Proposing a new system for top tournaments

by Miguel Illescas Córdoba
6/22/2019 – Entrepreneur, trainer, and Grandmaster Miguel Illescas puts down some thoughts on how to make elite chess touranments more fan friendly. Inspired by the Norway Chess Armageddon experiment, he offers a proposal to improve the format with tweaks to the scoring and time control. We encourage you to offer feedback in the comments below.

YOUR PERSONAL CHESS COACH - Whether you’re taking your first steps into the world of club chess, or already playing at a tournament level: with FRITZ, you can train more efficiently, intelligently and with a more personalised approach than ever before.
FRITZ is more than just a chess engine – it’s a training revolution! Whether you’re taking your first steps into the world of club chess, or already playing at a tournament level: with FRITZ, you can train more efficiently, intelligently and with a more personalised approach than ever before.

An improved Armageddon?

The Norway Chess tournament is finally over, so it is a good time to analyze the conclusions that we can extract from the “Armageddon” experiment. First, we must praise the organizers for achieving this wonderful show, and commensurate audience records. 

Draws: A real problem

The first aspect that grabs our attention is that the percentage of draws in the Norway Chess has been increasing year after year, apart from 2015, and in this edition, draws have beaten all records. Three out of four games ended up in a draw.

Norway Chess

Year 1-0 Draws 0-1 Games Draw %
2019 7 34 4 45 76%
2018 8 28 3 39* 72%
2017 10 31 4 45 69%
2016 12 30 3 45 67%
2015 15 25 5 45 56%
2014 10 28 7 45 62%
2013 16 21 8 45 47%

* Ding Liren retired after three rounds 

A logical explanation for this fact is that the players that take part in the Norway Chess are world-class chess players and, as a consequence, they make very few mistakes. In addition, these players are facing constantly throughout the year, and they know each other very well. 

When the field is mixed, the percentage of draws is lower, as it occurs in the Tata Steel tournament, for example:

Tata Steel Masters (Group A) 

Year 1-0 Draws 0-1 Games Draw %
2019 20 54 17 91 59%
2018 25 55 11 91 60%
2017 23 57 11 91 63%
2016 22 61 8 91 67%
2015 28 45 18 91 49%
2014 23 30 13 66 45%
2013 28 50 13 91 55%

And if we analyze tournaments with younger players or lower Elo ratings, this percentage of draws goes down, as we can see in Tata Steel Chess Challengers (Group B): 

Tata Steel Chess Challengers (Group B)

Year 1-0 Draws 0-1 Games Draw %
2019 24 49 18 91 54%
2018 18 55 18 91 60%
2017 32 37 22 91 41%
2016 28 36 27 91 40%
2015 31 36 24 91 40%
2014 35 35 21 91 38%
2013 34 34 23 91 37%

We understand now why there are so many draws in tournaments like Norway Chess, and a look to the past will confirm it. In the famous tournaments in Linares, back in the 80s and the 90s, the percentage of draws was much lower, between a 40% and a 50%. But when the number of players was reduced, and the level of the field increased, this percentage rose dramatically, reaching more than 70%. Once again, we had the same problem: all players were world-class and they knew each other too well.

The Sofia-Corsica rule, prohibiting draw offers before move 30, has improved this situation. However, there are still many short or boring games, especially when white plays in a very solid way. 

A lot of draws will occur when the best players face each other. Thus, the idea of declaring a winner in each round using rapid chess makes a lot of sense, especially if we want chess to be more attractive to the general public. Here are my suggestions for improvement:

A winner every time

In each round a match winner shall be proclaimed. If the first game is a draw, another game will be played with the remaining time, and so on, until someone wins a game. The time spent on each move takes enormous value. 

Colours balance

The player who is white in the first game will play black in the rest. In this way, the chances of each player are balanced, and short draws in the first game are discouraged. 

Scoring

The winner in each round receives one point, the loser zero. Easy to understand for the public and the press.

