Fightless Birds

I agree Mig. This nonsense that a game would be drawn anyway because its equal by move 18 is ridiculous. If elite players agreed to a draw everytime the position was equal why not just draw on the starting position?

We've had these debates earlier and almost everyone seems to be unequivocally in favor of banning the draw offer. Indeed, some tournaments have done so with success. I wonder why that inestimably wise institution, FIDE, and its inestimably sage president cannot incorporate it into the rules.

Mig Wrote:

"Just ban the draw offer and play chess as it was meant to be played."

and

"The draw was never intended to be a strategic element or a sporting device; it's an occasionally dull necessity."

I agree with much of you say. However, to suggest that the sporting aspects of draw offers (as they relate to tournament tactics) don't gel with how chess was 'meant' to be played opens up a big kettle of fish. Many would argue that intensive opening preperation doesnt fit with chess idealism, for instance.

I, for one, find tournament tactics specific to chess fascinating, as well as opening preparation, and would lament the loss of either.

John Nunn has "little sympathy" for people who feel that -- the high draw rate in grandmaster chess is so high that chess would be better if the rate were lower.
Nunn only objects to "the frequency of short draws".
( http://www.chessbase.com/newsdetail.asp?newsid=2729 )

Mig wrote "I'm not against draws, just short, non-game draws".
Mig and John appear to agree. In contrast, I believe chess would be better with a substantially lower draw rate.

The grandmaster draw rate is approximately 63% (among recent champship tournaments and matches).
Taken a literal face value, Mig & John are implying a 98% draw rate would be no problem, as long as they were all long hard-fought draws. Since that would be an absurd claim, let me explain how I interpret their remarks.

I believe Mig and John are implying that the 63% draw rate is fine, even though a 98% draw rate would be too high.
The only counter-point I wish to make is that their's is an *arbitrary* judgment. There is nothing wrong with arbitrary, as long as there is no pretense that the criteria are anything more substantial than that.

Mig and John have in mind some unarticulated maximum draw rate that is okay with them, and apparently that unarticulated rate is above 63%. Fine for them, but nobody's arbitrary judgment can speak for or dismiss the opinion of the masses. I doubt I am persuading the guys who have "little sympathy" for "a bunch of patzers [who] want to see more blunders or ... don't understand endgames".
BTWay, what is that unarticulated max percentage?

Yes the outcome of a chess gamed played perfectly by both colors is a draw; so there is nothing wrong with an individual drawn game.
But it is cheap sophistry to say ergo there is nothing wrong with the fact that the majority of grandmaster games are undecisive. There is no logic bridge for that jump from the singular to the group.
Besides, most sports give both competitors an even chance, something today's chess fails to do. Yet instead of thereby being even more draw/tie prone than chess, most sports have a draw/tie rate in the single digits.
So please stop using as an excuse for a high draw rate the fact that chess is a theoretical draw.

If hard-fought draws are loved by the chess public, then why did John Nunn include zero draws among the 30 games in his book "Understanding Chess Move by Move"? Something about draws ain't popular.

I visit Mig's blog because I love chess. But today's implementation of Abstract Chess does have a deep flaw in its draw rate problem, in elite grandmaster play (and in its most famous and important events).
I hasten to agree that any change capable of substantially reducing the draw rate would change the game and it would no longer exactly be chess anymore. In other words, the draw rate problem in chess will never be fixed, it is unfixable. Trying to fix it is folly.

Sofia 2005 had a 60% draw rate, so these attempts only affect the frequency of short draws, not of overall draws.
Nonetheless, Tim Kett had an interesting idea: when requiring a minimum number of move-pairs before allowing draw offers, also require a minimum of 3-4 hours of clock time.

For academic progress, I would like to see ChessBase.com run an article in 2008 describing a few well engineered implementations of Abstract Chess that both
(A) avoid the draw problem (no mere wishful thinking allowed), and
(B) are as much like today's chess as is plausible.
It is too early to know what we might learn from such an academic exercise.

Not all variations of Abstract Chess are funky, even though the funky ones dominate ChessVariants.com.

Thank you.

"Taken a literal face value, Mig & John are implying a 98% draw rate would be no problem, as long as they were all long hard-fought draws. Since that would be an absurd claim,"

Why?

Mig has put it right: The extended draw offer is interesting in itself, but a little beside the point. It does eliminate the fear-of-declining-draw-offers factor, which I feel would be very beneficial to my own play. But it does not address the Kramnikian casual fightless draw. Plus, the need to count moves would make things complicated.

The banning of the draw offer is more to the point. Sofia has shown it is a feasible solution. FIDE should give it a try, and encourage its usage.

What's the point in banning the draw offer? If players want to fight or win the will not use draw offers. They are professional players and they know what they are doing.

If you want to win somebody, eventually you just have to fight. It is a simple as that. I really don't see any point in denying so called GM draws. You really can't tell a marathon runner not to lessen pace if he is exchausted. Draw is already a half loss to both players, in an even field that is also expected result, but also most likely not enough to win a tournament.

The easiest way to increase the number of wins is to invite more players with different rating levels. Even though there would be more draws there would also be more wins and pre-arranging would be more difficult to do. Young players with a lot of passion will definitely create sharp games.

But if you instist in watching the best players in the world... I'm not sure if it is a good idea to force them play dull an not interesting positions to the end. I'm really not sure if that is a good idea to the spectators either.

There is a simple, fair solution that doesn't mess with chess tradition, rules, ratings or anything else:

If the game is drawn (for whatever the reason, including pre-arranged draws), successively shorter, "overtime" games (reversing colors after every game) should be played until a decisive result is reached. Only the initial game would count for rating purposes. Only the last result would count for tournament scoring purposes.

Very simple, doesn't compromise the quality of the INITIAL game (the one that counts for historical purposes) and brings far more excitement to fans who know they wil get their "blood" no matter what...

Simply encourage the players to continue the game, as an
exhibition, from the agreed drawn position, with the clocks set
back to zero, at a rapids time rate. This doesn't put too much of
a burden on them and should satisfy the onlookers' interest and
bloodlust.

I like draws in some situations. What do you do at move 63 of an uninteresting game and you have a headache, and two more tough games scheduled for tomorrow? Hope that your opponent falls asleep and drops a piece? Is that chess? After sitting at the board for several hours with someone who is clearly seeing as much as I am, a draw is a good way to save energy for the next round.

"The easiest way to increase the number of wins is to invite more players with different rating levels. Even though there would be more draws there would also be more wins and pre-arranging would be more difficult to do. Young players with a lot of passion will definitely create sharp games."

Ironically, the ELO system has contributed to the drawish nature of Top Level Chess, since it provides an incentive to the organizers to put together an even of the highest category that they can muster, and it causes Top players to play cautiously out of a fear of losing rating points. FIDE should toss out the rating system.

Even if the explicit draw offer (and agreement) is banned, there is still the problem of the tacit draw offer, in the form of repetition of moves.

"In other words, the draw rate problem in chess will never be fixed, it is unfixable. Trying to fix it is folly.

Sofia 2005 had a 60% draw rate, so these attempts only affect the frequency of short draws, not of overall draws.

Nonetheless, Tim Kett had an interesting idea: when requiring a minimum number of move-pairs before allowing draw offers, also require a minimum of 3-4 hours of clock time."

Yes, this should be incorporated in efforts to deter draws.

Short draws are bad. Even long draws, without a sharp fight, are not much better. If you merely rely on move minimums, then you might have a situation whereby both players will at some point tacitly agree to a draw, and simply race through the moves, reaching the designated number of minimum moves (be it 40, 50, even 60 moves)--yet not having used their allocated time.

So, perhaps there could be a requirement that in a Classical game, BOTH players will have had to use up at least 3 hours EACH of Clock time (which entails playing past move 60) before draw offers can be made OR accepted.

Or course, this would only serve to limit draws by agreement.

