Fritz Fails to Find Best Winning Try
I've been reporting on the progress of the Steve Ham vs. Computer chess engine matches that have been carried at my chess web site. The Senior Master chess player has encountered greater competition than expected in these four correspondence chess games. Nimzo won its match with a win and a draw. Fritz has a draw with one game remaining.
The Fritz-Ham draw was especially interesting since the game followed the Kasparov vs. Rest of World game for many moves. In the other game Ham livened up a drawish-looking ending by sac'ing the exchange. He may lose the game as a result, but he was determined to play as he would against a human player, as he did in the other three games. He was also determined to make the games as entertaining and exciting as possible, for the sake of the many spectators. This experiment, matching a very strong cc player vs. two of the better computer chess engines played under standard correspondence chess conditions, has been the most popular feature I've had at my web site.We have seen both the strength and the weakness of the chess engines in these games. Undoubtedly, Steve Ham could have secured much better results if he had focused on the weaknesses instead of playing exactly as he would against a human. One of the weaknesses of the computer was shown in the draw by Fritz. Several readers submitted lines leading to an endgame win by Fritz. Ham and others later found drawing resources in these lines. I'm not sure the final work is in, but Fritz certainly never found any of these intriguing winning tries. Instead, it simply continued checking till a draw by 3-fold repetition occurred.
Something I found quite interesting was the effect of installing the Nalimov 5-piece endgame tables, which contain solutions to all endings where five or fewer pieces remain on the board (this counts Kings). I did this after the game simplified to six pieces, King, Rook & Pawn vs. King, Bishop & Pawn. This meant that as the chess engine calculated variations leading to the exchange of a piece, the position simplified to a solved endgame, so the software would search the tables for the solution. I don't know how the tables are organized or how the chess engine uses these tables, but I do know that my disk drive was buzzing away like crazy after I installed these Nalimov Tablebases. It sounded like a bee hive under my desk, and I wondered at the advisability of leaving my computer cranking away when I went to bed (I keep my computer pretty busy with my other work, so I let the computer chess engine take over when I go to bed, calculating all night and all the next day while I am at work). I couldn't help but think that I was risking damage to my disk drive with such a heavy load of disk activity!
So Fritz handles the brute force calculations with great precision, but the long-range planning demanded by a tricky ending appears beyond its capabilities. Most interesting!
Another observation I made, which may not be totally valid, is that the computer generally zeroed in on it's move within a relatively short time, say in less than an hour. Letting it continue to calculate up to 24 hours seemed to provide little improvement. It might calculate another ply (1/2 move) in the extra time, but this didn't seem to really matter much. This makes me think that we'll see a limited improvement in computer engine performance over the next few years. Even improving computer speed 50 times will only allow for the engine to look forward another ply or two. Of course, it is possible that the programmers will find a more intelligent approach to computer lines.
Here's the final position and the complete score for the Fritz-Ham game. Instead of penetrating with the White King to a8 and forcing the Black King to a less defensive position Fritz just continued to check with the Rook and occasionally attack the Bishop. Ham's defensive idea of "setting up a fortress" to prevent progress by his opponent succeeded in obtaining the draw.
Fritz6 - Ham,S [B52]
The remaining game versus Fritz follows, with the diagram showing the position just before Ham's exchange sacrifice. This appears to be Ham's only remaining try to inject some life into the position, and I give him credit for going into a line that may allow Fritz to calculate a win. Would you try this line? I recently gave up the exchange against the lowest rating player in an ICCF section, only to have the position eventually simplify to a probable loss (I'm still trying to draw). Pretty tricky stuff, these exchange sacrifices.
Ham,S - Fritz6 [E86]
66.Rxg4 fxg4 67.f5 Rh6 68.Kg3 h2 69.Bh1 Re8 70.Kxg4 Rg8+ 71.
