What's the matter with draws?

by Donny Gray
12/20/2017 – At the beginning of the London Chess Classic, in December 2017, the first 19 out of 20 games were drawn. Only Fabiano Caruana was able to win a game. Folks writing about the high percentage of draws called it disappointing and absurd. But why? Why is having a draw in chess such a bad thing? Only three things can happen in a chess game: White wins, black wins, or draw. Why such a hatred towards one of the out comes? | Photos: Lennart Ootes / Grand Chess Tour

ChessBase 15 - Mega package ChessBase 15 - Mega package

Find the right combination! ChessBase 15 program + new Mega Database 2020 with 8 million games and more than 80,000 master analyses. Plus ChessBase Magazine (DVD + magazine) and CB Premium membership for 1 year!


The draw life


Many times getting a draw in a chess game is celebrated. Play someone 300+ rating points higher than your self and see if you are not happy if the game ends in a draw!! Of course you would be happier if you would have won, but you will be happy just the same with a draw.

The draws to me are just as amazing many times as the wins when I look at the games of a super tournament, or any tournament for that matter. For example if someone of any rating draws against Magnus Carlsen the current world champion, that is an accomplishment.  To not lose to Carlsen borderlines on a miracle these days. So why wouldn’t I want to see how someone survived their encounter with Carlsen? Why wouldn’t I want to see a game where two 2700+ players tried to annihilate each other? Just because it ended in a draw does not mean it was not a good game.

Fabiano Caruana and Sergey Karjakin by Lennart Ootes

The handshake the broke the draw streak: Fabiano Caruana and Sergey Karjakin | Photo: Lennart Ootes / Grand Chess Tour

To me if I am following a tournament and there is a high percentage of draws, this just makes the tension about the tournament outcome that much higher. Sooner or later some player will surge ahead. Sooner or later someone will win and take 1st place. Isn’t that worth waiting for? Why do folks just automatically think that if two GMs draw their game that they are not trying?

Over a hundred years ago, Capablanca and others bemoaned the fact that there were too many draws and soon the game of chess would be played out. How wrong was that belief? Extremely wrong! If Capablanca had had a chance after he made that statement to play matches with Fischer, Karpov, Kasparov, Anand, and Carlsen in their prime, do you think he would have changed his mind? I certainly think so.

Former world champion Kramnik once made the statement that the depth of chess scared him. If chess is so played out why would one of the best players ever say such a thing? There may be periods of time in chess history that draws are common among the grand masters but there is always one of them that pushes ahead and takes chess to new levels. Carlsen is a great example.Magnus Carlsen by Lennart Ootes

Always trying to assign new ways to score black extra points for draws is in my opinion bogus and irritating. Just a case of organizers and others to try to force their way into the results. By making white take chances and play incorrectly they claim is the way to go. I would rather see a hard fought draw any day of the week than some wild crazy game that is full of mistakes just because some organizer or some group of people wants to tweak the game of chess.

As for myself I am for sure old school. At 62 years of age I have seen a lot of changes in the way the game of tournament chess is played. For one there was no such thing as scholastic chess at least in the way it is now when I was young. My first rated tournament I was at the ripe old age of 18. Now days you are almost considered over the hill if you don’t have a GM title by age 18. It was not until I was 35 that I made it to the ranking of United States National Master. There was no such thing as computers or DVDs, just books and tournaments. We learned chess the old fashioned way!!

Of course not all changes in tournament chess in the past 25 years have been bad. What we did have in those days was no increments time controls and plenty of adjourned games. There are several 9-10 hour games in my history due to never ending games thanks to no sudden death time controls. If you made a time control there was an never ending supply of others. If you ran out of time to finish the game and you had to start the next round, you adjourned and finished it later. There was no such thing as increments. You played and if you ran out of time you lost.

Chess is about who is the more talented, fastest in calculations, and better prepared.  Adding increments to time controls in my opinion does nothing to change who will win.  Still the players that are more talented, faster in calculations, and is better prepared will still win.

Chess is fine just the way it is. We don’t need to realign the pieces at the beginning just because you don’t want to study openings. We don’t need to assign a win to stalemates just because you don’t think it is fair. And we certainly do not need to force white to play recklessly and give black extra points just because you don’t like draws.

Just be patient. You can be rest assured that sooner or later some player will break out of all the draws and will start winning. It always happens this way.

Topics: Draw, opinion

Donny achieved the title of United States National Master back in the stone age (1988). Now he teaches chess full time for the ICC (Internet Chess Club) under the name “Curmudgeon”. He also writes a monthly article for the Georgia Chess Association, georgiachessnews.com, and is in charge of the ChessThinker website, ChessThinker.com
Discussion and Feedback Join the public discussion or submit your feedback to the editors


Rules for reader comments


Not registered yet? Register

fgkdjlkag fgkdjlkag 6/5/2019 04:30
"I would rather see a hard fought draw any day of the week than some wild crazy game that is full of mistakes just because some organizer or some group of people wants to tweak the game of chess."

So presumably the author is against blitz tiebreaks as well, the point of which is to induce mistakes so as to get a decisive result?

I'm surprised it's not been mentioned, but IMO frequent draws are not a problem in tournaments, but are a problem in matches. The question becomes, what to do after a 6-6 tie like in the last WC match? ALL proposals just admit that the 2 players are equally matched at classical time controls, and so why not introduce mistakes in rapid or blitz games so that the CLASSICAL world championship can be decided. Or why don't we just give an advantage to one player before the match has started (ie draw odds), since the players are actually equally matched in classical time controls.

Saying that a certain % of draws is too much (eg was commented below that 90% is too much) is arbitrary. Why 90% and not 89%? Or 87%? There needs to be some rationale.
Petrarlsen Petrarlsen 1/10/2018 02:55
@ Resistance : I globally quite agree with your last post !
Petrarlsen Petrarlsen 1/10/2018 02:51
@ lajosarpad : I would just point out that, as my two last posts were written in December 2017 and as I haven't changed anything in them, there are several points for which this must be taken into account, as, for example, the Elo ratings, that use the December 2017 list and not the January 2018 list (which didn't exist at the time...).