This accomplishment would be mere "window dressing" since the players could tacitly collude to play draws without any real struggle. Professionals could concoct Triple Repetitions via Perpetual Check, or they could simply systematically trade off all of the pieces, leading to a draw by Insufficient Material (aiming for Bare Kings).

Of course, fans WOULD appreciate the display of ostensibly fighting chess, even if they understood that the game *might* have devolved into something closer to an exhibition by the players. In particular, spectators AT the tournaments would be less likely to begrudge the players their draws if it were manifest that the players were required to invest time, energy, and effort in the result.

"If the game is drawn (for whatever the reason, including pre-arranged draws), successively shorter, "overtime" games (reversing colors after every game) should be played until a decisive result is reached. Only the initial game would count for rating purposes. Only the last result would count for tournament scoring purposes.

Very simple, doesn't compromise the quality of the INITIAL game (the one that counts for historical purposes) and brings far more excitement to fans who know they wil get their "blood" no matter what..."

I think that this ought to be tried. Of course, tie-breaks at progressively faster time controls are already used in Knock-out events.

However, it might be more appealing to mandate tie-break games in Round Robin events.

In Knock-out events, each match is "Winner take All". The loser is eliminated, which the winner advances (or wins the event). A lot--perhaps too much--is at stake on a final Armegeddon Blitz game.

In contrast, in a Round Robin, all that would be at stake is a full point. Losers would not be eliminated. There would be an incentive to play aggressively in each game, since there is no intrinsic benefit to drawing the Classical (and Rapid) games, and thus having to play more chess, and be less rested for future rounds.

Both players would have to take into consideration complex tournament factors (for instance, one of the players might be relatively stronger in Blitz or Rapid chess, and hence might be happy to draw until he believes that his practical chances are superior). That strategy might work for one or two opponents, but it probably would not be sustainable for the length of a tournament. Agressive players who press and win in the Classical game will be rewarded with greater rest, and more energy for future rounds. Organizers can also add Prize money to be paid out for each Classical game victory, which would provide a further incentive to fight.

Ironically, in the FIDE Knock-Out events, there is a perverse incentive to draw the 2 slow games, since if the mini-match goes to tie-breaks, the loser gets an additiona sum of 10% of the winners share of the Prize for the next round. Players logically want to hedge their bets: The extra 10% is nice if you get eliminated, but a modest sum for the Match winner to concede, in exchange for a continued chance at the Jackpot that comes with the tournament title.

I believe that incorporating Tie-Breaks that will entail settling ties over the board actually would cause less competitive distortion when applied in a Round Robin format event, than it does in a Knock-out format.

The best tournament players in such a system would be those with good physical shape (stamina), fighting spirit (Aggressiveness in playing for victory), and superior rapid/blitz chess skills.

3 hours of clock time? I don't think I've ever used 3 hours of clock time in a competitive game, not even my game against Roberson at Swansea that went on for 144 moves.

Perhaps not Jack, but would you really deny those who think the interesting part is seeing you sitting around doing nothing?

At organizer's discretion, a less than 35-move draw can lead to a 5 percent appearance fee/final prize penalty. The penalty fund goes into the prize for the eventual winner of the tournament. Rewards playing for the win, hits the GMs where it matters and does not change the rules or nature of the game.

I've asked before, but never heard any response to this question from anyone:

- Is the problem REALLY the short draws or could the problem RATHER be that we share prizes at tournaments?

The practice of sharing price money rather than using tiebreaks or buchholzes or whatever makes it sadly comfortable (almost irrestistible) for top players to go soft on each other in the last round (when short draws are the most annoying for spectators).

Sweden's finest open, Rilton Cup, just ended with *nine* players at first place. Last round saw draws at tables 1-5, 3 of them before move 20. http://www.rilton.se/id_16.asp

So:
How about banning shared price funds and appointing ONE winner of our tournaments before banning draw offers?

Taking away the incentive for draws (/quick draws) - if that can be done sans rule changes that please patzers but damage us chess-lovers by making chess look more like football - would be more logical than simply banning draw offers.

The problem with Jens Petersson's comment, though, is how would you award prizes if not split between equal scores?

He's probably talking about making Armageddon tiebreaks universal. But there's already too much gambling in chess, to my mind. I think it's the patzers, not the chess lovers, who'd enjoy watching a sport where all the money goes to the guy with the fastest hands, rather than the guy with the sharpest mind. (If you've ever played a park hustler for bucks, you'll know what I'm talking about. If you haven't, then I don't care what you think, and neither do the readers worth reaching or worth listening to.)

If not some form of fast chess as the tiebreaker, then what? System-based tiebreaks would obviously piss off the players even more, without even satisfying any of the fans.

The only other possibility, it seems, is some kind of change in the scoring system - either the mentally retarded, anti-chess sort (draws count as 1/3 point), or the more sophisticated sort proposed by someone on another thread here (draws count as 1/2, but for tie-break purposes only, you lose a fraction of a tie-break point for each cumulative move below 40, or 30, in each short draw you play during the tournament).

On that earlier thread, I praised that last idea, which I'd never heard before. I'm still not convinced that any change is needed at all. But if one is made, then obviously the best option would be one that:

- has a logical justification,
- comes across as fair by penalizing only behavior that all agree is undesirable (i.e., short draws, not all draws),
- is simple to understand, and
- isn't motivated by the futile and stupid wish to transform chess into a different game with greater appeal to football fans than chess players.

Gene M: "the outcome of a chess gamed played perfectly by both colors is a draw"

Says who? Chess has not been solved, so it is not possible to determine the result of a "perfect play" game.

I would be interested to know what the draw rate is between computers, as I assume that all computer draws are "dead drawn". Anyone got the stats?

Jon Jacobs assumes that I'd like to see armageddon tiebreaks as the solution to my question on how to divide prices between players on the same point.

This is not the ONLY solution, however. Kasparov and Topalov tied for first in his last tournament and shared price money, but the actual Cup was given to Kasparov because he had better quality points, buchholzes or more blacks or better success in his individual encounter with Topalov (I don't remember which).

Couldn't we live with winners getting both the Cups and the full price check in the future?

Ofcourse there will be situations where nothing else than and armageddon blitz or the drawing of lots could divide players. At such situation maybe we should accept divided victories, I'm not yet sure what I think myself. I do, however, prefer lottery over *armageddon* blitzes, so don't pin me down to something I haven't suggested!

Jens

Jens, the obvious point of my last comment was to get you to come out with a specific recommendation. You just failed to do that - again.

Your wish (to have a single top prize winner instead of money equally split) sounds appealing in principle. But like other, equally appealing concepts like "Human rights" or "Democratic elections," the devil is in the details.

I don't know what a "buchholz" is.

"Better quality points?" Don't understand that either; surely you aren't talking about the chess quality of the winners' games? (Who would judge that, when hundreds of thousands of dollars might be at stake, and when the competitors in some cases would inevitably be higher-rated than the judges?)

Both head-to-head results and "most wins" have been used to break ties on occasion. Both methods meet the "fair and logical" test I outlined above, but often they won't succeed in breaking the tie (i.e. in a double RR, the co-winners usually will have split against each other, and also often will have the same number of wins).

Anyone who doesn't know what "a buchholz" is can benefit from Wikipedia's wide coverage:

http://en.wikipedia.org/wiki/Buchholz_system

As for quality of points, this would sound normal to most Europeans. Someone with a higher tiebreak should in general have a higher quality of points, i.e gained against stronger opposition.

I've played in Swisses for 40 years and never heard of "Buchholz" tie-breaks. The ones I grew up with were called Harkness, Harkness-Median, Solkoff, and Sonnenborn-Berger. Must be a trans-Atlantic thing.

Anyway, maybe it's different "over there"....but, the feeling over here seems to be pretty overwhelming that system tiebreaks are unsatisfactory. For the U.S. Championship, for instance, the authorities (during the fat AF4C period) went to great lengths to avoid having the winner decided by any form of system tie-break.