Rating

For rating calculation each game is computed separately in the corresponding modality (classic, rapid or blitz), depending on the duration of the game. 

Time control

handshakeEach round has a maximum duration, which the organizers establish according to their convenience. 

In all games, the players have 3 seconds per move, not cumulative, plus an initial time:

  1. For the first game each player has two and a half hours, or less, according to the attached table.
  2. For the second game, each player has the time not consumed in the first game, plus 15 or 30 minutes.
  3. For the rest of the games, the players will have only their remaining time. 

The time of 3 seconds per move is not cumulative. If a player moves in 3 seconds or less, his clock remains stopped. After the first 3 seconds, it begins to consume the main time, until it is exhausted, in which case he loses on time. There is no need for the players to record the games, the electronic boards or the arbiters can do the job. 
 
Each player has a time based on the maximum duration of the round, set by the organizer.

Round
max. duration
Time each player receives ... ... ...
  1st game 2nd game More games In all games
7 hours 150 minutes Remaining time + 30 min. Remaining time 3 seconds per move
6 hours 120 minutes Remaining time + 30 min. Remaining time 3 seconds per move
5 hours 120 minutes Remaining time + 15 min. Remaining time 3 seconds per move
4 hours 90 minutes Remaining time + 15 min. Remaining time 3 seconds per move
3 hours 60 minutes Remaining time + 15 min. Remaining time 3 seconds per move

The idea that we have seen in the Norway Chess is good, but I think it can be better. To give 1½ points to the winner of the Armageddon game makes things unnecessarily complicated, not to mention that we must explain that a draw is good for black in the second game, etcetera.

What do you think?


Miguel is a Spanish chess player born in Barcelona. He is the editor of the chess magazine "Peón de Rey" and contributes to the daily newspaper "La Vanguardia" where he tests readers with different chess moves.

Discuss

Rules for reader comments

 
 

Not registered yet? Register

ricardoalves ricardoalves 6/23/2019 04:07
I think this ideas - shared by M Illescas or those tried in Norway or St. Louis - over complicate the game. For me it would be a matter of adjust the total thinking time per game and the relation points between drawing and winning, for example:
- game duration: 90 min + 30 sec per move
- points: 5 point a win, 2 a draw (2,5 to 1 relation) if you don't want to overate a win; or 3 points to win, 1 point a draw (3:1 relation) if you want to increase the value of a win. You can even think of a 2.25 to 1 relation - 9 point for a win, 4 points for a draw.

Another idea is that you win different points according to what colour you play with. For example, 5 points if you win as black, 4 points if you win as white, 3 if you draw as black, 2 if you draw as white. Of course this alternative won't save the game as probably players playing as black will try even harder to draw. But on the other hand probably it will be risky to do that!

What do you think?
dumkof dumkof 6/23/2019 02:05
High quality chess inevitably results with higher draw rates. Computer chess proves this also. The stronger the engines, the higher the draw rates.
If some arbitrary formats "succeed" to increase the decisive game rates, it's only because of increased blunders and lover chess quality.
If this is what the audience wants, then go on with your "creative" formats.