Rereading these posts, and this sentence in particular : "So I must say I really very much disagree with the very common opinion that a 2600+ GM would be nothing more that a "common target" for everyone else", I think it could be said in a way that, yes, a 2600 GM would be a "common target" for the top-level players in such tournaments, but a quite special target : a target that can shoot back !!
Petrarlsen Petrarlsen 1/10/2018 02:34
@ lajosarpad (3/3) :

Two examples from January 2017 :

1) David Anton Guijarro, at the 2017 Gibraltar tournament :

Anton Guijarro finished 2nd of the tournament ; below him, there were 1 2800+ GM (Caruana) and 10 2700+ GMs (including 3 GMs between 2750 and 2800...) ! And he won 2 games against 2700+ GMs.

2) Baskaran Adhiban at the Tata Steel 2017 tournament :

Adhiban finished 3d of the tournament ; below him, there were 10 2700+ GMs (7 of them being even above 2750 !). And he won 4 games against 2700+ GMs, 2 of these GMs being even above 2750 ! (...one of his "victims" being Karjakin, whose defensive abilities are known by everyone - Adhiban certainly hadn't an "easy task", in this game, but, nonetheless, he won it...)

And I would point out that Anton Guijarro and Adhiban weren't even particularly "on the rise" players ; in January 2017, at the time of these tournaments, they were between 2650 and 2700 ; today, approximately one year later, they are still both between 2650 and 2700.

So a "normal" 2600+ player (not, for example, a young player following a meteoric rise) can be a very serious opponent for a 2700+ player (and even for a 2800+ player, as Caruana - who was 2827 at the time - finished the Gibraltar tournament at the 13th rank with 7 points, while Anton Guijarro finished the tournament at the 2d rank, with 8 points). This don't mean that this would happen very often, obviously, but, quite as obviously, this can happen from time to time.

Another last example from 2017 : In April, Caruana, 2817 at the time, lost one classical game against Zherebukh, 2605 (with an Elo difference of 212 points), and another against Hou Yifan, 2649 (with an Elo difference of 178 points) ; this is another (double) example that shows that a 2800+ GM can even perfectly well lose a game against a GM with a rating between 2600 and 2650...

So I must say I really very much disagree with the very common opinion that a 2600+ GM would be nothing more that a "common target" for everyone else, in a tournament where most of the other participants are above 2750 ; practice has shown that, from time to time, 2600+ GMs can demonstrate that they can quite well compete with top-level players...

And so I think that, yes, their presence tend to motivate top-level players to try significantly harder to win their games, but I think that, even if they aren't more than 2 or 3 2600+ GMs in a given tournament, this would have for effect to lower the draw rates without making necessarily at all these players the "appointed victims" of all the top-level players...

On that respect, it will be interesting to follow the next Tata Steel Masters tournament, featuring precisely 3 2600+ GMs out of 14 participants, 9 of them being above 2750, including two ex-World Champions (Anand and Kramnik), and Carlsen and his (current) 2837 rating...
Petrarlsen Petrarlsen 1/10/2018 02:22
@ lajosarpad (2/3) :

- "However, if there is no rapids, but a set of rules, then at least one of the leaders will have everything to fight for and almost nothing to lose at the end of the tournament."

I agree. Personally, I don't think that this could have a really big effect overall, but it would certainly have a slight impact.

- "If we include lower rated players to a tournament, then we need to make sure approximately half of them is lower rated, since, if there are too few there, then I'm afraid the lower rated will be little more than box sacs."

On this, I don't agree !

I think that, as many other persons, you underestimate (in my opinion...) the chances of a 2600+ player against top-level (above 2750) players.

In fact, the level difference between, for example, a 2680 player and 2780 player is the same as the difference between, for example, a 2050 player and a 2150 player.

But I rather think that many players would think that the difference between a 2050 player and a 2150 player isn't THAT enormous. This because a player around 2100 is, in a way, completely "in the middle" ; between a rank amateur and players with a IM or GM level - 1450, 1550, 1650, 1750, 1850, 1950, 2050, 2150, 2250, 2350, 2450... 100 points by 100 points, there are many, many steps, and, for example, the step between 2050 and 2150 is "one step amongst others", and doesn't seems to be particularly meaningful. Whereas, as, for 2600+, 2700+, and 2800+ GMs, these steps are the very last, each step seems to be very meaningful, and, in fact, more meaningful that it really is. Yes, a 2780 player is better than a 2680 player, but not so much as many persons would think, in my opinion.
Petrarlsen Petrarlsen 1/10/2018 02:15
@ lajosarpard (1/3) : VERY late post... but "not my fault !!"... because of a technical problem on ChessBase, I couldn't post anything for several weeks on ChessBase, and so, the two posts I had prepared the 23rd of December 2017 stayed in a (virtual...) cupboard until today. Anyway, there they are !...
Resistance Resistance 12/24/2017 10:35
Good conversation; many good posts, and replies. Many good ideas. First of all, I'd like to say that I agree with the overall sentiment of the author of this article regarding draws: there is nothing wrong with them. And I, on my part, thank God they exist. For they provide even greater mystery (-- greater depth --) to our beautiful game: they show you that no matter how much you pursue victory, or that no matter how big yours or your opponent's Elo (or ego!) is, there will still be (many) occasions in which you'll end up realizing that, contrary to that which you originally intended or thought, it was ultimately NOT POSSIBLE TO WIN (or lose) from this or that position; you realize there was a higher logic at work behind many of the situations that took place over the board, which you previously thought you had under control (-- This is one of the most amazing things about chess, which also happens to grant the possibility of unending series of creative achievements: that it never seems to get exhausted --). Eventually, you learn that chess is bigger than you (-- just as with life --).