Hard to believe anyone would be proposing a stepped-up use of system tie-breaks as a fan-friendly improvement, when all the momentum I've seen seems to point in exactly the opposite direction. (After all, if many people felt that system tie-breaks were a good way to break ties, Armageddon would never have happened, would it?)

Well, Jon, maybe it's my English and the lack of common terminology that confuses you from seeing my concrete proposal? Any way: to sum it up:

My thesis is this:
The practise of splitting of price money fuels players will to take last round easy draws and that these are specifically disliked by fans.

My specific proposal is:
Try very hard to avoid splitting price money.

Apparently something called "system tie-breaks" in the US, which I guess we in Sweden kall "kvalitetspoäng" and which GG gave a very good example of in a Wikipedia-link above could be used as a part of this solution.

Jon seems to argue (correct me if I'm wrong) that Buchholzes etc are not fair enough to judge a winner from. In some instances I'd be happy to agree - in some tournaments it would be better to have tiebreaks games played (for instance to sort out who's the US or Swedish champ). In other instances such tiebreaks are impractical and we will remain stuck with either "divide the money and love the draws" or "love the buchholz and learn to play to win!". You now know where my heart is in that choice.

I agree that banning draw offers is a needed reform of current tournament chess, but I think that another step needs to be taken: penalize the higher-rated player of the drawn game by awarding him or her a smaller share of the split point.
A difference in chess rating implies a difference in chess skill - the higher-rated player ought to win. If that player fails to win, he or she should not receive the same reward as the lower-rated player.
This would eliminate the financial incentive for draws since the possibility of ties would be greatly reduced. Higher-rated players would be forced to play to win in the final round since they could not split the prize money via the "GM draw".

So... in Jon Jacobs' opinion, anybody who has never been scammed in a park by a guy with 1 minute on his clock has no opinion worth anything and is thus ineligible to participate in a civilized discussion. Seems like it should be the other way around, as he was the one that got ripped off. " (If you've ever played a park hustler for bucks, you'll know what I'm talking about. If you haven't, then I don't care what you think, and neither do the readers worth reaching or worth listening to.)" MAN, some people from NYC make me mad.

Evan, it sounds like you're proposing something along the lines of "whoever has the best tournament relative to their current rating should win". For example, the winner is chosen based on highest rating gain or best differential between tournament performance and pre-tournament rating. This would be madness I think; it's nice when a 2500 GM has a 2700 performance, but does he really deserve the title over a 2700 GM with a 2800 performance? Tournament winners must be judged on an absolute scale. Good showings by lower rated players are already rewarded with rating points and (hopefully) better invitations in the future.

Jacobs' posts have become increasingly arrogant over the last year or so. I don't expect the trend to reverse anytime soon.

kgd and Cynical: the arrogant posters are the ones who don't play chess (not necessarily you), yet not only spout opinions about it, but expect their opinions to actually influence the rules governing top-level chess and/or amateur chess.

I can see where the two of you are coming from when you pin the "arrogant" label on me instead of on the opinionated know-nothings that I often expose. You guys have spent way too much time on the Internet, which has elevated the Rush Limbaugh ethic of "all opinions are created equal," and no one knows or bothers to evaluate anyone's level of knowledge or experience.

Perhaps you would both feel more comfortable posting comments on the USCF blog, USCF Forums, where the powers that be have decreed something called an AUG (Acceptable Usage Guidelines). In a nutshell, USCF's AUG says anyone who writes anything impolite about anyone else, will get sanctioned.

The fact is, the act of making and expressing (sometimes frank) judgments about others' opinions is not arrogant - or even impolite; in fact, it is all but inseparable from the process of reasoning/arguing in public.

For instance, we have all seen - on a variety of chess sites, but here on Dirt most of all - people call for anti-draw rule changes like making stalemate a win, or adjudicating all non-matable endings as a win for the side with a material advantage.

Quite simply, anyone who makes such a recommendation thinking it will "improve" the game (and many of these "reformers" seem to sincerely believe their proposal will eliminate draws at little or no cost), is blissfully ignorant of what makes chess chess, and why so many people love it. And no one should be ashamed of calling it like it is.

Go back and read the context for my comment that you were so quick to condemn. It was made in reference to people who think a decisive blitz game between GMs is somehow more exciting, more deserving of attention from chess lovers, than a drawn conventional chess game.

And the guy I was arguing with when I wrote it - Jens Petersson - even explicitly agreed with me on that specific point in his next comment (about the undesirability of Armageddon playoffs).

It seems that there isn't any acceptable solution so far in resolving this quick draw issue "during game play", and there might not be any. It might be for a valid reason from the players' strategy perspective (e.g. to secure GM norms for Masters, or during the game prior to rest day, etc). Eventually the players care more about winning/cash prizes than public opinion (including from patzers like myself).

However, "post game play" consequences might work. The following has been proposed before: "Quickdrawers should get less invitation to big tournaments."

Do I care what category a tournament is? Not really. If top ten invitees results in 70%+ draw rate, why not get some from the top 20, even top 100 that have more fighting spirit in the following year's tournament?

Of course, the challenge is whether tournament organizers would get the sponsors/funding needed to run the tournament if, for example, they invite Fridman (#100 in Jan 08 list) or Vladimirov (#101) instead of Svidler (#5) or Leko (#8).

Another downside is that the games might be more prone to blunders (assuming that top 10 players knows chess much better than the rest of the top 100). I don't know the solution to this, except that it would be a hard lesson learned for the loser.

Personally, I'd rather study a 1-0 or 0-1 game, than a 1/2-1/2 game (actually, only those kind of games are the one I replay and study from big tournaments). Maybe if I got promoted to Master Patzer, I'd appreciate 1/2-1/2 games more.

Jon, even if you were right as often as you think you are (all the time) you could still be arrogant.

About reducing draw percentages, i think the solution is to either:
1. Invite players with some diffrence in rating positions
2. Give monetary incentives for wins.
3. Invite players who play "fighting" chess.
I don't think draws should be banned, they do serve their purpose.

Interesting post GeneM!

One comment grated:
"Yes the outcome of a chess gamed played perfectly by both colors is a draw".

I disagree. I have seen this claim made often yet there is never any evidence to support the claim.

Maybe because it is assumed to be axiomatic?

The claim *may* be true for very weak players who may approximate random move generators. However as skill level increases there is plenty of evidence that the win percentage for white predominates and is quite marked at the GM level.

That colours alternate in Swiss tournaments and matches is tantamount to saying chess is not *fair*, in the sense that having the first move is advantageous.

Therefore it is false to claim the logical outcome of a game between equally rated players is a draw. Current statistical evidence disagrees.

To use this false claim to excuse fightless "GM draws" is dopey.

(Incidentally, I'm not saying this is your view GeneM!)

Well Jon, maybe ”the obvious point” of me agreeing with you on Armageddons was ”to get you to come out” with a less arrogant tone in the way you argue? ”You just failed to do that – again.”

Jens ;-)

The silent majority here feels there is nothing wrong with an 8 move agreed draw to conserve energy and increase chances to win an event. It is your game, play it as you want to.

dysgraphia at January 8, 2008 03:31 wrote:
{
"Yes the outcome of a chess gamed played perfectly by both colors is a draw".
I disagree. I have seen this claim made often yet there is never any evidence to support the claim.
}

In John Watson's 1998 book titled "Secrets of Modern Chess Strategy", on page 232b under the mini-heading "Remis?", Watson claims that all grandmasters agree that "the proper result of a perfectly played chess game" is a draw. Watson says "Of course, I can't prove this, but I doubt that you can find a single strong player who would disagree." According to Watson, Kasparov says "chess is a draw".

Yes White's unfair advantage does cause White to win more games than Black wins. That only means there is less room for imperfection when playing Black.

Guest at January 6, 2008 20:33 wrote:
{
"the outcome of a chess gamed played perfectly by both colors is a draw"
Says who? Chess has not been solved, so it is not possible to determine the result of a "perfect play" game.
}

Gene thanks for the reference.

