Monday, September 22, 2008

How do you solve the problem - with Sofia?

The brief draw that was agreed by Adams and Short at Liverpool, and the subsequent shortness of the leading games in the final round, have raised the question, once again, of the problem of the short draw in chess, whether something needs to be done about it, and if so, what.

I've never felt it very urgent to address this question, for a number of reasons, one of which is that it's a mite hypocritical for me to do so seeing as my ratio of draw-offers-made-to-draw-offers-received runs at something like five or six to one. Another is that a lot of commentary on the subject tends to come from The Man In The Bar, who has no interest in understanding why short draws happen or indeed in anything but mouthing off about it. (Sometimes this commentary comes from people who do know and understand why they happen, but mouth off anyway and pretend they don't.)

I found myself agreeing very much with John Saunders:
the players themselves... have a living to make and a limited shelf-life (very few players maintain their status and earning power after the age of 45-50 these days). They simply adapt their approach to the prevailing conditions. They carefully assess their share of the prize and how many rating points they stand to win or lose. Where the two players' aspirations coincide, a draw is going to happen regardless of the demands of the spectator. They agree draws because they can.
It's easy to complain when somebody plays for a draw: less easy to explain why they should listen to you when it's their living at stake rather than yours.

Players have other reasons than money, too. They are trying to chase IM and GM norms, for instance, or trying to avoid risking too many rating points. (The present writer threw away a dozen Elo points at a single tournament this summer, by losing the last two games to lower-rated players. A couple of short draws would have suited my needs much better.) As ratings, and titles, have a sizeable impact on whether players get invited to tournaments and what conditions they receive if they are, it is asking a lot to expect the professional player to ignore them. And it is too, too easy to simply accuse them of a lack of fighting spirit.

A third reason is that I remember similar discussions taking place in the world of football when it was at a very low ebb twenty years ago. We needed to change everything, to rip up scoring systems to make this declining game more exciting before it died.

Points for goals (Jimmy Hill, I believe) or awarding penalties for every foul (Hill again) or deciding drawn games on the basis of a corner count (Peter Corrigan in the Observer) or having shoot-outs at the end (because the North America Soccer League was such a roaring success).

Eventually none of this happened: the points system was tweaked a bit (three points for a win) the playing rules a tiny bit (the backpass rule) and football turned out to be enormously popular after all, not because of those changes but because it always had been. The game was what it was.

Still, there is a point of view that we are in a different world from the one in which we used to be, and that while it offers chess an opportunity to attract spectators which has never really happened before (or not in the West, at any rate) a serious part of that audience will necessarily come from people who are not regular chessplayers, who like the game and will follow it live on the internet (especially if there is sympathetic commentary) but are not likely to show much understanding, in any sense of the word, if they log on in great anticipation only to see 1/2-1/2 come up beside the board twenty minutes in.

Which is not to say that the regular players are that fond of the grandmaster draw either - but we'll come back next time, because it's still our game. Just as football fans will almost always come back next time regardless of the performance they see this time, because it's still their team that's playing. The casuals, however, may likely not.

So if that is to be considered a problem, it's a problem first and foremost for people organising professional chess tournaments and as well as for people concerned with how chess relates to people in the wider world and how it fares within our culture. John continues:
Delivering entertainment to spectators is not an issue for professional players themselves to decide – it is for tournament designers to address. The need to get tournament formats and prize structures right is becoming urgent because we now have a new – and huge – internet audience to consider... the problem obviously needs to be addressed if professional chess is to move forward and attract sponsorship
So if it is to be addressed - how? With what specifics in mind?

One can of course find any number of past discussions of the issue by means of a short search of the internet - sometimes involving the most convoluted ideas as to how the problem should be addressed - but this is today's. So today I'll identify and discuss five different potential ways of addressing the short draw problem, these being:
  1. the Sofia rule
  2. cash prizes for winning individual games
  3. skewing prizes towards the top
  4. three points for a win
  5. invitations policy.
I'd reject on principle more complicated ideas that involve splitting the points with a second, tie-breaking game or awarding a larger portion of the point to Black than White. You want tournament tables to be readable: you also want the rules to be comprehensible to the outsider. What is not simple, in this situation, is not good.

1. the Sofia rule

This has been tried out in a number of major tournaments now: it seems popular and is likely to be seen more often. Simply, it refuses the players the right to make draw offers directly to one another: they must be made through the arbiter, which can only be done in cases of perpetual check, threefold repetition and "theoretically
drawn endgames", the last of which requires some expert knowledge on the part of the arbiter.

The main success of Sofia is surely in eliminating the draw that occurs when either equality has been achieved or when an unclear position has been arrived at: the "cowardice" draw, if you'll forgive my coining a phrase that is somewhat short of fair. What it doesn't help with so much is the prearranged draw, since it's always possible to go down a line with an early threefold repetition or perpetual check provided you trust one another (or the line has been agreed on beforehand). There were some early repetitions in Liverpool - and I'm not thinking of the Adams-Short game here - that did make my wonder whether they had happened naturally or were just ingenious ways of achieving a quick draw while looking like you'd tried to do the opposite. I can imagine some people might think that's clever.

