM B.Annakov: Perfect calculation. After previous games we cant see here any terrible mistakes.
Both sides had really perfomed a good chess and after solid battle both sides principally had made a draw.
GM A.Kharlov: Both programs were really very equal partners. And difficult to give advantage to one of them. But game is game. Draw was nearly sure after a middlegame-just white had slipped a little. They wanted to trade black's queen as a dangerous piece I understand it so.
As a result white had run into a worse endgame. And as both sides were very similar it was easy to predict that the only mistaken decision of one side will be used by another side.
So it happened. The chance was given to NIMZO in a drawish finish and it had used this chance.
GM B.Annakov: It happened as it happens in a human's game: After one mistaken decision
FRITZ had started playing the endgame worse and worse.
You may be surprised why after so solid game FRITZ has screwed up the game so fast in the finish.
People often are starting some desperate decisions after the first mistake.
Program may start trying to return some chances .Seems FRITZ was feeling the endgame is lost and started searching for some fast chances. But those attempts had given opposite result-
After the first big mistake FRITZ's mistakes were made faster and faster and it had lost very fast.
GM A.Kharlov: It seems as after queen's trade FRITZ has resigned and continued the game just for fun:Nimzo must be happy after FRITZ has given a win in the endgame so easy after a such solid and equal game ,played earlier.
Marijus Kulvietis: Thanks. Ok then-another win by BLACKS:
Game n11: JUNIOR 7-ZCHESS2.2 .0-1
GM B.Annakov: Rather surprising result. JUNIOR 7-rather new ,well advertised and really strong program had lost to ZCHESS which is not known to wide chess audience.
GM A.Kharlov: The more surprising is the fact that JUNIOR had lost in theory.
Difficult to believe-but such new and solid program as JUNIOR 7 had lost the opening
And played positionally rather weak. ZCHESS thus had a perfect chance for a typicall tactical computer battle with a clear advantage after the opening.
IM Oleg Krivonosov: We may remember that GMs like to analyse with HIARCS as it sees human decisions. And JUNIOR is a good variation counter.
Theory was really that fatal and weak point where JUNIOR had lost in that game.
And of course even it's counting skills were difficult to save the much more worse position.
And ZCHESS looked really nice and haven't made any blunders.
GM B.Annakov: Last tryings of JUNIOR to save the game were interesting , but too late.
Marijus Kulvietis: Game n 12: CENTURY4-GAMBIT-TIGER. 0-1
IM Vaidas Sakalauskas: That battle simply had to be very ambitious as both programs are the children of the same REBEL company. And both are created as a totally different fighters.
CENTURY improves itself trying to find some interesting and maybe not always the best features-but those possible to imitate the human's game. While GAMBIT as we all know is created as a tactical killer.
GM A.Kharlov: And a tactical killer had won: Anyway this is a programs tournament.
Senbernar may be a fine dog, but pitbull or rotveiler have more chances to win the dog's fight.:
IM Vaidas Sakalauskas: The most important moment 1/ that the main mistakes
Made by CENTURY were positional mistakes. Thus we cant say the result is such is because
GAMBIT-TIGER was too strong in tactics.
2/The endgame is typicall weakness of computer chess and the field where humans have advantage in it's understanding. And endgame was played better by GAMBIT-Tiger as well.
Perhaps two important moments about this duel.
IM Oleg Krivonosov: I'm sure it was all up to the endgame. The draw was possible and would not be surprised by it. But Gambit-Tiger had played the endgame really well and had realized slight advantage very well.
This game reminds me a corrgames when it is so difficult to realize even a pawns advantage in a clearly better endgame. Now I see that GAMBIT-TIGER can be good advicer in such situations:.
GM B.Annakov: What seems strange to me-I would say that white's game was more active ,
Aggressive as if GAMBIT-TIGER was white. And Black had lost in it's field-more human understanding while trying to fight sharp chess. Maybe programmers were to concentrated about fight vs typicall tactical program as they knew GAMBIT -TIGER. In short-an impression would be ok if white were
GAMBIT and CENTURY -black: I would say programs have changed their typicall styles in that game.
Marijus Kulvietis: Really-interesting fact-CENTURY playing as GAMBIT-and GAMBIT winning due to those features that are created for CENTURY!