I thought you were stating as a fact that a perfectly played would result in a draw. You should really re-phrase your comments show that this is just an opinion. (e.g. It is the opinion of Watson / Kasaparov that a perfectly played game of chess will result in a draw).

And don't always take what Kasparov says as gospel, IIRC he once said "no computer will ever beat me"

GeneM, thanks for the reference to Watson's book and the quotation from it...appreciated. Watson uses the "appeal to authority" argument rather than provide any proof as such.

I wonder how many of the GM's he asked play for a win with white, draw with black......ie act the opposite of their belief!
I wonder how they explain the evidence that as rating level increases (ie play presumably gets closer to perfection?) the preponderence of white wins increases?

When and if 2 humans play who have reached say Elo 4000 will white always win?...or computers for that matter?

In Watson's quote one could replace the reference to "the proper result of a perfectly played chess game" with "the earth is flat" in an earlier era and find all experts agreed at the time.

Numerous Nobel Laureates told Leo Szillard his patent for a nuclear reactor was baseless...."moonshine" the great Rutherford said, Einstein pooh poohed it too. A handful of years later Szillard and Fermi et al built the first pile in a disused squash court at Uni of Chicago!

It can be useful to know what the authority figures say on a subject...as a starting point for one's own analysis not for hero worship.

I notice the just completed Reggio Emilia had 67% draws, 29% white wins and....4% black wins.
OK the high draw rate could support Watson et al....but then the disparity of white to black wins says?....one tournament alone proves little of course.

I would be interested to see graphed cumulative results of W/B win % for similarly rated top GM's when they play each other. I suspect (guess)monotonically increasing with rating.

We are nowhere near perfectly played games of chess, so why worry about that scenario? At the moment the decisive percentage is what, a fourth of the games? Throw in another third of draws which are fighting and a small percentage which is unavoidable and what you are left with is the simple realization that the game of chess between two Super-GMs still has tons of entertainment to offer a fan. The people who are unhappy with the current draw percentage meanwhile can entertain themselves with complaining that every fight does not end in a bloody knockout, that the goalie sometimes does a good job blocking the opposing team's attacks or that 4 balls in a baseball leads to a walk rather than a game of "armageddon" baseball where opposing teams try to club each other to death with bats.

I think that there is overwhelming evidence that chess is a draw with best play. Not once have I seen a strong player annotating a game say "after 1 e4 it appears black is lost!". I think the statistics clearly point to the fact that white has a hard time converting his first move into something more than a slightly better position.
Some people seem to confuse the issue when they consider a theoretically drawn position to be equal in play. Most if not all top players would say that white has an advantage because he moves first. This is why some players try to draw as black. This does not mean that they think white is winning!
In my database of elo 2700 games. White wins 28% of the time. Black wins 16%. This seems consistent with white having a small advantage from the start.
Regards

The only real problem is fighless draw. There are wonderful examples of terrific defenses to gain a draw, or beatiful combinations (stalemate combinatios and so on).
Changing the scoring system will destroy the essence of chess.
Only we need is something like Sofia rules.
Best regards
Ramiro, Argentina

"... or that 4 balls in a baseball leads to a walk rather than a game of "armageddon" baseball where opposing teams try to club each other to death with bats."

This is the cleverest comparison I have seen in a long time. Nicely written!

Bones wrote:
"In my database of elo 2700 games. White wins 28% of the time. Black wins 16%. This seems consistent with white having a small advantage from the start."

Thanks Bones for providing this stat from your db.

Your figures suggest to me a huge advantage to white...almost 2:1 compared to black.

This perhaps does suggest the game is not *fair* in the commonly accepted meaning.

*If* this is the case then players esp professionals will inevitably factor this into their play.

Hi Yuriy, A nice word picture you give!

With respect, the debate has never been about draws as such as we all know there are abundant wonderful fighting and creative draws in the annals of chess.

The debate is about the fightless draw, "GM draw" or "non-chess" draw. Playing a few desultory moves of theory and shaking hands is not chess.

Using your baseball analogy if it suited two teams to have a few swings then agree to go home would this still be "baseball"....how long before baseball's governing bodies stepped in with sanctions....two milliseconds?

Jon Jacobs wrote:

"Quite simply, anyone who makes such a recommendation thinking it will "improve" the game (and many of these "reformers" seem to sincerely believe their proposal will eliminate draws at little or no cost), is blissfully ignorant of what makes chess chess, and why so many people love it. And no one should be ashamed of calling it like it is."

Jon, I like your posts, they are always a good read.
However, in my view chess has evolved and continues to do so, the "reformers" have had some pretty spectacular results in the past so who's to say what the future holds?

I suspect the reformers in past eras copped plenty of flak too.
Imagine!... some klowns wanted to introduce changes to the strength of the Queen and make it the strongest piece instead of the weakest...what a crazy reform...it'll ruin the game as we know it ....

Then some weirdo wants to bring in the en passant rule and stop us having all those lovely blocked pawn chains and constipated draws...it'll drive players away from the game in droves!

Not satisfied with the damage so far some drop-kick wants to introduce clocks!!...fergawdsake is this a Swiss scam to sell double-sided clocks?...chess doesn't need clocks...if my opponent or I want to sit on a move for 36 hours so what?...that's chess as we know it and you crazy reformers can like it or lump it!

Then some fruit-loop wants the players to write the moves down on a sheet of paper as tho we don't have enough to do fending off an opponent!

And the touch-move rule...what's wrong with fondling all the pieces before I move one?...what an insult!!

....and so on. The history of chess has many examples of major and minor rule changes as the game evolved. Players try to take advantage of any new situation, fairly or unfairly as their ethics dictate, if disquiet rumbles enough (eg computer cheating), the game tries to adapt.

IMHO, there is and has been disquiet over the phony draws for a long time with some viewing it as akin to cheating eg a self-awarded rest day or rest-round.
Fischer campaigned vociferously against the phony draws and made no secret of his view that it was outright cheating.

A recent paper coming out of Washington U Economics Dept , St Louis,is interesting:

Did the Soviets Collude? A Statistical Analysis of Championship Chess 1940-64
by Charles C. Moul and John V.C. Nye

They used one of my pets, probit analysis and their conclusion re Fischer's allegations was that there was a 75% likelihood that in the 5 relevant Candidates matches Soviet players colluded against Fischer.
Inter alia they commented that Fischer was not strong enough to be a serious title aspirant at this stage of his career anyway and that Sammy Reshevsky had been the greatest victim of collusive behaviour.

So the supporters of the "phony draw", "GM draw", "fightless draw" or whatever one wants to label it can take comfort that their cherished "draw" has possibly led to some phoney tournament outcomes and maybe even phoney "champions".

This disquiet will not go away until it is addressed and effectively dealt with. Some of the suggestions for tinkering with the scoring etc etc are complicated and likely to be ineffective if not just plain silly....our chess ancestors would laugh at our feeble efforts and point to their own decisive interventions.

dysgraphia: In my quotation you began with, I was referring only to specific kinds rule changes such as: stalemate is a win, or K+N vs K is a win, or K+P vs K is deemed a win for the superior side. These changes would eliminate draws while indisputably making chess less interesting overall. (Aside, for Cynical and kgd: Go ahead - complain about the uncompromising tone of that last statement. Which is to say, make my day... by suggesting there might be room for anyone who knows anything about chess to think that a "whoever dies with the most toys wins" approach would make the game more interesting, draws or no.)

So, it's misleading to equate that statement of mine with the idea that any sort of reform in chess is bad. In fact, my article in last month's Chess Life (http://main.uschess.org/content/view/8043/426/ ), in which I urged hidebound types to dump their obsolete analog clocks in favor of delay-capable digital models, made me a target of flaming on USCF Forums from two readers who branded ME a "reformer." (See http://main.uschess.org/forums/viewtopic.php&f=24&t=5840 ; start at the top of page 2 with the comments from "Condor" and "ILfish", and enjoy the fireworks.)