However, that's no reason enough to avoid fighting nor to stop enjoying the game, which I think is the fundamental issue at stake here, though I don't think that forcing players into adopting certain attitudes towards the way they approach tournament play is the right thing to do here, either.

My suggestion for an improvement is more on the line already suggested by other commentators here on this forum: if you want fighting, lively, interesting chess in these big tournaments, then bring in people that fights and that enjoys playing the game. If that means getting some lower rated players, then that's the way it's gotta be. Organizers, especially those who also provide live broadcasting of the games, and which worry so much about audiences, should be more careful then when choosing their playing fields (-- do not blindly invite people solely on the grounds of their high position in the Elo rankings: see the way they play, think of best matches among the available players, etc, etc --).

Establish some sort of prize for the best game of every tournament and/or best game of every round (a la Wijk aan Zee); have some other prizes, too. For example, 'most beautiful game', or 'toughest game', or 'hardest-fought game'; or 'best theoretical novelty', 'best ending', and so on, and bring some quality judges (well-known and/or highly regarded players) in order to do the selecction of these games (-- Do not allow some random group of people from the internet to choose The Best Game of the Tournament --). Perhaps you can have people voting for best game each round (from the internet), but for Best Game of the Tournament have some known faces in that panel of judges (-- for example, for the 2017 London Chess Classic, you could've had some of the locals: Speelman, Short, McShane; King, Ward, Malcolm Pein, etc. --), much like they do in the votings of Chess Informant: every judge's selections are published at the end of the tournament; everyone of them picks its 10 favorite games, then they rank them from 1 to 10, the best one getting 10 points, the second 9, etc, etc. The game with the highest score among the judges wins the prize. After the tournament has finished, also, and in an effort to make of every tournament something to remember, publish a book (a quality book (!)), presenting the best or more important games, annotated by the players themselves, or at least by some of them, etc, etc.

There's a lot of stuff you can do to keep players and audiences motivated...
lajosarpad lajosarpad 12/22/2017 10:16
The fighting spirit could be raised by the regulation of the tie break. If they play rapids and the leaders are comfortable with rapids, then they will not leave their comfort zone. However, if there is no rapids, but a set of rules, then at least one of the leaders will have everything to fight for and almost nothing to lose at the end of the tournament.


If we include lower rated players to a tournament, then we need to make sure approximately half of them is lower rated, since, if there are too few there, then I'm afraid the lower rated will be little more than box sacs.

@Angelo Pardi

I agree with you.
Petrarlsen Petrarlsen 12/21/2017 11:43
@ Angelo Pardi :

"However I think a general way to get more interesting games is to go for (rather more) mixed fields. Like have only one or two tournaments with most of the top 10, and most tournaments being either opens or closed tournaments with a mix of top 10, 2700 and high 2600 players. Then you would see more hard fought games, because top players would need to get a high score against the weaker players, and thus would need to take risks."

I am rather under the impression that, like you, many people are in favor of including more slightly lower-rated players (I would say, typically, 2600+ GMs) in top-level tournaments.

Personally, I quite agree. This obviously doesn't change anything to the games ; 2600+ GMs are very strong players in their own rights, and I've always find very interesting to follow the games between top-level (2750+ or 2800+ GMs) players and 2600+ GMs ; the top-players' goal is necessarily to win, and they play differently ; it shows another facet of their talent. And, sometimes, as a 2600+ GM is clearly VERY far from being a patzer, the 2600+ GM wins the game (frequently because the top-level player overpushed, or took too much risks), and these games are also quite interesting in my opinion.

And, necessarily, this tends to lower the draw rates, so everyone can be pleased, including those who would want lower draw rates...

As an aside, I rather THINK (I am not sure at all...) that it isn't necessary to add many 2600+ GMs to lower the draw rates ; I am under the impression that the inclusion of even 2 or 3 2600+ GMs has a significant impact.

In my opinion, this could be due to the fact that, when all (or nearly all) a given tournament's participants are 2750+ GMs, the players tend (in my opinion) to play in a rather defense-oriented manner, as they know that all their opponents are particularly strong, whereas the inclusion of 2600+ GMs tend to generate a globally more offensive approach to the game by the players.

In that respect, it will be interesting to follow the next Tata Steel tournament : in the A group, 3 players out of 14 are 2600+ GMs, and it will be interesting to see what the draw rate will be !
Petrarlsen Petrarlsen 12/21/2017 11:15
@ genem :

I both agree and disagree !

- "To say simply "Draws are not a problem" is to evade the legitimate question, and to imply that a 99.8% draw rate would be perfectly fine."

I quite agree. In fact, I already said so indirectly on this page ; in a post to ICCF Grandmaster, I exposed my opinion : that the 90 % draw rate of correspondence chess is too high, and that I don't see any problem with the current draw rates in on the board chess (giving as an example the London Chess Classic 2017's 77.8 % draw rate, which represents a rather high draw rate compared to other high-level tournaments ; in my opinion, it is quite acceptable to have a more or less 80 % draw rate in the tournaments having the higher draw rates).

- "At a majority 60% draw rate, and often higher in World Champ matches, to much scoreboard action is being drained by draws."

With this, I must say I very much disagree.

The rather high draw rates in chess are precisely one of the points I find particularly interesting in chess. For me, the fact that draws occur frequently precisely mean that wins have a much stronger meaning : to win a game, you must escape the "central draw zone" ; in each game, there is quite a real "drawing maze", and to win a game means that the player managed to escape this "drawing maze", and this, in my opinion, give much more value to wins, in chess. If, for example, there were only wins, and no draws, a win wouldn't mean much ; only, really, that the winner played at least more or less at the same level that his opponent ; a coin must fall on one of its side, and the win must go to one of the players ; it wouldn't mean more than that. Whereas, in chess as it is played presently, when you win, this certainly doesn't only mean that you played at more or less the same level as your opponent : you have to EARN a win ; really ; when you see a win, you know that, necessarily, the winner really battled for it. And, in my opinion, this makes all the difference.