Also, Sofia is only suitable for a small and presumably closed tournament: where there's dozens of boards the arbiter will have many better things to do than act as a letterbox for people's draw offers. So that would have counted it out at a tournament like Liverpool: or for that matter, the British. (Or at least, until I get my way.)

2. cash prizes for winning individual games

In some tournaments there are rules encouraging attacking play by awarding cash prizes for wins. I assume there are different arrangements in different tournaments, but in Benasque, for instance, I believe it applies during the last four rounds, on the top fifteen boards only. (How much the prizes are, I couldn't tell you - you think I'm that high up the pairings that late on?) How effective this is, it's hard to say: I'd have thought it unlikely to induce a player to take risks that might cost him or her a shot at a four-figure prize if a draw will keep them in the running.

But if this is an issue of professionals playing safe to keep themselves in the money, possibly the answer does lie in changing the structure of the money.

3. skewing the prize money upwards

Prize money is, of course, already skewed upwards, but I mean doing so to extremes, having a much smaller number of much larger prizes, so that nobody with a record consisting largely of draws is likely to get close to it.

I believe that this is common, or more common, in the US, and it makes a certain amount of sense. In poker, it's not unheard of to have winner-takes all tournaments - sometimes these are even on the television, which is something chess might like to bear in mind - and if it leads to a more enthusiastic and a larger audience, one can argue that if the size of prizes increases and the number of opportunities to win them increases also, then many professionals will find themselves better off in the long run. Especially if they show more fighting spirit than they do now.

Moreover how badly would anybody be affected if the bits-and-pieces prizes were to disappear? Annoyingly, I can't find the entire prizes list from Liverpool, which I remember seeing previously on the tournament site, but if I recall correctly then after the top four, nobody won more than £337. Not peanuts to me or you, perhaps, but I don't really believe that anybody lives by winning bits-and-pieces prizes every week. So if they have to win the big prizes anyway, why not make 'em bigger?

Worth thinking about, perhaps, though one suspects that as a philosophy it conflicts with (and will not work with) the practice of giving "conditions" to titled players. And if you do not offer conditions, the titled players will go elsewhere, to a tournament that offers them more security. If you do offer conditions - then the players are not so desperate for the prize money. That's why conditions exist.

Well, you say, then abolish conditions - people shouldn't be paid just for turning up. But in fact, most people are paid, in some sense, for turning up, so it is not such an outrage that a professional chessplayer should be. But more importantly, if we consider outcomes, then how good is the US model at producing and sustaining professional chess? Not very. And without so many professionals, you do not have much of a professional circuit. So there will not, perhaps, be so many big prizes to win, and those big prizes will not, perhaps, be quite so big. (Besides, how do you encourage players to keep playing when they're out of the running for the small number of players?)

It might be added - the tournament at Sofia, which prides itself on producing attacking chess, has no prize money. Everything is paid in appearance fees. Which provides a completely different lesson, or at least a completely different point of view.

4. three points for a win

I believe that when this was used at the recent Bilbao tournament it was the first time it has ever been the system at a major chess tournament anywhere: I'm not even sure that the system has been used at any great number of minor tournaments. It seems to have functioned well, without any real complaints or any real anomalies: then again, none of the competitors at Bilbao has a particular propensity to take a lot of early draws.

It seems to work well in football, despite the claim that it originated in the mind of Jimmy Hill and despite the illogicality of splitting two points such that one side receives three of them. This in itself has made it easier to contemplate in other fields. It is simple, widely understood and provided an obvious incentive not to take the draw. Though of course, that is not quite so straightforward as it may seem: it's not as if defensive play has entirely disappeared from football. It's not a simple matter to understand or explain, but to look at one aspect alone, it is still perfectly possible to defend in the knowledge that the opposition has an incentive to attack - and is therefore, perhaps, more likely than before to allow a good defensive side to score goals on the break.

Similarly in chess, there may be just as much incentive as before to play a cagey game in the hope that your opponent, greedy for the win, will commit themselves to attack when perhaps they should not. But we can live with that: from the spectators' point of view, they see a struggle all the same. The idea is not to insist that everybody plays like Tal: it's to discourage or eliminate the early draw.

There are, of course, other objections to three-points-for-a-win. I am not sure if it will mess about with rating systems, but another objection was voiced on this site by Richard on Friday:
3pts for a win is not a viable option when there are serious and material differences between playing with the black and white pieces. It would certainly necessitate a remodelling of the pairing system to prioritise getting a colour balance over pairing players on the same scores.

It also puts far too much importance on the "luck of the draw".

Therefore it can only be considered viable in all play all events (as, indeed, in football)
I wonder, though, how far is this true? Or rather, not how true, but how important? Most closed tournaments, for instance, already have a White/Black imbalance, as do Swiss tournaments with an odd number of rounds, and to do well in the latter it is already necessary to try and win some games with Black. Moreover while there clearly is some injustice involved, we might well decide on pragmatic grounds that this was relatively small and a small price worth playing for a more attractive game. And maybe, given that necessity is the mother of invention, grandmasters will get better at playing for a win with Black.