And in an earlier comment on this thread, I indicated one anti-draw proposal I could live with, which was introduced by someone on another Dirt thread. (In essence, it called for docking a player's tie-break total for each short draw they made in a tournament.)

Finally, I'll point out, as many others have, that Fischer's (now well substantiated) complaint about the Soviets involved COLLUSION - not draws, or even pre-arranged draws, per se.

His whole beef was that the Soviets' prearranged draws among themselves didn't consist of two players avoiding a fight to advance their own individual interests; rather, the deals were part of a team effort by the entire Soviet squad, aimed at ensuring that a favored Soviet player would finish 1st.

This is entirely different than the argument over draws today (whether pre-arranged or not). So, invoking the Fischer situation as evidence in favor of anti-draw changes today is a complete non-sequitur.

The statistics dysgraphia quotes to discredit the idea that "a perfectly played game results in a draw," in fact do not discredit that idea at all. His claim that they do, is a basic error in logic.

All they show is that White starts with an advantage. Just about everyone, from super-GM to novice, agrees that's true. That's why pretty much all of opening theory starts from the premise that Black's task is to neutralize White's initial advantage of having the move.

Whether a perfectly played game should end in a draw, or a win (presumably for White), is logically a different question...which boils down to, if both sides play perfectly, can Black neutralize White's opening advantage? As Gene and others noted, virtually all knowledgeable people believe the answer is yes.

Neither proposition is proved - that White enjoys an inherent advantage due to moving first, or that a properly played game should end in a draw.

But the statistics dysgraphia quotes (and that agree with more comprehensive stats I've seen quoted on Dirt and elsewhere) clearly and strongly support both propositions. Most high-level games are drawn, and the higher in strength you go, the higher the frequency of draws. (Games between 1200 players, or even 1700 players, almost never end in draws.) And, of the minority of games that do have a decisive result, White wins far more often than Black.

Contrary to the conclusion dysgraphia drew, these figures, as a matter of logic, do not imply that a perfectly played chess game would end in a White win. If anything, they are evidence for the reverse, that it should be a draw.

As an aside, I've seen (people who claimed to be) math professors really go at each other on old Dirt threads over the logical/philosophical issue of what would be needed to "solve" chess - to determine "perfect" play and what the result of a "perfect" game should be.

The only thing that got clarified from those amazingly heated (and sometimes incomprehensibly technical) debates was that a solution is likely to be many decades away - even centuries away, perhaps. The difference in computing power between what's needed to work out an 8-man tablebase and to construct the holy grail, the hypothetical 32-man tablebase, is simply enormous.

So, for the rest of all of our lifetimes, we're stuck with relying on human reasoning and probabilistic judgment, rather than scientific proof, to decide whether chess is logically a draw or not.

Jon wrote:
"Most high-level games are drawn, and the higher in strength you go, the higher the frequency of draws. (Games between 1200 players, or even 1700 players, almost never end in draws.) And, of the minority of games that do have a decisive result, White wins far more often than Black."

Thx for the useful reply Jon!

Perhaps it did not come thru clearly what I was on about but it is encapsulated in your last sentence which I have reproduced above and which agrees with my post.

I have no quibble with the demonstrable fact that the majority of high level games are drawn, everyone realizes this.

My point was that there is a growing disparity of White to Black wins as we ascend the rating tree.

Therefore if chess is a "fair" game then I am curious why is this so?
Then the logic follows like this: as ratings increase, and I speculated at say Elo 4000 for discussion purposes but key in any value that may be achievable by humans or computers you like, then the outcome of chess games played at this level may be only a) Drawn or )b White wins.

As the opening discoveries from these games permeate down the rating tree then I would hypothesize that the super_GM W/B win ratio will move down the rating tree to the not-so-super GM's and so on.

This hypothesis is back-testable even now.
That is: "For a given rating band with similarly rated opponents the win ratio W/B has been steadily increasing".

Anyone with a suitable db to test this?

Also if anyone can generate them I would be interested in figures for the average no. of moves and SD for W wins compared to B wins, assuming of course players were "similarly" rated.
One of my few remaining functioning neurones dimly remembers that rating differences of up 16 points may be statistically insignificalt but any other defensible figure is welcome.

This leads me logically of course to the next step: how will chess "reformers" deal with this situation?...how to redress the balance back towards black or do we just ignore it?

One attempt has been the Armageddon, the time handicap route but I don't see this as satisfactory....it must tend to favour certain gifted blitz players.

Jon, you wrote:
(snipped for brevity) "....can Black neutralize White's opening advantage? As Gene and others noted, virtually all knowledgeable people believe the answer is yes."

Then perhaps someone from this mass of logically competent and knowledgeable "virtually all" can explain why is their such a disparate W/B win ratio and why does it increases as the rating band increases?.

I would prefer a logically argued analysis rather than the droll "appeal to authority"..."have faith my son" type of reply.

Where is the logic in saying virtually all chess experts say B can neutralize W's opening advantage
when the stats say this is just not happening in the real world. Sounds quite illogical and collective denial.

http://www.chesscafe.com/dvoretsky/dvoretsky.htm
http://www.chesscafe.com/text/dvoretsky88.pdf
-------

It seems like some unknown crazy person named Dvoretsky hacked into a web site I never heard of until tonight, a "ChessCafe.com", and posted a new article today that says extremely crazy things.

One claim I gather from Dvoretsky's article is that --
-- many short draws are not unfought, they were fought hard but only during home analysis and memorization. These draws are an inevitable result of reusing the same one starting position for every game.
Okay I admit that is fact, the opposite of crazy, but listen to his next idea...

Dvoretsky says --
-- an alternative form of chess should be added into the brotherhood along side today's chess.

Dvorestsky says this new form could be FRC-chess960, or Dvoretsky prefers a compromise where dice determine the first two plies.

[GeneM: Bobby got one major thing wrong in FRC -- it turns out to be a bad thing to have bishops starting on corner squares (only one way to develop). And 40% of setups have a bishop on a corner square. Just ban those setups.]

Dvoretsky says that with the heavy analysis of the chess opening position "the very concept of the game is destroyed".

Dvoretsky says "let’s dream ... that opening information has all disappeared ... How much fuller and more interesting the chessplayer’s life would be!".

Dvoretsky says "chess in its present form is moving gradually towards a dead end".

Dvoretsky says "many of my friends and students have taken part in the traditional Fischer-random tournaments in Mainz. Most of them liked the new game".

Dvoretsky noted that some Rapid FRC grandmaster games from Mainz Germany showed tactical errors as early as move-pair 6. Dvoretsky implies that he believes this rate of errors would not decline much, even if the grandmasters played with long time-controls and played in several such tournaments each year.
I disagree with Dvoretsky about this. I believe necessity is the mother of invention, and that cool new opening principles will be developed.

- - - - -
Dvoretsky says "such an enormous change in the rules should be examined from all sides and tested, with all aspects considered".

[GeneM: Or as I very recently put it -- alternative implementations of Abstract Chess, ones that are as similar to chess as possible but which avoid its problemmatic aspecs, should be proposed and studied to see what we might learn from them. Alternative ideas like Dvoretsky's Dice-2-Moves Chess, and FRC, are what ChessBase.com should call for from its web visitors; instead of the folly of trying to fix problems within the strict confines of unchange-able chess rules.]

I was curious if stronger players win more with white than weaker players. Here are the results from my database search:

2700+ W-28% B-16%
2699-2600 W-27% B-16%
2599-2500 W-26% B-17%
2499-2400 W-29% B-20%
2399-2300 W-33% B-24%
2299-2200 W-35% B-28%

If anyone is proving that chess is a win for white, It certainly is not the Super GMs. However, It does look like it takes all of black's efforts to neutralize white's advantage of moving first.