After its last move Fritz suddenly changed its evaluation of the position from dead equal to almost a full point edge (one point equals one pawn). Steve Ham's comment on the current position follows:
"The text move seems forced, preventing White from playing Kg5 to support the advance of the connected passed pawns. Fritz 6a calculated for 18 hours and 20 minutes to a depth of 21/21 ply, at an average speed of 424 kilonodes/second. The chess engine "thought" that my pawn capture motif unbalanced the position too much since it's formerly even/draw assessment just changed to 0.94 pawns in its favor. This is a major devaluation, signifying a large Black advantage. But my belief is that my calculations are correct and, instead, the chess program is once again showing error in its ability to evaluate positions. Let's see who is correct here, the human or the machine."
For the latest match situation check my web site. The address is given at the top of this column.
Chess on Television
It seems to me that chess has become more visible than ever recently. As I sat here typing away (and flipping TV channels) I stopped momentarily on "Who Wants to Be a Millionaire" just in time for the $8,000 question, "Which chess piece only moves diagonally?" Well, not very challenging for us, maybe, but it was still rather neat seeing chess mentioned. I'm happy to say that the fellow got that one right (he stumbled on the $64,000 question).
Another ad begins with a man and woman playing chess. As the camera pans back it is revealed that they are on a moving trail, and the optical effect of the camera zooming back is quite impressive. The board was even oriented properly. I saw another movie with a chess game and, once again, the board was oriented properly with a reasonable looking grouping of pieces. I swear they get the orientation wrong more often than right, but recently that has been reversed in my observations. Perhaps people are starting to learn more about chess.
One of my personal favorite ads featuring chess shows football fans facing withdrawal from the end of football season. In one of several episodes the fans are cheering a play. The camera pans back to reveal two men at a chess board. "Send in a replacement", one of the men shouts. When I was at MIT it was said that the tiddly winks team had cheerleaders, but I've never seen a chess match with cheering spectators.
Of course, the big surprise was seeing a Pepsi commercial during the Super Bowl football game featuring chess. Now these Superbowl ads are commercials that cost a lot of money and are meant for massive viewing audiences. At the local university (Michigan State University) a group of professors gather each year to analyze the Superbowl commercials. I heard one say that the commercials reflect a great more effort than the game. Like many others, I generally listen to the commercials during the superbowl instead of pressing the mute button. Imagine my surprise when I suddenly saw GM Kasparov sitting at a chess board!
He was facing a computer and was just finishing off the game with a win. After that he encountered various mechanical problems, such as the elevator stopping between floors. The machines finally got their revenge for a "comrade's" chess loss when Kasparov was knocking into the open elevator shaft. This is probably the most expensive, most widely viewed commercial featuring chess ever seen. Even with the antics of the FIDE people, chess continues to be held in great esteem by the public, obviously considered to be a game that requires great intelligence (and maybe a certain amount of weirdness).
ICCF-U.S. Secretary Max Zavanelli
For many years Professor Max Zavanelli has sacrificed his personal correspondence chess goals to work on cc opportunities for his fellow countrymen. His organization of the USA office of the ICCF and work for the benefit of all USA cc competitors (and for all cc players in general) has been appreciated and has been highly successful. Thanks for everything, Max!
However, it appears that recently he has been able to find the time to return to active competition, and his return has been spectacularly successful! He has achieved the norm for his International Master title with a score of 10/11, and there are still three more games yet to be finished. Amazing! This performance has come in the ICCF Reg Gillman Memorial "amici sumus" group. Max should receive his title and medal at the ICCF Congress in Rimini, Italy later this year.
Tõnu Õim Voted Top CC Player in History
Tim Harding, editor and publisher of the excellent Chess Mail magazine, has recently published the results of his reader poll for the five top correspondence players in the history of the game. Dr. Hans Berliner, the fifth world cc champion and first from the USA, took second place in the polling. He was my first real cc hero, with his remarkable and convincing win. His book on this world championship event was my bible for years. It was my great privilege to meet this giant in American correspondence chess at the ICCF Congress in Florida last year, and he graciously autographed my old and well-thumbed edition of his book. He is retired from his university work and has already published one new book, The System. From what I heard he will be probably be producing another book in the future, and I'll certainly order one as soon as it becomes available.