So, my personal opinion on this is that, for a win to be really significant, the draw rate must be clearly above 50 % for top-level chess ; to give a number, I would thus say at least a 55 % draw rate.

- "Sofia rule, and extending the life of each draw offer, could help a little."

I wouldn't be at all against Sofia rules, even without any move limit, as it doesn't in fact change anything in the game. Obviously, it doesn't do anything against none-fighting games, but, personally, this doesn't bother me too much ; at least, if the game is rather long (and to implement Sofia rules necessarily tend to generate longer games), when the players are 2700+ or even 2800+ GMs, it is a very well-played game, and I find rather interesting such games.

- But, as for changing the rules, for castling, or using different starting positions, it really changes the game, and, as I consider that chess as it is currently represents a very refined equilibrium, and as I also consider that, for the moment, the draw rates are quite acceptable, I wouldn't find at all such changes to be something positive.
adbennet adbennet 12/21/2017 06:51
Excellent post by Angelo Pardi.
Angelo Pardi Angelo Pardi 12/21/2017 11:30
In the draw rate debate, I think there are two very different phenomenon to consider : you have to ask what is the cause of those draws.
There are two main causes : aversion to risk-taking, and high defensive skills. The first one leads to boring games and tournaments, the second one to good games (which are not exactly the same as interesting games, because, let's be honest, we all get more excited over a sac & mate East-Indian than a 80-moves hard fought Berlin).

The first problem is a really big problem, and I'm not sure it can actually be solved at all. Chess players hate losing, I think generally speaking we tend to hate losing ever more than we love winning. But for professional players, there is also lots of money at stakes, so they tend to favor tournament strategy that will decrease variance as much as possible : hence playing solidly rather than aggressively.

However I think a general way to get more interesting games is to go for (rather more) mixed fields. Like have only one or two tournaments with most of the top 10, and most tournaments being either opens or closed tournaments with a mix of top 10, 2700 and high 2600 players. Then you would see more hard fought games, because top players would need to get a high score against the weaker players, and thus would need to take risks.
And by contrast it would also make the encounters between top 10 players more interesting, like in Wijk aan Zee the Carlsen-Caruana or Carlsen-Aronian games are the climax of the tournament, while in Sinquefield or London it is just an other game.
lajosarpad lajosarpad 12/21/2017 11:00

you are cherry picking some tournaments which you think underlines your point, yet, you call people who disagree with you "blind". I am not saying that you are wrong with your point. I am saying that your methodology is wrong. Examples you think supporting your argument are not empirical evidence. They are making your argument stronger, since they prove it is possible, but not more than that. And of course this applies only if the examples are meaningful. The difference of playing strength in those tournaments, as pointed out by Petrarlsen is quite a possible explanation of the result.

Is it likely to fix the results to win one and lose one under the 3-1-0 scoring system? It is highly unlikely in an arbitrarily picked example, but it is very likely that it will happen once in a while. Do not mix the chance of an event occurring at a given example with the chance of that event happening at all.
genem genem 12/21/2017 10:51
A more useful question is: HOW HIGH can the draw rate go before it becomes a problem?
To say simply "Draws are not a problem" is to evade the legitimate question, and to imply that a 99.8% draw rate would be perfectly fine.

In a healthy sport, action should occur in the game, but also on the scoreboard. At a majority 60% draw rate, and often higher in World Champ matches, to much scoreboard action is being drained by draws.

The solution? Sofia rule, and extending the life of each draw offer, could help a little. To make a significant change in the draw rate, the castling rules could be modified to make opposite wing castling more attractive.

Author Donny Gray alluded to chess960-FRC or Fischer Random Chess, negatively. I agree. But the better option would be to "Discard the *Random* from Fischer Random Chess!", and to let one non-traditional start setup be added and reused in some games in some tournaments. The draw rate might be reduced by a 2-3 year deluge of clever opening novelties able to give one player a potentially decisive advantage (all games are draw from the start until one player achieves a big enough advantage).
Petrarlsen Petrarlsen 12/21/2017 12:37
@ koko48 (2/2) :

To answer rapidly (...if anyone is interested in more details, I gave the links to our previous conversations on this subject previously on this page...) to some of your arguments :

- "The 2009-2012 London Chess Classics had drastically improved win rates. The 2015 and 2016 Bilbao tournaments did not, for several reasons."

And so, it isn't possible to conclude from completely contradicting data...

- You attempt to discard the 2015 Bilbao Masters because it is a four player double round robin ; I don't find the argument very convincing at all ; you don't even try to prove that such tournaments played under the "1 - 1/2 scoring system" also produce high draw rates...

A much more significant difference between the London Chess Classics 2010 - 2012 and the Chess Masters is that, in the London Chess Classics, there was a wide range of players (several players were even under 2650), while in the Chess Masters 2015 - 2016, nearly all players were above 2750 (with only one exception, Wei Yi in 2016 - and all the players were above 2750 in 2015). In my opinion, this is the obvious explanation, for the extremely different draw rates, between the London Chess Classics and the Chess Masters : when there is a wide range of levels, the draw rate always tend to diminish, while it is quite the opposite when there is nearly only top-level (+ 2750) players : draws tend to abound... So in fact, in my opinion, it is very probable that the main reason for the low draw rates in these editions of the London Chess Classics is quite simply the wide range of levels of the players, and not the use of the 3 - 1 scoring system...

- "I also noticed that not ONCE did anybody mutually throw games in this double round robin to "split the point" - some people's (highly unlikely) concerns about football scoring."

I never stated anything about this, so this isn't a problem for me...

- And, for the rest, your main argument is always, one way or another, to argue that the games are more interesting with the 3 - 1 scoring system. Quite obviously, as the "interest" of the games cannot be converted into numbers, this is indeed the "easy escape" for you ! But also, as this is a completely subjective argument and quite impossible to prove rationally, it is quite obvious that it isn't really an argument ; only a way to say that, whatever everyone will ever say, you are right, and every other possible person wrong...