I fully agree with the last comment of michelhoetmer. And almost fully agree with malfa (except the 3-1-0 points rule. I prefer the classical 2-1-0 point rule plus price money as a stimulus) Thanks for both comments.
michelhoetmer michelhoetmer 6/23/2019 12:38
Bad idea! Chess does not get more interesting for the public simply by shortening the thinking time. With a faster pace people will understand less and not more. If you want mayhem? Just give every player a minute and you will get total chaos. In fact Illescas pointed already a solution into the right direction: mix the playing strength like in Tata. That makes for more interesting chess. It forces strong players to try to win. And don't invite GM's who play too many uninteresting games. I even dont follow those 'top class tournaments' anymore. It's always the same bunch of players with rather predictable results. Boring!
malfa malfa 6/23/2019 10:42
Forcing a game to go on with increasingly short time controls is as stupid as having two tied marathon men run for further 800 m, then 400 and so on. Draws are part of the game and are not bsd in themselves, since a lot of interesting games have always been played which ended in a draw. What must be done for sure is to minimize uninteresting draws, which means in the first place to simply forbid draw offers and to allow only draws by repetition and stalemate. This of course is not enough, since two players who are determined to draw will always find a legal way (just look at the games between the Muzychuk sisters...), therefore I agree that playing for a win should be favoured by properly changing the allocation of prizes, rather than introducing the 3-1-0 points rule: money is always the most convincing stimulus.
bishopric bishopric 6/23/2019 09:28
I must complement M Illescas on making a positive suggestion and Norway Chess for experimenting.
I think the worst aspect of the Armageddon experiment was the low quality of some of the Armageddon games. A result of fatigue, both in a long tournament and playing a second game after a drawn classical game. If Illescas's idea were taken up this would get even worse, with some really poor 3rd, 4th etc games, especially later in a tournament.
On Illescas's idea I have a possible unintended consequence I think likely. Imagine Carlsen playing under these rules. He is a fast player and exceptional at faster forms of chess. The pragmatic thing to do is for Carlsen to play absolutely risk free chess, especially in game 1, and just rely on his ability to play good chess fast.
It is not just Carlsen, others would adopt a pragmatic take no risks policy as well.
Sure, there would be no draws, well sort of. But would it result in more interesting chess- maybe the opposite. Between players of near equal strength it makes speed of play by far the most important thing. Followed by being able to spring opening surprises in the 2nd and later games. Is this what we want? I don't.
hansj hansj 6/23/2019 09:28
This one is really easy. Do not invite players, that do not want to play chess.
jsaldea12 jsaldea12 6/23/2019 09:19
I still contend, make it simple and easy to understand, make it STANDARD, not complicated, more condition, and more confusing. Reiterating, the standard I am suggesting is: whether in classic, rapid, blitz, gtive a win 3 points, draw 1 point, and loss zero..
fgkdjlkag fgkdjlkag 6/23/2019 08:49
@niccroad, not true that organizers are not interested in chess960. The Fischer Random world championship is this fall and a lot of top players are competing - Carlsen, Nakamura, Aronian, etc. Mr. Sinquefield has sponsored 2? top chess960 events recently. In last year's online blitz championship there were a couple chess960 games each round. There was the big Nakamura-Carlsen chess960 match last year. And based on interviews/comments, the vast majority of the top players (GMs and super-GMs) are big proponents of chess960.
GrayDuck GrayDuck 6/23/2019 08:17
The proposal is interesting but it is dead on arrival because it is too radical and complicated. I think a better approach is to advocate for incremental changes leading to marginal improvement. For example, we could push for modestly shorter time controls--maybe a ten percent reduction--and aim for a fifty percent draw ratio.

I would like to take issue with this idea that the absolute quality of the game--without considering speed of play or time management--is the only factor that determines whether a game is exciting. If that was true, tic-tac-toe (noughts and crosses) would be the most exciting sport on the planet. The problem with that thinking is that routinely high level of play leads to focus on the rare mistake rather than exceptionally brilliant or innovative play. I think a better approach is to aim for moderation--there should always be a good mix of terrible play, mediocre play, and brilliant play.
Lilloso Lilloso 6/23/2019 07:41
The system proposed by M. Illescas looks a bit complex to deal every day in tournament's strategy and will be exhausting for the players. Are draws such a big problem ?
judge57 judge57 6/23/2019 07:38
I agree w/ Wok and peterfrost. Highly restricted super-GM tournaments are the problem, not something to be preferred. Mix the list up a bit each time.
rasagyan2018 rasagyan2018 6/23/2019 07:08
If you're gonna make the players work so hard in each round.. Give them a rest day after every 2 rounds.
vipiu vipiu 6/23/2019 06:45
Have stalemate considered as half win(scored 2/3 or 3/4) for the side with more material. In this way you would force many games to be played until the end. Also the perpetual, have it considered as half loss(scored 1/4 or 1/3) for the side which is forcing it.
Imagine so much existing theory in place still valid but yet allowing new theory to be written (up to the level of pawn endings)...
peterfrost peterfrost 6/23/2019 05:57
Agree completely with Wok. The draw problem results from all players being elite and too evenly matched. A mix of elite players and merely "strong" GM's (including interesting young players) is a much better solution than radical approaches such as Armageddon games.
johorsky johorsky 6/23/2019 05:09
@Masquer