I have no problems with a drawn game per se - the outcome is logically irrelevant to the quality of the game. I am one of those sad people who actually enjoy Kramnik's draws with the Petroff! There is this idea that if you play the petroff it will draw itself. You have to be a great player to make the Petroff work like Kramnik does or Gelfand did recently - and even Kramnik came very close to getting a lost position with it in Mexico. For those people who do not think that a perfectly played game of chess is a draw (which IMO it very obviously is) try playing high level cc and you will rapidly change your view! One of the reasons that it is (imo) a draw is that that there are those asymetric draws some stalemte dependent some blockade dependent. You can sweat your guts out to get a material advantage then bang your head against your opponents drawing possibilities.

It quite common at cc to abandon e4 because playing against the Petroff or Caro can be futile. None of this is relevant to the problem of short draws which should be solved by banning communication between the players during the game. We have not reached the stage where chess knowledge has reached the level where perfect chess can be played over the board in classical chess ie there is a pretty sure route to a dead drawn position (ie prospectless between top players.) However it will eventually happen against e4 - for certain. Leko could not quite manage it against Kramnik in there match when he tried to draw his way to victory against e4 losing to the advance in the caro but the time is coming....

The more serious problem (at the very highest levels) is the one raised by Dvoretsky in chess cafe which boils down to chess becoming my computer prep v your computer prep and his solution a balloted single pawn moves 1 for each side at the start of the game. Under this system all opening thory is out of the window but all the principles linekd to the classical starting position of the pieces, remain. Again irrelevant outside of the elite of top chess professionals but very interesting and quite possibly the future of chess

Draws are not a problem. Pre-arranged games and fight-less games meant to manipulate tournament standings and player ratings are the problem.

The solution is very simple: force players to play until someone wins, by playing progressively shorter time controls (alternating colors) until someone scores the point.

The main objection to this idea is that the "quality" of play "suffers". I'd rather see players performing at 80% capacity than drawing in 15 moves. Wouldn't you?

I'm amazed by patzers talking about "bad chess" bewtween, say, Kramnik-Topalov or Svidler-Kamski if they are forced to play 15-minute playoff games (or 10 minutes or even 5 minutes). The truth is that, at this point in history, a 30-minute Leko-Shirov game is much better, technically, that any of Paul Morphy's brilliancies.

I'd rather see games with errors and drama than lifeless accurate, computer-like draws between Svidler and Leko.

Make them play. It's the only sensible solution.

Advantages:

1. It doesn't mess with the quality of the game - the initial game is played at classical time controls

2. It doesn't mess with chess rules or tradition - no crazy variations of opening setups or any of that silliness

3. Does not mess with ratings - only the game played at classical time controls counts toward rating

4. Makes sure players can's safley pre-arrange draws for the last (or any) round of important tournaments when fans want to see real action.

Bones wrote:

"It does look like it takes all of black's efforts to neutralize white's advantage of moving first."

Many thanks Bones for doing the db search and for providing the table....appreciated!

The figures suggest that Black's win% are in constant decline as the rating scale increases as I expected.

Any chance of imposing on you further Bones and reducing the granularity a little as the rating spread for each record in the table is likely to violate the premise of having approximately equal strength players competing?....maybe 40 point bands?
With that data table I will graph the rating vs win% and see if and when it extrapolates to zero for B.

So far your table does suggest that at some point above current maximum ratings that the outcome of a chess game will be either a) Draw or b) Win for White.

Does your db enable you to generate the number of moves played? I suspect the average number of moves for a B win is greater than for a W win....ie not only is it less likely for B to win
as you ascend the rating scale but that when those wins are achieved they are harder work.

This list includes classical, blitz and rapid games.

Rating Win Win Moves
2750-2800 W-27% B-12% W-44 B-46
2700-2750 W-27% B-16% W-45 B-45
2650-2700 W-28% B-18% W-46 B-45
2600-2650 W-26% B-15% W-45 B-46
2550-2600 W-25% B-15% W-44 B-44
2500-2550 W-26% B-17% W-43 B-44

Regards

Hi Jon,
Thx for the reply. My apologies if I misconstrued your position on fightless draws.
I will have a look at the link you gave for the USCF forums.

You wrote:
"Finally, I'll point out, as many others have, that Fischer's (now well substantiated) complaint about the Soviets involved COLLUSION - not draws, or even pre-arranged draws, per se.

His whole beef was that the Soviets' prearranged draws among themselves didn't consist of two players avoiding a fight to advance their own individual interests; rather, the deals were part of a team effort by the entire Soviet squad, aimed at ensuring that a favored Soviet player would finish 1st.

This is entirely different than the argument over draws today (whether pre-arranged or not). So, invoking the Fischer situation as evidence in favor of anti-draw changes today is a complete non-sequitur."

The article I drew attention to was:

Did the Soviets Collude? A Statistical Analysis of Championship Chess 1940-64
by Charles C. Moul and John V.C. Nye

so I was well aware that the allegation was COLLUSION...!!
Collusion was of course the alleged wrongdoing, one aspect of the alleged collusion involved pre-arranged games and fightless draws.

My point was clearly stated and that was that the fightless draw complaints of nowadays are nothing new, they were going on long ago and *may just may* have led to rigged tournaments at the highest level.

I wrote "IMHO, there is and has been disquiet over the phony draws for a long time with some viewing it as akin to cheating eg a self-awarded rest day or rest-round. "

I gave Fischer's complaints of 45(?) years ago as an example of how long phoney draws may have afflicted high level tournament results.

How you claim this is irrelevant to today's debates about phoney or fightless draws escapes me!

Jon, do YOU consider phoney or fightless draws that may determine the outcome of high level tournaments or championships to be cheating?

(Irrespective whether by an individual or individuals acting alone or collusion by a team or officials).

Thx Bones for those new figures!

From your data it looks like I'm wrong re B wins requiring more moves than W wins.

Extrapolating the B win% suggests that at the rating level of 4077 Black will have zero win%.

Chess will be either a)Draw or b) Win for White.

My guess was 4000....oh well...wrong again!

As I remember it that article rather ignored the obvious possibility that good friends frequently prefer to draw rather than play and that good friends tend to come from the same country.

Of course short draws aren't cheating. Get a grip.

If you want to see fighting chess, stop paying for draws!
Stop paying players based on where they finished in the standings!
Pay top GMs attractive appearance fees - award bonus money as "beauty " prizes : "best game", "best game with black", "best ending" etc. etc etc
The rating system will sort out the skill ranking of the players -- in tournaments it is not so important where they finish as it is how they play.

Does anyone have soem stats about the draw rate between the Super GM level programs (e.g. Zappa, Rybka, Junior, Hiarcs etc).

Has a GM tournament ever been run using the number of wins as the deciding factor. I think that would be an interesting experiment to see how many games ended in a draw when both players had such an inncentive to play for the win.

What about a change to the scoring system that awards 0 points to both players if the draw is less than a set number of moves (e.g. 40 or 50). Agreed if the player are intent on drawing they will just pump out the requires moves and then draw, however unless there is collusion between the players, there is a that the game may progress to a position in which one of the two players sees an advantage in playing for a result.

No I don't consider today's "phony or fightless draws" cheating, regardless of their impact on the results of a tournament. Assuming of course, that no money changed hands - paying your opponent to give you a draw IS of course cheating. Neither of these contentions originated with me; they are simply laws of tournament chess, as practiced by all legitimate TDs I'm aware of (obviously excepting tournaments where "Sofia rules" apply).

Re Fischer vs the Soviets: My point was that playing as a team in what was set up to be a tournament of head-to-head competition among individuals, is what incensed him (and ultimately, moved the powers that be to change the Candidate selection system). Yes, phoney draws were a key tool employed by the conspiring Soviet GMs. Absent such conspiracy, the case against fightless draws simply wilts (well, your animus doesn't, dysgraphia - but please don't drag Fischer into it, his battle was fundamentally different). And no one contends that any similar schemes underly the GM draws of today.