And I suppose that now - unfortunately, as I know you rather well now, this is all too easy to predict... -, as you will not know anymore what to answer, you will throw at me a full bucketful of insults... you do it absolutely each time, so it doesn't seems very probable that you will do otherwise... one more time, this is quite a pity, really...
Petrarlsen Petrarlsen 12/21/2017 12:31
@ koko48 (1/2) :

Quite obviously, even if I would - to say this metaphorically - put a white dog just in front of you, you would insist to affirm that it is a black cat... quite incredible (but, unfortunately, quite true, also)...

Your first two paragraph are quite a topical example ; you say in them :

"And how can you once again focus on 'draw rate'? When I explained to you several times that is not the underlying issue (This is why I wonder if your obtuseness is deliberate).

For possibly the third or fourth time: The underlying issue is the rate of GM Draws, or NON-GAMES. Not the 'draw rate' as a whole."

But, 12 days ago (not 12 years : twelve days !) you said, in a post to lajosarpad (https://en.chessbase.com/post/london-chess-classic-2017-live) in which he affirmed that you were "cherry picking" :

"How is it 'cherry picking' when I gave you four consecutive London Chess Classics - 2009 through 2012 - that used Bilbao Scoring and had drastically higher win rates than this one? And drastically higher win rates than most elite tournaments in general?"

Quite typical ; when it goes in your direction (or, rather, when you THINK it goes in your direction...), you use abundantly the "draw rate argument" (a full paragraph, in this case), and when it doesn't, you even insult other commentators ("I wonder if your obtuseness is deliberate", in your last post to me) to affirm that the "draw rate argument" you used yourself 12 days ago is completely stupid !!!

One thing is clear, taking all this into account : you are not seeking the truth, but only to defend, even with the most absurd and contradicting arguments, your thesis ; this is quite a pity indeed...
koko48 koko48 12/20/2017 10:45
@Petrarlsen And how can you once again focus on 'draw rate'? When I explained to you several times that is not the underlying issue (This is why I wonder if your obtuseness is deliberate).

For possibly the third or fourth time: The underlying issue is the rate of GM Draws, or NON-GAMES. Not the 'draw rate' as a whole.

The 2009-2012 London Chess Classics had drastically improved win rates. The 2015 and 2016 Bilbao tournaments did not, for several reasons. However, the rate of GM draws were still virtually ELIMINATED in the Bilbao tournaments as well

The 2015 Bilbao tournament was a four player double round robin incidentally. A much different dynamic. With only two games played in a given round, it's easier to take into account what's happening on the other board. If you see the other game is heading to a draw, you can also keep pace with a draw.

I also noticed that not ONCE did anybody mutually throw games in this double round robin to "split the point" - some people's (highly unlikely) concerns about football scoring.

The 2016 Bilbao Masters also had virtually no short unplayed draws, with the notable exception of Sergey Karjakin, who played an 18 move, a 19 move, a 23 move, and a 26 move draw. This tournament was also three months before Karjakin's WC match with Carlsen, so I believe Karjakin's main concern was avoiding showing prep, not winning the tournament

The only short boring draw I remember in 2016 Bilbao, not played by Karjakin, was Wei Yi- So. But the point is these types of games are an extremely rare exception. They are not regular occurrences, like they are in current elite tournaments with traditional scoring
adsqueiroz@hotmail.com adsqueiroz@hotmail.com 12/20/2017 10:24
Years ago Bobby Fischer said that many of the games between grandmasters were pre-arranged to draw each other quickly. That´s disgusting because takes aways our passion for this game.
Petrarlsen Petrarlsen 12/20/2017 09:35
@ koko48 : On the "Live" page about the London Chess Classic (https://en.chessbase.com/post/london-chess-classic-2017-live), lajosarpad and I proved extensively that you were "cherrypicking", i. e. that your were chosing only the elements favorable to your thesis, while systematically discarding disfavorable elements.

At the time, I sincerely thought (and said so) that you didn't do it on purpose, but, now, it is unfortunately completely obvious that, now, you are deliberately suppressing the facts that doesn't go in the "right direction"...

In fact, you say : "We have several examples of the 3-1-0 system in practice, and the results were strikingly positive in this regard.", citing exclusively four old (2009 to 2012) examples from the London Chess Classic. How can you, ONE MORE TIME, suppress completely the fact that the Chess Masters Final 2016 had a 76.7 % draw rate ? And that the Chess Masters Final 2015 had even a 83.3 % draw rate ? You know it ! I explained it to you (with links to these tournaments pages) only 12 days ago !!

Are really 76.7 and 83.3 % draw rates such "strikingly positive results" ???

You spend posts after posts to explain that the London Chess Classic 2017 had an inacceptable draw rate, your beloved "3 - 1 scoring system" give comparable and even higher (for the Chess Masters Final 2015) draw rates, and you don't find any problem with this ??? At this point, I really think that, even if a tournament featuring the 3 - 1 scoring system gave a 95 % draw rate, it wouldn't change your opinion on this system ! This is really quite astounding, I must say...
koko48 koko48 12/20/2017 09:13
Except for one thing people. We have empirical evidence the 3-1-0 scoring system works. I invite each of you to look at the 2009, 2010, 2011 and 2012 London Chess Classics. Each of those tournaments had much higher win rates (one of them over 50% decisive games) and there were virtually no GM draws in any of those tournaments.

No, the quality of the games did not go down. They went UP. The lowest quality game is an unplayed game, or a game that never deviates from the equal path.

This is the blind traditionalism and conservatism among many chess players, I mentioned earlier. "Everything is fine the way it is".....Or worrying about about some unlikely hypothetical problems of football scoring (poor quality games, mutually throwing games in a round robin), while they ignore the real life evidence that's right in front of them. We have several examples of the 3-1-0 system in practice, and the results were strikingly positive in this regard.