Obviously I was not clear enough :-(
A win in the classical game: 3 - 0, no tie-break.
A draw in the classical game and a win in the tie-break: 2 - 1.
We would have a winner every time, and winning the classical would be worth more than a tie-break win.
Masquer Masquer 6/23/2019 04:20
@johorsky

I am all for weighting the classical wins more -- but, if the first, classical game is won, why would there be a need for a tie-break?
I suppose one could have a multiple time control format, or a faster tie-breaker after a drawn classical match.

In my proposed system, if the classical game is won, no second game is played, and the winner gets a 2-0 score from the round (see my earlier post below).
Masquer Masquer 6/23/2019 04:04
@niccroad
1) if White gets more wins as a result of Black playing (more loosely) for a win himself, then that surely can't be a bad thing if we want to see more wins and better entertainment
2) if you don't ban the short draws, you will have a lot more of them - not good
3) chess960 has very unsightly castling rules and lacks the elegance and harmony (not to mention the popularity) of conventional chess
4)modifying basic rules such as stalemate would mess with the whole fabric of the game -- all of sudden just being a pawn down can be a losing proposition, and who would take any extra chances then? On the contrary, we need to find ways to encourage and reward risk-taking.

Always beware of unintended consequences.
Masquer Masquer 6/23/2019 03:50
@deasla
Too radical solutions can have unintended consequences, like one player lying down for a friend and letting a drawn game slip away since it's all the same for him, but gifting to the opponent an undeserved win.
There are better ways to discourage draws - if two draws count a little less than a win.
A draw can be the result of a well played game and it is simply not fair to count it as a loss.
Neither simple nor fair.
Masquer Masquer 6/23/2019 02:58
I like the colours balance concept in case of a draw. If a classical game is drawn, then a second (rapid) game is played with colours reversed.
If the initial classical game is won it counts as 2 points. Otherwise, add up the results of both games (if a reversed colours game is played, in case of a draw in the first game). So, if both games are drawn the result is 1-1. If the second, rapid game is won, then the winner scores 1.5 points from the encounter.
Very fair and logical and avoids the Armageddon mess which encourages Black to draw at all costs.
Also, no need to play additional games with the remaining time left (a ridiculous and undignified idea.)
daftarche daftarche 6/23/2019 12:54
I always felt someone who has the upper hand and stalemates his opponent should be given some kind of advantage in chess vs someone who gets stalemated. The only thing I change in chess is having a better tiebreak for someone who has stalemated more or a worse tiebreak for someone who has gotten stalemated more. this must be the first tiebreak in all tournaments and even matches. In a world championship match two players end up 6-6 after 12 games lets look at who has stalemated more and he is the winner. this way you will see fighting chess and players have to play accurately until the very last moves.
Aighearach Aighearach 6/22/2019 11:43
The fact is, chess is not going to make money the way soccer or racing do.

3-1-0 with limits on draw offers has been shown to be effective at encouraging dynamic play when appropriate, without ruining the traditions of the game.

