Fascinating article covering the question of whether Magnus Carlsen back in the World Championship cycle?
Most interesting is the rule on draws. Not a surprise there is a rule on draws, but rather than the historical “Must make X moves before an offer,” now FIDE has leaned towards other methods usually favored by privately sponsored tournaments.
At the world championship level, allowing theoretically known draws to be claimed may make sense. Would that include table-base checking and their sometimes tortuous drawing methods?
At lower levels I’ve seen 1800+ players mis-handle and lose a drawn K vs K+P ending, even in non-time-pressure situations (including black K on e7, white P on e6, white K on e5 - black played Kf8 rather than Ke8).
When players want to draw, they will, regardless of the rules. I guess we will see the “spectacle” of players shuffling pieces back and forth and creating more draws through a repetition of position.
These attempts to avoid draws have all been done before. In the late 1800’s and early 1900’s, a number of tournaments had odd rules that cancelled draws, required replays of games, or gave 1/4 points for draws. In one tournament, a draw by Black was scored as a win for him. All of these rules by reformers were derided by the players and the public after some shocking results that deprived favorites of prizes. My favorite weird tournament result was the one notched by US Champion Frank Marshall. He entered a round robin tournament in England that awarded one point for a win against a lower ranked player, but more points against a higher ranked foe. Marshall played 3 games a day at normal time controls and won all 44 games. He came in third place!
The best way to cut the number of GM draws is to provide a financial incentive to play fighting chess. That is how i would attack the problem, make it worth their while to risk. Maybe a prize for “aggressive play” even for a player who didnt score high, but tried hard to win.
That would work if it’s unrated or if the top 10 is only playing the top 10. Conditions and appearance fees based on ratings will outstrip tournament-based financial incentives for a long time.
I was referring to top-level tournaments, like “top 10”. I don’t see chess having a draw problem in lower-end tourneys…
If financial incentives produce fighting chess, more fans will keep and gain interest, and everyone benefits.
Tablebase positions have been misplayed as recently as game 2 of the Kramnik-Topalov match. 6k1/1p2r3/3K4/6N1/3P4/8/8/8 b - - 0 53
Geller-Fischer, 1970 interzonal, Fischer-Taimanov match game 2, & this position from the Korchnoi-Kasparov match 8/8/4K3/6r1/3P1k2/3R2p1/8/8 w - - 0 1 were all tablebase draws that were lost.
Let the humans play.
I was watching Ivanchuk-Topalov a few days ago on ICC. According to the computer, Ivanchuk had an advantage. Then Topalov turned it around and had a similar advantage, about +.5 according to Houdini. Suddenly it comes across: DRAW. Give me a break. My point isn’t to blindly follow computers, but after this, I can’t listen to the people that say there is “no draw problem”. THIS is a problem. The game was dynamic, obviously with the advantage gone back and forth was interesting, and had an unclear ending. The players should have no right to just walk away from this game. There should be a penalty for such action. Heck, imagine if I walked out of work after 6 hours instead of 8!
Speaking of World Championships, I witnessed a really bad example of this. I was at game 6, Kasparov/Anand, at the World Trade Center, 1995. Kasparov was up the exchange for a pawn, and most spectators and experts believed he had winning chances in the endgame. In any case, the position still gets discussed occasionally. Fritz 13 believes Kasparov is up over 1.00. Suddenly DRAW. A few spectators, including myself, actually booed. I love endgames, and they should have been forced to give me my money back. It’s a travesty to believe that just because these players are great players, that they have the right to subvert the game. I’m sorry, I just don’t buy it. I did not go to another game, for fear of the same frustrating thing happening.
To deny that this is a problem just in order to “respect GM’s” is just silly. I hope for chess’ sake it gets addressed seriously. It’s cheating, except it’s the fans getting cheated.
Kasparov-Anand 6 is an excellent example: didn’t two US GMs play out the ending for the spectators’ benefit? Also see Anand-Kramnik WC game 2 and a famous incident from the Lyon leg of the 1990 Kasparov-Karpov match.
Even at the WC level, Sofia rules make sense. (What is interesting to us may be trivially drawn to them, but that’s not much of a downside.) The players in a match are more worried about winning than about entertaining the spectators, so change the rules to make “entertainment” an integral part of the match.
Perhaps the above statement alone provides sufficient reason to question the remainder of the argument. The key phrase is “Then Topalov turned it around…”
Immediately, we have an issue. For you see, one cannot “wrest” the advantage. The advantage is bestowed. It comes either because White has the first move, or because the opponent makes an error.