I provided the links before, I don't feel like doing it again. You can google those tournaments on chessgames.com.

You don't need to play through every game. Count the number of total games, and count the number of decisive games.

Then pick the shortest draws you can find in terms of number of moves, and play through them. See if you can find one non-game in that tournament. Because I played through almost all of those games, and I was unable to find ONE. In four consecutive London Chess Classics
Petrarlsen Petrarlsen 12/20/2017 08:09
@ JayPBee : Quite interesting post !... and, globally, I quite agree with you...

One little point :

- "I have seen it quoted below that top tournaments have 90% draws. I do not think this is correct (...)" : This draw rate is correct for correspondance chess, but NOT for on the board chess (even a tournament much criticized for its draw rate as the London Chess Classic, for example, has only a 77.8 % draw rate).
A Alekhine A Alekhine 12/20/2017 06:47
Several of these draws could have been won if not for mistakes made by players who had the advantage.

Counter-balancing errors are examples of "the equalizing injustice of chess."

But the point is, these were games that could have been won. Which suggests that the real problem was not the nature of chess, or the nature of the scoring system, but simply human fallibility...as usual.
adbennet adbennet 12/20/2017 06:23
What the mass-market spectator wants to see in sport is not the win per se, but "action" (goal, point, run, etc.). The big problem for chess as a spectator event is that most of the action takes place only in the players' heads. It takes a knowledgeable spectator to begin to understand why a typical move is made. The less knowledgeable spectator can understand the combination, but not much else. What the mythical mass-market spectator would want to see in chess is the combination.

Changing the point-scoring system (e.g. 3-1-0) will not necessarily produce more combinations between evenly matched opponents, but it might produce more oversights due to one player rolling the dice. If that sort of thing is appealing to spectators, chess is better off without that kind of spectator.

Changing the points awarded for draws might have the exact opposite effect from the one desired. Instead of combinations, we will get to see tedious endgames played out where one player is aiming to stalemate the opponent, and the other player is aiming for bare kings. Nobody wants that.

No amount of "tweaking" can change the fact that most of chess takes place in the players' heads. Reti proposed a different solution long ago. To encourage more combinative play, organizers need to invite weaker players, so that there are bigger mismatches. Of course it was not a serious suggestion, he was just illustrating that the problem is built-in and has no easy solution.
jjmolina jjmolina 12/20/2017 05:44
"What's the matter with draws?" Correspondence chess is the matter, or more precisely, computer chess. Try playing a tournament were computers are playing (directly or indirectly, trough assistance) and you'll see soon enough.
JayPBee JayPBee 12/20/2017 05:38
Ok ... for my two penneth worth. I do not believe that draws per se are a problem in chess - although I will agree with one commentator below that our friends across the Atlantic in the USA tend to have a much greater dislike for a drawn result in any game / sport. Where there is a problem is when there is an 'un-fought' draw. This is not just about a draw offer after 10 moves but includes games where the players play a line known to end in a repetition. This latter also gets around the Sophia rules. For instance Nepo v MVL in the last round at the Classic. That said - I completely understand both players' motivations for what they did and both are really good guys. In fact even a draw after 10 moves might make perfect sense from the point of view of the individuals concerned. But as a spectator or sponsor this would not be ideal. I would personally make tournaments such as the Classic forbid draw offers before move 40 and not 30 as was recently the case as move 30 doesn't even get to the first time control. But that does not completely remove the problem of a repetition bashed out in 5 minutes.

I wanted to say that I don't think the scoring system of 3-1-0 will make a difference. The 'problem' here is that ratings are far more important for invitations for top players than the number of wins you achieve. Yet we all like our ratings - us amateurs as well are just as guilty here. I would also say that I have yet to see any evidence that the change in soccer scoring has actually led to more decisive games. Google it for yourself and see if you can find anything at all to support this theory - I would be interested. As an aside I do believe the standard of football is much higher now. [Sorry of course I meant 'soccer' - must remember the international audience here who believe that a game where the ball is kicked most of the time should not be called FOOTball because of a game where the ball is thrown or held by HAND for most of the time. Sorry couldn't resist! ;) ] Anyway the increases in the standard I think have much more to do with the money in the game and the scientific advancements rather than about any scoring system. Certainly even with the change in scoring I seem to recall having sat through a number of World or European Cup tournaments where every so often you get one or 2 groups where the first two games have all ended 0-0 or 1-1 and you get the soccer equivalent of this thread.

I have seen it quoted below that top tournaments have 90% draws. I do not think this is correct - although I fully accept that it sometimes can seem like that! In reality I think 20 - 30% decisive results are the norm for 'top' events. But this is hardly surprising. The guys at the top are all so good that it is difficult to beat them even when they are worse. So as they are all so closely matched draws are very likely. If a decisive result is what is craved above all else then rapid games or more mixed opposition will achieve this no need to tinker with the stalemate rule - which for me would change the game from the one in which we play. So personally I like the classical time control and am more than happy to have a well fought game end in a draw.

I would like to end with one point that some commentators seem to forget. That is that these guys at the top are all immensely competitive. Even in the pro-biz events the professionals want to win - particularly if they have a partner that can play a bit. So when they play each other the desire to crush the opponents ego is quite high ... it is just tempered by their desire not to have their own ego crushed! Sometimes it seems to me that some commentators put too much emphasis on the result rather than the effort of the game. So the first 19 games were drawn at the London Classic .... but were they hard fought draws? Personally I wouldn't even have mentioned it other than as an interesting anomaly. But these are just my views of course.
Petrarlsen Petrarlsen 12/20/2017 04:15
@ ICCF Grandmaster :

In my opinion, the situation isn't really the same for correspondence chess and for on the board play.

I agree that, in correspondence chess, the 90 % draw rate is too high, and that it would be important to find a solution.