But simple fact: chess is still not big money. That's the only reason for these new ideas; people still think chess might be big money with the right formula. That has nothing to do with improving chess for chess fans, though!
FlannDefence FlannDefence 6/22/2019 11:27
Grandmaster Illescas' proposal is interesting, and deserves some serious experimentation, but the discussion so far demonstrates that most so-called solutions amount to futile and arbitrary means to deny what chess is: a game where it is natural to have some wins, some draws and some losses. I like chess. I don't really care so much if the general public don't pay much attention to it. In all honesty it would be strange if they did. Chess can never be like football, tennis or even boxing. Let us who appreciate the game enjoy it - stop trying to make it into something it is not.
niccroad niccroad 6/22/2019 11:27
I think there are a number of issues with the various proposals tried,
1) Changing the scoring to reward winners 3 and draws 1 doesn't really work well enough. Ultimately at the very top levels playing to win with Black is always giving White a much increased chance of winning in return. Also ratings are too important for players to really abandon the traditional scoring systems incentives.
2) Banning draw offers to 30/40 moves results in many still content-less games but with more perfunctory moves.
3) Chess 960 has a much lower draw percentage and should at the highest levels but doesn't appear to be favored by tournament organizers to any degree.
4) The proposals for giving half wins and half losses for final positions where one side has a minor piece or forces stalemate should improve the number of results, but involve changing the rules of chess.
airman airman 6/22/2019 10:17
Every Sport goes through this. Eventually the rulemakers run everyone off. From small club level to national level sports .... Many follow this trend.
Chess has had the doomsday sayers for over 100 years relating to draws and the death of chess.
Draws wont kill chess. Constantly tinkering with format and rules will.

Latest example is NASCAR. FIDE of the past tried real hard to kill chess. Will todays rulesmakers succeed?
wok wok 6/22/2019 10:13
The best approach is described at the start of the article: Avoid "elite player only tournaments" and go for a mix of top players and prodigies. Fights are guaranteed and the discussion is closed.
Semyorka Semyorka 6/22/2019 09:49
@turok. That makes chess join sports like gymnastics and Formula 1. How much points will you give to black in case of a draw? 0.5632 or 0.6123? Don't do this!
turok turok 6/22/2019 09:32
very simple: Put a draw with black more point value than white-this gives incentive for white to play on-especially in a championship title game
Semyorka Semyorka 6/22/2019 09:28
The tables in the article do not really convince me that the amount of draws is a problem. So what is the problem? That a chessgame can end without a winner in most cases? No problem to me cause I think it is in the soul of the game!

When most players do have a problem, consider this: adapt the basic rules to stalemating your opponent is a win. Problem solved? Would you like that?
PCMorphy72 PCMorphy72 6/22/2019 09:06
As this article states itself, “The time spent on each move takes enormous value”. In this way, the normal reflection time of a normal classical game (as it ought to be!) tends to be unbiased in order to satisfy the pseudo-urgent need to have “A winner every time”, losing a lot of possible quality of the moves.
Of course, it would have a serious impact on the way how top-players play their classical games, although, as stated, it may improve the Norway Chess Armageddon format, as well as may have pros and cons compared to similar proposals ( https://en.chessbase.com/post/the-kasimdzhanov-shipov-draw-proposals-readers-feedback ).

Many years ago I proposed armageddons among many detailed overarching measures for both top-tournaments and top-matches, carefully considering low impact and low bias: https://sites.google.com/site/pcmorph72/articoli/wcc-cycle ( BTW, I think that Smerdon’s paper sucks: https://www.academia.edu/24439310/Fixing_Flaws_and_Stopping_Draws_A_new_chess_tie-break_system_based_on_directed_network_analysis )
In the meanwhile, being me an unknown nobody, I tried to support Seirawan’s ideas: https://en.chessbase.com/post/seirawan-a-radical-solution-for-the-world-championship
After years, we are still discussing the same “flying” ideas.

The truth is that nobody cares about any bias in the exotic systems adopted or proposed till today, at least until it will be usual to approach 90% in the percentages seen in this article, if classical time control will not be snubbed before than that.
Keith Homeyard Keith Homeyard 6/22/2019 08:41
telemwecomin Very similar to my suggestion about 5 comments lower. :)
tangabell tangabell 6/22/2019 07:55
Win: 4
Loss: 0
Current Draw = Half of 4 = 2
Proposed Draw = Half of Current Draw = 1
telemwecomin telemwecomin 6/22/2019 07:00
Why not have players play each other without knowing who they are playing? Granted, that means they are not sitting across each other and instead are looking at a computer screen, but I bet you that would reduce draws more than any other artificial system.
Rune Eliasen Rune Eliasen 6/22/2019 06:48
How should ratings be adjusted when you play a sequence of games as suggested?