The best available evidence indicates that chess is indeed a "fair game" (in the mathematical sense that it's a draw with perfect play). Dysgraphia, each statistic you point to hoping to undermine that message, in reality only reinforces it.

If one side having an easier time winning when play is less-than-perfect means a game isn't mathematically "fair," then tic-tac-toe isn't a fair game either! ("White" also has an advantage of sorts in tic-tac-toe: he can still hold the "draw" even after failing to make the optimal first move - whereas the second player loses by force if he makes a misstep on any turn.)

Both your message and your tone remind me of Clint Ballard. You and he each seem to want to solve with one blow the "draw problem" and the "White advantage" problem. You each are entranced by statistics about win and draw percentages. And in the way you argue, you're both exceptionally good sports: polite, cheerful, and with a long fuse. (In case you really don't know about Mr. Ballard, your fellow readers will be glad to enlighten you; and of course there is Google.)

Two further points about extrapolating toward superhuman rating or strength levels:

Ratings are condensations of probable results; they aren't designed to chart or benchmark quality of play. Therefore the linear regression you used, superimposing an estimated ratings line over an imaginary "quality" line ascending toward theoretically perfect play, is logically flawed.

I read somewhere that Engine tournaments produce very low draw rates. I don't recall what the ratio of White wins to Black was; but Black did win some games, and the ratio might even have been higher than in human GM competition. Anyway this says next to nothing about the outcome of a perfect chess game, because even the strongest engines still don't come close to perfect play (sans tablebase). Just ask anyone who's worked with them in some depth, such as a top-level correspondence player.

A draw is he natural outcome of the game at the highest level. If the intent is to minimize draws then giving each player .4 of a point would elimate them.

Jon: In regards to the draw rate between engines, I wasn't suggesting that they play perfectly, but rather that they play at a level similar to GMs; however engines don't (as far as I know) offer short fightless draws.

Thus we could roughly estimate what the draw rate would be if fightless draws were eliminated from GM tournaments.

Just wondering if the difference was significant enough to warrant bothering to try an eliminate short draws (i.e. would it result in more decisive games or only longer draws?)

If you want to see good chess, watch good chessplayers play and stop whining about their results. Honestly, about two or three exciting (=seriously flawed) games per round is the maximum I can digest; I don't really see why anyone should care about the length of the rest.

Jon Jacobs wrote: "No I don't consider today's "phony or fightless draws" cheating, regardless of their impact on the results of a tournament.
Assuming of course, that no money changed hands - paying your opponent to give you a draw IS of course cheating."

This attitude surprises me! There may be more to gain than money though eg qualification to a higher event, a FIDE title etc etc.

Re Fischer vs the Soviets: Jon, you invoke Clint Ballard and tic-tac-toe in response to me giving Fischer's complaints of phony draws as an example of the practice having gone on for a long time.

Well!....with all due respect to you surely this is feeble and irrelevant!...a sure sign to me you know you are on shaky ground and want to divert the discussion to Clint Ballard or tic-tac-toe.
I'm happy to but perhaps not relevant to this thread..."Fightless Birds".

You wrote, eventually!! "Yes, phoney draws were a key tool employed by the conspiring Soviet GMs."

Agreed!!

You wrote: "The best available evidence indicates that chess is indeed a "fair game" (in the mathematical sense that it's a draw with perfect play)."

I read and hear this repeatedly yet why the marked disparity between W versus B wins?

The proponents of this claim always use the convenient get-out of "perfect play"....of course if the evidence doesn't fit well...it wasn't perfect play!!
The best available evidence is that W/B win ratio markedly favours W esp as rating rises.

You wrote: "Dysgraphia, each statistic you point to hoping to undermine that message, in reality only reinforces it."

In that case Jon have a look again at the tables Bones posted. I see possible trends, you don't....I can live with that!

BTW, Tic-tac-toe is much more interesting when played in reverse ie you have to force your opponent to place three in a line. Bright kids quickly bore with the std game but get entranced by the reverse game.

You wrote: "Ratings are condensations of probable results;"

Wrong! Huh?... are you aware of how Prof Elo devised his rating system?

Ratings are NOT a "condensation of probable results at all"!.....ratings are calculated using an algorithm from ACTUAL results.
Here Google is your friend.

You continued: "they aren't designed to chart or benchmark quality of play."

Who made this claim?
Well, your rating at any given time IS just a "benchmark" of your ACTUAL results against other rated players up to that time.
If it doesn't "benchmark" the quality of your play in some way then what do you think it does?....are you saying high quality players could have very low ratings or low quality players could have very high ratings?

You wrote: "Therefore the linear regression you used, superimposing an estimated ratings line over an imaginary "quality" line ascending toward theoretically perfect play, is logically flawed."

Not at all!....it's called trend analysis or forecasting. Certainly I agree I would not sell one of my properties and bet the proceeds on a chess futures market.
This is a chess blog discussing "Fightless Birds" not the monthly mtg of the ethics ctee of the American Statistical Assocn.

A proper discussion would require me to have the actual data, not merely the aggregate tables of Bones. I don't have such data but am happy to analyze any provided to me.
If you are interested in this stuff there is a nice programmable public license stats package, R, (legal knock off of S+). Three Harvard profs have kindly made available an additional superb module named Zelig (yes!....after Woody Allen's character).

With R/Zelig one could use probit or logistic regression on the actual data rather than the aggregate tables.

I don't know what line of work you're in Jon but I do quite a bit of risk analysis and forecasting in areas where data is much less certain than the data in chess databases and, yes, do back myself with $$.

You wrote: "I read somewhere that Engine tournaments produce very low draw rates. I don't recall what the ratio of White wins to Black was; but Black did win some games, and the ratio might even have been higher than in human GM competition."

Well!....this starts to look like you agree with me!!.....but then you write:

"Anyway this says next to nothing about the outcome of a perfect chess game, because even the strongest engines still don't come close to perfect play (sans tablebase)."

Wow!....this is weird logic here!.....the cop out again...ouch!...not perfect play.
Jon, please look at the TREND here described by yourself.

My thesis was: as we ascend the rating tree ie approach "perfect play" chess is either A) Draw or B) Win for White.

You wrote: " Just ask anyone who's worked with them in some depth, such as a top-level correspondence player."

I played cc at a reasonable level (c 2600) BEFORE computer cheating became widespread.

Computers, no matter how strong, always suffer from the "horizon effect". If I sensed my opponent was a computer cheater in a cc game I went for positions where I could exploit this. It's a while since I played cc seriously. Maybe this no longer works.

A brief reply; I haven't the time (or the justification) to respond to each and every point in the above, rambling post.

You wrote: "My thesis was: as we ascend the rating tree ie approach "perfect play" chess is either A) Draw or B) Win for White."

I have no quarrel with that. And the statistics you and others presented clearly indicate that the most likely answer is A) Draw. Of course, it must be one or the other, and the correct answer isn't knowable at this time (and very likely won't be knowable within our lifetime).

As for "fair game", mea culpa. I just now looked on numerous sites for a rigorous definition, and discover that I unwittingly confused the issue by injecting this term, which seems inapplicable to chess altogether (by "inapplicable," I don't mean chess is NOT a fair game; rather, chess doesn't seem to fall within the category of games that could be properly classed as either fair or unfair).

I'm familiar with the term from finance theory. But since finance theory often assumes players make optimal decisions (in aggregate at least), I made the unwarranted assumption that in a fair game each side has an equal chance of winning, GIVEN OPTIMAL DECISION-MAKING. In fact, my checking revealed that the entire concept of a fair game may be meaningless in the context of chess - because each and every example I found (including in mathematical and financial dictionaries) gave only games of chance, never games of skill or even partial skill (i.e. backgammon). In a game of pure chance, there is no such thing as optimal play (although there may be optimal betting strategies, independent of the "play" itself).

If the term is applicable to chess at all, then based on the definitions I saw, chess would NOT be a fair game, because all the examples relied on actual results (rather than assuming best play) to gauge the players' expectations of winning. Nor would tic-tac-toe be a fair game, I might add.