But, in on the board play, even in tournaments with a high draw rate as the last London Chess Classic, for example, there are more than two times more wins than in correspondence chess (the last London Chess Classic's draw rate was 77.8 %, for example).

So, for on the board play, I agree with the author of the present article : I don't think that, for the moment, there is a "draw problem"...
koko48 koko48 12/20/2017 04:00
@tourthefarce Thank you.

C'mon people you're chessplayers, use your heads.This isn't that difficult to understand.

There's absolutely nothing wrong with fighting draws. A good, fighting game that's correctly (or mostly correctly) played on both sides, will end in a draw anyway.

The problem is with unplayed GM draws, which is nothing less than collusion. The problem is also with draws that may not be as short, they may even satisfy the Sofia Rule requirement of 30 or 40 moves played - but neither side is doing much to break the equilibrium. Another, less blatant iteration of the GM draw. And of collusion

"Chess is fine just the way it is"....Not according to a lot of people. Not according to many serious players and devotees of of the game, not according to the casual/potential chess follower that gets turned off by these cowardly little snooze fests (no other sport has so many unplayed, non-games as chess, incidentally). And chess is certainly not fine among potential sponsors - otherwise chess wouldn't have so few of them
tourthefarce tourthefarce 12/20/2017 03:23
As somebody already commented, the problem are not draws, but listless games without a true fight that end in a draw - and we had plenty examples of that in London - that is what irritates chess fans the most. We tune in to see the greats play, not just push pieces around and agree to a draw, which is what many of the games in London looked like. This is where we need to reward those that take risks and sometimes loose because of it, but sometimes score. So, if you play 10 games, lose 5, win 5 - you should have more points in the end then someone who has 10 draws. Sure, it is possible that the 10-draws guy tried to win every single one of his games and pushed really hard, but come on - how often does that happen (ever)?
A7fecd1676b88 A7fecd1676b88 12/20/2017 03:03
"Still the players that are more talented, faster in calculations, and is better prepared will still win." -- Donny Gray.

Mr. Gray, another victim of the talent myth.
ICCF Grandmaster ICCF Grandmaster 12/20/2017 02:32
In the ancient chess, when playing for money, you lost half of the stake, if getting stalemated. Stalemating the opponent meant you played better than him, even though you couldn't mate him. This is what also happens in modern chess. You may have played better, almost achieving a decisive advantage, but nevertheless many endgames end up in a draw due to stalemate, if you couln't convert a positional advantage into a decisive material advantage in order to force a mate.
Lasker, Réti and others recalled the ancient rule as "wise". Tartakower, Nimzowitsch, Capablanca, Vidmar a.o. thought similarly.
Awarding stalemate with a score of 3/4 point to 1/4 point wouldn't change too much, but makes chess even more complex and dynamic. A "sport" with an extreme high draw rate is just unsatisfctory for all, esp. for the media.
Réti wrote: "It is the same as if with a race it has been agreed that a small difference of time, say a second, should not decide the race and that it should count as a dead heat. Such a result according to Capablanca would mean that the best runners could not beat each other." (Réti, Modern Ideas in Chess, p. 176)
By the way, we had this discussion already two years ago with reference to the situation in correspondence chess, where the draw rate nowadays tends to 90% in top tournaments. Lots of interesting and exciting draws... no question, but doesn't that sound rather academic?
Changing the score for stalemate could help a bit. And don't think, it's generally much easier to achieve an advantage that allows you to stalemate the opponent. It's still a tough thing when playing on a high level.
See also: https://en.chessbase.com/post/correspondence-chess-the-draw-problem
AgainAgain AgainAgain 12/20/2017 02:07
All the funny scoring systems (i.e. 3-1-0) are perfect for match-fixing...
Petrarlsen Petrarlsen 12/20/2017 01:23
@ melante :

In fact, I think that the big difference between chess and football (and other similar sports) is that, in a way, in chess, a draw is a normal result between two top-players of exactly the same level (for example, I think it is rather obvious that if Caruana plays Mamedyarov this month - both have a 2799 rating, presently -, no-one can be surprised if the result is a draw). In chess, a top-level game without any clear error, even really hard-fought, will most of the time result in a draw.

Whether in football, I don't think that it is possible to say that a 0 - 0 score is a "normal result" ; one more time, I am not very knowledgeable in football, but I rather think that to score a goal is something more or less "normal", in football ; that the "defense level" isn't high enough to consider that a "normal match" is a "0 - 0 match". And I don't see any reason to consider that the number of goals for each team would necessarily tend to be the same (with a 1 - 1, 2 - 2, etc. tie for a result), if there is one or more goals in the match : if the two teams have scored one or more goal, logically, both of them can hope to score still one or more goals. So the situation doesn't seems to me at all to be the same as in chess, in that respect...
Alan Smith Alan Smith 12/20/2017 01:13
Thank You! I totally agree. All the whining about draws is so tedious.
lajosarpad lajosarpad 12/20/2017 12:14
I agree with the article. I am quite interested in tournaments and I do not lose my interest in chess just because there is a high percentage of draws. I do not understand those who hate draws so much. If they are interested in chess and already watching tournaments, then it proves that chess is interesting for them despite the drawish tendency, which is considered by them to be too high. And I don't see many people coming to chess if the draw ratio would be reduced. Yes, tournaments need to be interesting, so if we do not see only super high level players, but some of them are slightly weaker, that ensures fighting chess in every round. Higher rated players want to score against lower rated and those higher rated players who failed to win against lower rated opponents will need to get out from their comfort zones to remedy the situation.
Petrarlsen Petrarlsen 12/20/2017 11:43
@ melante : I quite agree that there isn't sufficient data to prove for good that the 3 - 1 scoring system doesn't change significantly the tournaments' draw rates (but it isn't either possible for the moment, in my opinion, to prove the usefulness of this system, as I explained previously on other ChessBase pages - cf. the links I gave in one of my previous posts on this page).