Today the rating adjustment for each player is the K-factor times the difference between the actual score (0, 1 or 0.5) and the expected score (calculated from the players current ratings).

When you play a sequence of games until a winner is found with a maximum of 4 games, the actual score can be defined the following way:

Win: 1
Loss: 0
Draw, win: 3/4
Draw, loss: 1/4
Draw, draw, win: 2/3
Draw, draw, loss: 1/3
Draw, draw, draw, win: 5/8
Draw, draw, draw, loss: 3/8
Draw, draw, draw, draw: 1/2
mamago92 mamago92 6/22/2019 05:53
Win : 3 point
Draw : 1 point
Loss : 0 point
ChessTalk ChessTalk 6/22/2019 05:48
Not discussed is the problem of following the armageddon game. It's often hard to follow fast games online. I would love to see the games. I love the Classical games but also enjoy the fast time controls. If you want to make chess more of a spectator sport, fix the system used or make armageddon games with some increments.

To respond to the article's suggestions: We aren't dummies. We can wade through a 1/2 point score. I think some system that rewards wins would incorporate money incentives. This is not my suggestion but many others have brought it up.

I've followed chess a long time and I really don't like seeing draws. I love the tie break system that encourages risk by making decisive games more important than steady draws. In the recent Karpov tournament, I was glad to see the young powerhouse, Artimiev win on tie breaks against the strong Jakovenko due to more wins. Both players are deserving but that last round win by Vladislav was impressive. And the win was crowd worthy.

The suggestion of a full point in armageddon isn't terrible. It includes the draw in classical and gives a 1/2 point to the winner in armageddon. Okay. That works. The time used idea is hard to parse.
RaoulBertorello RaoulBertorello 6/22/2019 05:42
@klklkl, KevinC: klklkl is right here, classical chess is just one of the 960 variants given by 'Chess 960' ('chess' for brevity), not the other way around. And therefore classical chess can be seen as one hyper-studied variant of the 960 possible ways chess can be played. And of course, sticking always to the same hyper-studied variant brings about its turning into a proof of memory. I totally agree with klklkl, the only easy, natural possibility to avoid all these draw games is to adopt Chess 960, as all and only the rules of classical chess are still in place.
Keith Homeyard Keith Homeyard 6/22/2019 05:12
In the Hastings 1895 tournament, if you excuse my ailing memory, although an all play all tournament you didn't know either your colour or next opponent unti; half hour before start of game. This must have a great effect in reducing preparation as it would also be less clear whether to be playing for a win! :)
Crik Crik 6/22/2019 04:45
I agree with Rokko too, and not to forget that armageddon games favors youth over older players
deasla deasla 6/22/2019 04:45
I propose a radical solution.

Win : 1 point
Draw : 0 point
Loss : 0 point

Simple and effective. If you want to win a tournament you got to win games. If you're above the field you can't relax and make draw until the end. And you would say : Why on earth would players fight for a draw in difficult position then ? Simple, they would do it so that their opponent wouldn't gain point either. They would think of the tournament as a whole and make sure no one would clim up the point ladder before them.
fgkdjlkag fgkdjlkag 6/22/2019 04:34
As klklkl has pointed out, the problem is more fundamental than having a decisive result. Look at how many top games today are a memory test, a complicated line of 20-30 moves that is equal if the opponent remembers the right moves. Look at how many players admit this after the game in the interview, that they were just testing the opponent's memory. Look at how boring the commentary is for these games.

None of the proposed measures are going to work because it's not the player's fault. it is the nature of high-level chess, where as the author of the article points out, draws are a natural result. You can prohibit draw offers after move 30, pay more money for decisive results, etc, but the players can't get a win from a drawn position.