Therefore, if the computer when from saying Ivanchuk had the advantage to Topalov having it, one of two things is true:
Computer Analysis Incorrect:The computer may not have correctly evaluated the original position, and until it reached a position further down the road, the horizon effect of the assumed better position wasn’t wiped out. Certainly, I’ve seen this happen many times with some programs.
Computer Analysis Correct:In this case its clear that Ivanchuck must have made a mistake.
No where from the above can we conclude that the position is dynamic, or that a draw isn’t the correct result. In fact, I can show, from my first tournament in years this past weekend, that the first round game was similar. The position was basically equal, perhaps slight edges for one side or the other. Then my opponent erred and I won. But the position was never dynamic. In fact, the issue was largely that if was the kind of position that would be “tainted” if one tried too hard to win. Computers are notorious for showing that there is an edge even if that edge is not winnable. (It’s the chessplayer’s version of whether a tree falling in a forest makes a sound if no one is around to hear it. Is an edge truly an edge if its not winnable?)
I suppose it’s also good to remember that += essentially means “better side of a draw”. A .5 advantage doesn’t indicate that the correct result isn’t a draw. It indicates that one side has an easier time of it that the other. A .8 advantage doesn’t indicate that White “is won”. It indicates that White has a sizable advantage.
Let’s assume you are self-employed, or in a job that has essentially self-employment “rules.” (I work such a job.) You will be paid in respect to the value you provide. If I am self-employed and working on a project, I have days where I may work 20 hours on that project, and days where I may work 2. But I am paid by the value of the PROJECT, not the value I bring each day. Perhaps one day I will have an idea that will accomplish 20 hours of work value in 2 hours of work. Should I be paid less because I work 5 hours on that day?
The players are paid per project (tournament) not per game. It is well known that compensation (amount and structure) changes behavior. If one believes that there is a draw problem (I personally do not believe this) then the correct approach is to change the compensation structure, so that some compensation is delivered PER GAME rather than having it all delivered PER TOURNAMENT. I suspect any compensation consultant would see this quickly. Other sorts of rules are artificial, and easily allow for artificial results.
Why do you think the game was subverted? (One can put endings on different programs and obtain quite different results.) Was the analysis correct? Were both players feeling well that day? It may have been a very good pragmatic decision in respect to the overall project.
(BTW - Houdini [which I picked up this week] after a few minutes evaluated the position as += .30, while Fritz in the same amount of time had it at ± 1.29; so much for clarity of argument.)
I don’t deny it’s a problem just to respect GM’s. I deny its a problem because in tournaments and matches the goal is not (generally) to win a game. The goal is to win the tournament or match. Sometimes those goals coincide. But not always.
If we pay players for the goal of winning a tournament or match (the project) and then expect them to ACT as though the goal is to win each game then we sorely misunderstand human behavior.
It’s not often that I agree with Kevin as strongly as I do here. He nailed the whole problem.
If organizers want to reduce draws, they should reward winning games rather than winning tournaments.
Bill Smythe
Then a fairly strong player (top 16th but not top 32nd) can sacrifice the first game of a five round tournament and win the next four, rather than winning three and drawing two. That would be a variation on the Swiss Gambit. Alternatively, the player could win until reaching an unwinnable game and then pitch that game rather than fight for a draw so that the player would have easier pairings after that.
Agreed. This is why my original post noted that some of the compensation needed to be weighted to the tournament, and some to games.
An implementation of this concept is the Plus-Score event. (Bill Smythe, did you develop this or did you just refine it?)
In a Plus-Score, prizes are awarded for Score - this is essentially a combination of awarding for winning games and for winning the tournament.
I can’t claim to be the first to develop the plus-score concept, but I hadn’t seen it before when I came up with it. It’s an obvious enough idea that many people could easily have invented it independently.
My contribution, if any, was the exponential (or, at least, rapidly increasing) dollar amount as the score increases. For example:
4-0 $100
3.5 $50
3-1 $25
2.5 $10
Each difference is greater than the difference just below it. For example, the difference between $100 and $50 is greater than the difference between $50 and $25. (That includes the last difference – $25 minus $10 is greater than $10 minus $0.)
The decreasing differences tend to discourage draws. If two players are paired against each other at 3-0 going into the last (4th) round, they will win a total of $125 for a decisive game, $100 for a draw.