But, at first view, I don't think either that your comparison with football is quite adequate ; I am not very knowledgeable in football, but I rather think that to try to score a goal when you are tied, in football, is much less of a commitment than, in chess, to choose quite a double-edged continuation, rather than another continuation, very safe but completely drawish.

What seems to me to be very specific to chess is that, in many positions, if you choose to try really hard to win, you will also risk quite strongly to lose, and I don't think that, in football or in other comparable sports, there is such a strong tendency in this direction. So I don't think that it is simply possible to take the results this system gave in football, and to consider that, for chess, the result will necessarily be the same...
vipiu vipiu 12/20/2017 11:17
Try a superGM tournament like this:

0 points for being checkmated
1 point if you get stalemated or if you are the player forcing the repetition
2 points if it is a position where no player can make progress even both try to avoid repetition
3 points for stalemating the opponent or being forced to accept the repetition
4 points for a checkmate

I am not claiming this is for sure best but I believe it has its PROs and it would be very interesting to give it a try (a chance) in SuperGM tournament.

Some PROs:
- today's theory is still valid, but will need to increase to be taken to stalemate level.
-stalemate is not a loss but still not equal with a draw
-today games with end fast as draws will have a lot of fight in it because the margin between stalemate and draw would give plenty of play and resources

So don't change the rules, just try it once in a tournament :)
melante melante 12/20/2017 11:05
@Petralsen : I understand your reasoning, but I don't agree :) There are many league based sports that show otherwise. A player may indeed want to block his direct opponent on a draw, and that will definitely happen as does happen between teams in football, but he/she/they will know that, in so doing, they may also harm their tournament advancement chances if other players win. Play too much on the defensive won't pay in the long run.

When, decades ago, the 3-1 point system was adopted in my country for football leagues instead of the previous 2-1, I was really against it (and indeed my team was the first to be relegated because of that rule. If the 2-1 rue were still in place, we would have done well that year!) but in the long run I have to agree it was a good choice and the overall quality of matches improved as, in general, it does indeed motivate people to push for a win.

Any other reasoning without extensive data is just speculation: the proof is in the pudding and as long as we don't have, say, at least 5 years where all tournaments are played with the 3-1 scoring system, we can't really compare results and say it won't work.
mtm57 mtm57 12/20/2017 10:36
I totally agree! Draws are frequently great fight!
Petrarlsen Petrarlsen 12/20/2017 09:53
@ melante (2/2) : And, later, I added this, in another post to Raymond Labelle :

"I thought a little more about your last idea about the 3 - 1 scoring system.

I think there is a possible further consequence : with the 3 - 1 scoring system, I think that the motivation for "not-losing" a game (and thus, for accepting more easily a draw, rather than to take risks to try to win the game) is at its utmost when the two opponents, in a given game, are approximately of the same level.

Why ? Because, precisely, the main motivation for "not-losing" isn't, as in the 1 - 1/2 scoring system, to gain a half-point, but to prevent one's opponent to win the 3 points that winning the game would give him.

So, when two participants in a tournament are widely separated in level, to lose isn't an important problem, because you will not compete with your opponent of the day for the global standings.

But when, for example, you have Carlsen playing Aronian, even at the beginning of a tournament, they both know that they will very probably compete for the first place in the global standings. So if one of them loses the game, yes, he directly loses only 1 point, but this isn't the end of the story : he also gives 3 points to his opponent, and it can have very negative consequences for him, as for the global standings.

And the problem is : 1) that the 3 - 1 scoring system has been thought up as a means to avoid very high draw levels, and 2) that the highest draw levels occur when all the participants, in a given tournament, are (approximately) 2750+ GMs (because : a) the highest the global level is, the highest the draw rate will be, and b) the narrowest the global level range is, the highest the draw rate will be, so when these two elements are combined in a tournament, the tendency is strongly oriented in the direction of high draw rates).

So the consequence seems to me to be, rather probably, that the 3 - 1 scoring system is at its least effective, for lowering the draw rates, for precisely the tournaments featuring the highest draw rates, i.e. the tournaments for which the participants are exclusively (or nearly exclusively) 2750+ GMs.

And this could explain the quite high draw rates encountered in some tournaments featuring the 3 - 1 scoring system (as the Chess Masters Final 2015 and 2016 - the draw rates were 83.3 % in 2015 and 76.7 % in 2016) : these tournaments' participants were precisely nearly all 2750+ GMs... and the draw rates, even with the 3 - 1 scoring system, were, as we have seen, quite high..."

In view of all this, I really think it isn't obvious at all that the 3 - 1 scoring system would necessarily diminish the draw rate, in high-level chess tournaments.
Petrarlsen Petrarlsen 12/20/2017 09:52
@ melante (1/2) :

"I believe knowing that more aggressive play would be rewarded will bring in more wins in chess in the long run (...)"

In fact, following an idea by Raymond Labelle on this page : https://en.chessbase.com/post/london-chess-classic-2017-carlsen-wins-grand-chess-tour, I think this isn't necessarily obvious at all in chess.

Raymond Labelle explained this :

"In a 3-1 system, you also have to be as prudent as in a 1/2 system. In a drawish position, each player has a good reason not to want the other one go away with 3 points - which compensates the possible gain of 3 points in pondering whether you offer or accept a draw. The net incentive to draw is the same, even if the terms of the equation are more intense in one case than in another. 1-1 = 3-3."

I answered this :

"It can indeed be a strong motivation, not to let your opponent ran away with a win's 3 points.

With the usual 1 - 1/2 scoring system the motivation to favor a draw rather than to take risks for the possibility of a win is that a half-point represents something quite significant, and the players don't want to risk to lose this half-point, whether with the 3 - 1 scoring system, the same motivation in favor of a draw can quite well exist, but, in this case, because the players wouldn't want their opponents to gain the 3 points that a win would give them - quite a big advantage for the opponent. Quite interesting indeed !..."