This is of course irrelevant to the question that dysgraphia has been debating against Gene, myself and most other posters here, of whether or not chess is more likely a draw or a White win with perfect play.

Are we certain that the in the starting position white is not in zugzwang?

Jon wrote: "A brief reply; I haven't the time (or the justification) to respond to each and every point in the above, rambling post."

Apologies if I rambled...I merely responded to just some of your comments so am I the rambler or ramblee? (grin!).

Some progress at last!...glad you now see chess is not a fair game!! The first mover has a significant edge as all stats clearly demonstrate.

Imagine if in other professional competitions the first mover had such a marked advantage (W wins roughly twice as often as B)...would there be cause for change?

GasolineToTheFire made the incendiary comment: "Are we certain that the in the starting position white is not in zugzwang?"

Mostly.

Depends on the rating difference between W and B!

Re Jon's comment as to whether we will see in our lifetime whether chess is ultimately a draw:

Quantum computers are still relatively feeble, bit like microprocessors were circa 1970.
Only "toy" problems are being tackled at the moment.

An operational QC with several thousand qubits may "solve" chess in a few microseconds.

FischerRandom will take a few microseconds longer.

Depending on one's age, some will see this in their lifetime.

Any takers for your lazy $10K at 100/1?...chance to leave your kids/grandkids $1M...

Reminds me of a guy I know who makes a substantial living as a sports bettor. When Steffi Graf and Andrea Agassi were making their names in tennis and while both were involved with other partners he took out a wager worth to him $5M that they would one day marry and have a child who would win Wimbledon. I asked why. He replied he would leave the betting ticket to his grandchildren as their (possible) inheritance!

Enough rambling!...before you take me to task Jon, you brought up the issue of betting strategies!

Back to the thread issue of phoney draws!

I came across this item from Mig's blog in 2003.

(start of quote)

August 19, 2003
GMs Draw
You would think the sheer ignominy of pathetic non-games nicknamed "GM draws" would be enough, but no.

John Henderson brings to our attention this tidbit from the interesting notes of Jerry Hanken on the just-finished US Open in Los Angeles:

"We had our first test of the draw rule Monday in the 6-day schedule. In accordance with the Rulebook, we are requiring that players stay at the board and play at least 15 moves and 1/2 an hour before they can agree to a draw.

This is not a new rule. The Rulebook says "It is unethical and unsportsmanlike to agree to a draw before a real fight has begun." Penalties for such behavior are at the discretion of the TD. In keeping with this rule, we wrote and posted a notice to all players that this would be the way we enforced the rule.

Two GMs chose to ignore this rule and tried to draw in 1 move!

Admonished by International Arbiter Carol Jarecki, they returned to the board, played four more moves, and disappeared without turning in a scoresheet. marking the result as a draw."

For the rest of the story, go here and scroll down to August 13. What I really don't understand is why Mr. Hanken over-politely declines to name the culprit GMs. Why? Name them, shame them, nothing wrong with that at all.
If they choose to do it they should live with the repercussions of their actions.
Why protect them from their own destructive (to the game) behavior?
Celebrate them when they fight, criticize when they don't. It's the only way.

(end of quote)

There was no suggestion that any money changed hands between the "players" involved.

Were they cheating?

Apropos of the above: I'm proud to announce that I played a 10-move draw in a tournament earlier this month.

Pressman-Jacobs, St John's Masters, Jan. 2008, followed the well-known drawing line in the Pirc Austrian Attack introduced by Seirawan many years ago, in which black sacks his Q for perpetual check around move 10.

I can't recall if this was my shortest drawn game ever, or my first draw ever in under 20 moves (out of 2,000 or so tournament games). But it's my only one in a long, long time.

As for whether the GMs who drew quickly in violation of published tournament rules in that above-mentioned US Open were "cheating," that's a matter of semantics, since they obviously did violate the rules that were announced for that tournament. How and whether to punish their transgression, and whether to apply the very strong term, "cheating," are judgments best left to the TD.

By the way, I've noted above and elsewhere that I'm not opposed to Sofia-type anti-draw rules...although I'm not a strong booster of them, either. And it's very odd to see such rules applied in an Open tournament.

Also in response to dysgraphia's conviction that something needs to be "done" about the unfairness of White having a significant opening advantage: That is why a single game almost never constitutes a "match," in chess. In tournaments, yes each pair of players usually meet head-to-head only once (unless it's a double or multiple RR); but that's offset by the fact that they also play others, with attempts to equalize the number of Whites and Blacks each player gets. In other words, the appropriate unit for gauging fairness or unfairness is an ENTIRE TOURNAMENT (or match) - rather than each individual game.

ChessBase.com wrote, Feb 16 2008:
{
Morelia 08: Three wins in round one
16.02.2008 – The Mexican half of the Morelia-Linares tournament has started with a bang. Three of the four games ended in decisions.
}

A "bang"?! Why the extra excitement, if hard-fought draws really are just as desirable as any decisive game?
ChessBase is lauding more than the avoidance draws beyond merely the "short fightless" kind.

Mig's Jan 14 thread:
{
January 14, 2008
Corus 08 r3: Drawfest Monday
if you slept through this entire round of the Corus tournament you didn't miss a whole lot, ... all seven games in the A Group will finish without a winner today ...
Most of the games were fairly interesting
}

Decisive chess has a far more interesting narrative than drawn chess, almost by definition. Long draws are preferable to short ones, of course, but action in a chess tournament has to happen on the crosstable for it to make any difference in the sporting sense. The games don't take place in a vacuum. They are trying to win. A fascinating draw might be a great game and we should make sure to give attention to such games, but the decisive contests are always going to get more attention because of the inherent drama and because results matter. This is particularly true in news coverage, when the tournament is still in progress.

"Decisive chess has a far more interesting narrative than drawn chess"

Only to children, non-players, and the like.

Then you don't understand the word "narrative." I'm not talking about the intrinsic interest value of a single game. I'm talking about the tournament and the coverage of the tournament. Saying that a tournament with 50 interesting draws would have the same fan interest as one with 30 decisive games and 20 draws is ignorant. Everyone looks at the decisive games first, everyone pays attention to the results. Wins and losses automatically contain happiness and sadness, the ingredients of a storyline. A draw can contain both of those things, but many don't.

A player holding a draw in the final round with black to win the tournament is more interesting than playing the exact same game in the first round. That's narrative. Pretending you exist in a pristine world where only the quality of the chess moves matter is preposterous. And even were it possible such an attitude would merit only pity.

I would rather say I understand the word 'narrative' more subtly than you. Every chess game is a succession of small battles and following those is the interest.

It's true that a tournament with 50 drawn games would be rather dull, but that's a different question, and anyway we're not quite there yet.

Suppose the tournament arrived at a "critical" game between two players near the lead. The game lasts for 5 hours (not unfought).
Instead of a decisive outcome, Black "easily" holds a draw.

That is a dull "narrative". That Black had to play careful correct moves in move-pairs 17 and 20 does not rescue this narrative. The quote from UsChess.org:

- - - - - - - - - - -
http://main.uschess.org/content/view/8278/443/

"The critical game in round 4 was a showdown between Sevillano (who had 3-0) and Yermolinsky (at 2.5). The popular IM from Southern California was able to easily hold a draw with the black pieces ..."
- - - - - - - - - - -

Not exactly "the thrill of victory and the agony of defeat".

______________________________

At ChessBase.com, proposals to fix the draw problems in chess continue to pour in from all corners of the globe. Aside from Armageddon Chess, none achieve their objective.

(Armageddon is a point-fiddling scheme that, for draws, gives the whole point to Black.)

http://www.chessbase.com/newsdetail.asp?newsid=4522

- - -

Mark Dvoretsky (at ChessCafe.com) is the only person to recently propose an ON-the-board rule change to fix the draw problem (all other proposals are OFF-the-board).

http://www.chesscafe.com/text/dvoretsky88.pdf