And the final game n13: SHREDDER5-CHESSMASTER8000. 1-0
IM Oleg Krivonosov : the game where the result is not surprising at all.
SHREDDER5 is a clearly leader in this tournament. Perhaps everybody expected SHREDDER will win this duel without any bigger difficulties.
GM A.Kharlov: And the win was doubtful. After CHESSMASTER managed to trade queens the position was rather drawish and further CHESSMASTER looked not bad.
Just sure SHREDDER5 is a very powerful program and it had overplayed CHESSMASTER at the very slip endgame.
IM Oleg Krivonosov: It was not a fast win on opponent's blunders. Just winning step by step.
First improving it's position. Placing pawn on a6 and making whites position very difficult.
Computers don't like when their kings and other pieces have little space.
IM Vaidas Sakalauskas: We can say-the stronger program had won.
SHREDDER5 is really very strong!
But CHESSMASTER wasn't too weak and the game was interesting to study-all moves were rather clear
And it's good to show for people . Just we must agree that SHREDDER is seriously pretending to win the ODYSSEY tournament 2001!
Marijus Kulvietis: We will see If SHREDDER5 will manage to win this strong event.
Thanks for a discussion.
1. SHREDDER5, (1) - (6) CHESSMASTER8000 1.0.4., 1:0
2. REBEL-CENTURY4, (10) - (2) GAMBIT-TIGER14.6, 0:1
3. JUNIOR7, (9) - (12) PHARAON 2.5, 0:1
4. DEEP FRITZ, (8) - (20) NIMZO8, 0:1
5. MCHESS8, (23) - (24) YACE BERLIN, 0:1
6. PATZER311B, (17) - (5) SHREDDER4 CHESSBITS, 0:1
7. GANDALF432H, (11) - (19) CRAFTY 18.12, 1:0
8. VIRTUAL-CHESS2, (7) - (3) REBEL-TIGER14.6, ½:½
9. ZARKOV4.5T, (13) - (4) HIARCS7.01, 1:0
10. GROMIT 3.8.1, (26) - (14) CHESS SYSTEM TAL2.03, ½:½
11. WCHESS2000, (15) - (18) COMET B36, 0:1
12. LITTLE-GOLIATH2000V3, (16) - (22) SOCRATES X, 1:0
13. GENIUS6.5 CZUB-STYLE, (21) - (25) EUGEN7.92, 1:0
Download the games of round-9 in PGN format.
RANKING after ROUND-9
Place Name Sco MBch Buch Ws
1 SHREDDER5, (1) 7.5 34 43½ 6
2 GAMBIT-TIGER14.6, (2) 6.5 32 43 5
3-4 NIMZO8, (20) 6.0 34½ 46 5
PHARAON 2.5, (12) 6.0 33 40½ 4
5-6 REBEL-CENTURY4, (10) 5.5 36 47½ 4
YACE BERLIN, (24) 5.5 32½ 41½ 4
7-10 JUNIOR7, (9) 5.0 37 46 3
CHESSMASTER8000 1.0.4., (6) 5.0 32 41 4
GANDALF432H, (11) 5.0 29 36½ 4
SHREDDER4 CHESSBITS, (5) 5.0 28½ 37 2
11-15 DEEP FRITZ, (8) 4.5 37½ 49 2
MCHESS8, (23) 4.5 34½ 45 3
ZARKOV4.5T, (13) 4.5 31½ 39 3
GROMIT 3.8.1, (26) 4.5 31 41 2
REBEL-TIGER14.6, (3) 4.5 29½ 35½ 2
16-22 PATZER311B, (17) 4.0 37½ 49 2
CRAFTY 18.12, (19) 4.0 35½ 43 3
HIARCS7.01, (4) 4.0 29 35½ 2
CHESS SYSTEM TAL2.03, (14) 4.0 28½ 39 1
GENIUS6.5 CZUB-STYLE, (21) 4.0 28½ 36½ 2
LITTLE-GOLIATH2000V3, (16) 4.0 28 35 2
VIRTUAL-CHESS2, (7) 4.0 26 33½ 3
23 WCHESS2000, (15) 3.5 31½ 41 1
24 COMET B36, (18) 3.0 28½ 37½ 2
25-26 EUGEN7.92, (25) 1.5 28½ 36 0
SOCRATES X, (22) 1.5 28½ 35 1
Download all games of Odyssey 2001 in PGN format.