The top vote-getter was current world champion Tõnu Õim of Estonia, the only man to have won the world championship twice. I regret not getting to meet this outstanding cc competitor, but he was unable to attend the Daytona Beach ICCF Congress. Following is the complete list of the top five.
By the way, I strongly recommend Harding's excellent cc magazine, published in Ireland. If you haven't done so before, subscribe now. You can obtain a one-year 8-issue subscription to Chess Mail by sending $46 to the following address:
My Changing Approach to Competition
For many years I played according to a simple philosophy. To play strictly and completely by the rules was a non-issue. I followed the letter of the rules and expected my opponent to play the same way. Therefore, if the rules called for a 5-day penalty when my opponent neglected to write the time used on his card, I simply added the five days. If he failed to give the time received and sent, I asked him to do so on future cards. If I overstepped the time limit I simply indicated to my opponent that I had overstepped and applied the appropriate penalty to myself. This is what the rules indicated, so I did these things without much thought.
I thought of cc competition as a sport, such as basketball or football. There referees watch carefully and enforce the rules precisely. The cheering crowd may be disappointed when a decision goes against their team, but everyone expects the rules to be followed. If the ball is one inch short of a first down, so be it. If a player accidentally runs into an opponent, the foul is called. Nobody complains (well, not much) when some small infraction of the rules leads to the loss of the game. This isn't universally true in correspondence chess competition, though.
I ran into an opponent who refused to list the dates. He objected strenuously when I applied the prescribed penalties. Things got very unpleasant and, in the end, I dropped the penalty claims and agreed to continue without his cards being dated. I even stopped dating my own cards. I later regretted doing this, but it did return the game to some normality and our correspondence was interesting and friendly.
When I attended the ICCF Congress in Daytona Beach, Florida last year I found the prevailing opinion to be that rules were only there to use when absolutely necessary. The spirit of "amici sumus" (we are friends) was considered the most important philosophy. And this was a congress that was squarely focused on updating the ICCF rules of play. It was stated that the rules should not get in the way of playing chess. Even the President of ICCF stated that he didn't use the rule stating that the official "send" date is postmark date on the card. If the postmark was a day later, he ignored it and used the date claimed by his opponent. The rules were only really in place to allow for dealing with difficult opponents.
Now this was a new concept for me! Instead of enforcing the rules as a matter of course, instead of considering the rigid following of rules being a non-issue, it wasn't an issue at all! The rules are simply available for use when needed to deal with a trouble-maker.
I've been trying to follow this new-found approach to cc competition. As a first step I stopped using my opponent's postmark as the "sent" date in my time calculations. If I disagree with an opponent by a day or two in calculations, I just accept my opponent's calculation. If my opponent leaves off dates or time used, I just ignore it. I even allowed an opponent to overstep the time limit by a day without comment. This was a strong opponent who wouldn't overstep a second time, so I was confident it would make no difference in the final result of the game. I have determined to record the dates and time used on my cards, though, whether or not my opponent does the same. I won't insist on my opponent following these rules precisely, but for me it's still important to do so.
The rules lawyer is generally held in low regard, it seems. Is that what I've been all these years? Was I just being a trouble-maker? That certainly wasn't my intent. I just expected my opponents to hold the rules of play in the same high regard as I. The rules of competition seemed as important to me as the rules on how to move the pieces and affect checkmate. Apparently I've been wrong for all these years. I'm doing my best to adjust to this new attitude, but it is a struggle. In the end I'll undoubtedly enjoy a lot of chess and have few problems.
Now that I'm also directing an ICCF tournament I can appreciate another view of the rules. There are indeed players who insist on the enforcement of the rules of play and who are sending me official complaints concerning their difficult opponents. This is the one thing I don't enjoy about directing. The work of recording results, collecting game scores, publishing standings and confirming results for title norms are a pure joy. Dealing with player disputes is quite the opposite. Players pity the poor tournament director! Try to avoid disputes work out problems among yourselves. Only involve the TD when you simply can't work out your problems. And don't be a rules lawyer.
copyright © 2000, 2001 by J. Franklin Campbell
|Home||Column Menu||Previous Column||Next Column|