Bill Smythe
Thanks for your detailed response. But your arguments don’t make much sense at least not to me.
Not that it’s relevant to the discussion, but in my games, over-the-board, and with a clock running, players indeed do “wrest” the advantage from each other. If you choose to refer to a game in a theoretical sense, where a computer can determine an error after 20+ ply, then fine. To me, it’s semantics as to whether one player “made an error”, or the other player “wrested advantage”. By your theoretics, I guess Tal wasn’t much of a player, since computers can refute most of his combos.
"No where from the above can we conclude that the position is dynamic, or that a draw isn’t the correct result. " I agree. A computer can’t, or shouldn’t. But you seem to be implying that you bestow Ivanchuk-Topalov or Kasparov-Anand the godly right to determine that a draw is “the correct result”. What the heck does “correct result” mean anyway? Again, i have a feeling you’re referring to some 60-ply theoretical finding by some supercomputer. Who cares? The games I refer to should have been played out. Topa had a passed h-pawn in fact, and the fact that Kaspy-Anand is STILL talked about casts aspersions as to any talk of “correct result”. Correct result means nothing in those positions. As someone pointed out here, EVEN tablebases have been incorrect at times, and who even cares about that. Chess is a fighting game OTB, not a theoretic-adjudicate-by-computer-or-GOD technique.
You ask me why I think Kaspy-Anand was subverted. I wish you were at that game. Even the experts up front at the demo boards were absolutely speechless. There’s no better example of ripping off fans, or “subversion” of the game, as you call it, than what I saw that day. To this day, I would be reluctant to pay money for another championship match. In 1990 I attended 2 fabulous games K-K, game 3 and game 6 that were played to good conclusions (both draws). 1995 was a monstrous rip-off.
I agree with you that financial incentives need to be in place. But I also agree with the Sofia rules.
You continually give examples of computers (Houdini, Fritz) differing in evaluation of positions. Which further proves my point even more that these games need to be played out, that draw agreements should be prohibited.
Keep the draw; fix the flaw.
good one!
Another way of looking at it is if one person is going into the last round 3-0, the worst they can do is 3-1 winning $25. Therefore, with their 4th game, they win $25 for a draw, and $75 for a win.
Sorry, its more than semantics, its fundamental to understanding chess theory. The advantage is never taken. It is bestowed. Understanding practical play is different than understanding theory. The two go hand-in-hand, but fundamentally correct theory must trump.
You were the one who introduced computer valuation to support your argument. My argument was nothing more than to show that a computer valuation did not support your argument.
By agreeing, your argument is undermined.
Not at all. There is no requirement in chess that games end “correctly”. I once saw an expert defend a difficult position for hours (forgive me Dave, you know who you are) only to resign in a position where he had finally equalized. As an expert on the way to Master, I once played a sacrifice that led to a mating attack. For some reason I had an attack of chess blindness and missed a mate in one, eventually losing. (This was so bad, that FM Al Chow walked up to the board, saw the mate and left. When he later returned and we were still sitting there, he assumed we were analyzing and reached out and started moving the pieces, only to have us say “What are you doing? We’re still playing!”)
Would you argue that either of these were the “correct” result? In terms of the preceding play, of course not. In terms of the total play, of course they are.
Players don’t need to have a right to the “correct” result. They have a right to “THEIR” result. And if you want different results, then reward them differently. This is common practice for behavioral change. Why would we assume differently here?
You argued against a draw in part due to computer evaluation. My point was simply that there was nothing in the computer analysis to indicate that the theoretical result of draw (the theoretically correct result) was unreasonable.
Other than the reason that you insist it is so, why? The players weighed their risk/reward in the tournament standing, and didn’t.
If correct result means nothing, then why should they have to play it out?
Sorry, the players delivered what they were paid for. That’s precisely my point. Getting angry at the players won’t fix that. PAY THEM DIFFERENTLY. That will fix it.
I don’t agree with Sofia rules at all, and will never support any attempt to legislate the outcome, other than as a “theoretical” tournament. Why don’t we just legislate what moves players can make?
Chess is a game between two players. Each player makes their own decisions. If you want to change their decisions, change the risk/reward strategy.
I gave the example because YOU were the one arguing based on computer analysis. I was simply showing that such analysis is faulty. It therefore does not prove the point that the position needs to be played out. There’s no logical connection there whatsoever.