This might be true and it might not: but at any rate the idea seems worthy of wider consideration and I see no insuperable reason why it should be limited to all-play-alls.

5. invitations policy

This was the point made by John Nunn in an article for Chessbase on the draw question.
As for my own suggestion, it is really quite simple. I am constantly astonished at how often tournament organisers invite noted draw specialists to their event, and then throw up their hands in horror at the number of quick draws that ensue. We all know who the drawing experts are, and if you don't know then it doesn’t take much work with ChessBase to find out. It is up to organisers to invite players who show fighting spirit to their events. The category of a tournament isn't everything, and organisers could be more imaginative in inviting slightly lower rated players who show imagination and fighting spirit. When the drawing masters see their invitations dry up, it might encourage them to change their styles.
Nunn's point is a good one, though of course it's not quite so easy as that: I played in an Open in the Czech Republic in 2006 which also featured two all-play-alls, an IM and a GM tournament. The highest-rated player in the GM tournament, proceeded to play eleven draws, nearly all of them quick (and the exceptions being games in which he was losing early on and had to rescue the half-point). Was this a notorious draw-master? No, it was a world authority on the Black side of the Leningrad Dutch.

I have no idea why it happened or who was to blame. But I wouldn't be surprised if the organiser didn't take Nunn's advice and invite somebody else the next time.

I always thought it would be a good idea to hold a tournament and invite only players with a reputation for original and daring play - Jacob Murey, Mike Basman, Jonny Hector, players of that kind. Perhaps this will happen in the future: if organisers think it's going to get them (and therefore their sponsors) a bigger audience, they'll invite players people will log on to watch. And nobody will log on to watch somebody play nine or ten short draws.

I mentioned Sofia above - that's an annual super-tournament with an invitation-only policy which relies on the players producing fighting chess. Play hard or don't come back - it's a policy the organiser can only take if they mean it. It's also a policy easily abused. I thought, for instance, that Luis Rentero's behaviour as his Linares tournaments was high-handed and petulant, and disprectful to both players and the game. But if, after a tournament, you don't think a player has given of their best, then you surely have the right not to ask them back. And then other players know you mean it.

But - like everything else on the list it's an idea of limited applicability: it applies to invitational tournaments and only to these. It won't work so well, if at all, with opens, with world championships, with national and transnational championships. Even in super-GM tournaments it might not always work as well as Sofia might suggest. Because what happens if a drawmaster is world champion? What if some of the biggest names - and therefore the biggest attractions for the audience - aren't the most aggressively-minded players? The tournament organiser doesn't necessarily have the whip hand. The strongest and most famous players have the most clout. So what happens if they prefer to use that clout for peaceful purposes?

Does it really matter? Isn't it normal, anway, for the casual audience to find that a sport isn't all excitement and end-to-end action? I'm sure everybody else who follows football as a fan is used to the syndrome whereby people who haven't been to a game in their lives tune in for a World or an FA Cup Final - which game is usually more cautious than the average match - and then tell you that the sport you love is boring.

It's a syndrome that I can't necesarily see how we break. The more important a contest, the more likely it is to attract the casual spectator. But at the same time, the more important it is, the more likely the players themselves are to exercise caution and to consider no consequences other than what suits them best. They're supposed to - that's what professional sportsmen and women do. And you can't necessarily so anything fundamental about it without fundamentally altering the structure and integrity of the game. If you do that, the likelihood is that the casual audience soon gets bored and wanders off to the next fad.

Besides, does it really matter, in an Open, or even in a closed tournament, if some of the players are peaceably inclined? My early experience of watching chess in London involved the Candidates' Semi-Finals and the World v USSR match. I remember almost nothing about the former games - there were only one or two games to watch (or none at all, on one occasion, where both games were postponed). The other was fantastic, because with ten games going on, if it wasn't happening in one place it was happening in another. That's perhaps the way it is. Who cares if some players agree quick draws on this board or that? On another board, they won't.

Of course I underetand that it does matter to some extent, in the circumstance that the main attraction, the players people have logged on to watch, play quick draws against one another, as happened with Adams and Short at Liverpool. But if they're still going to be the players people will to see the next time, then what can you do?

Professor Robertson, chief organiser of the Liverpool tournament, comments:
Some thoughts on 'short draws'. No tournament promoter, myself included, likes to see games concluded without a fight. Sponsors get confused; and the vast hordes of online spectators are disappointed at prospective enjoyment spoiled. But efforts to reduce or eliminate the 'short draw' offend against other aspects of the contest. 'Sofia rules' ban draw offers altogether; variants ban draws in under 30 moves, and so forth. These are, in my opinion, artificial intrusions that compromise the integrity of the contest. 'Bilbao rules' attempt to avoid this by introducing an incentive to win that doesn't interfere with the contest itself. Where 'Sofia' is the stick, 'Bilbao' is the carrot....

...the type of question we need to answer is: when is a win not in itself a sufficient premium? Put another way: at what point do additional incentives equalise the risk of seeking a win with the security of taking a draw? If we knew the 'tipping pont', we could design the incentive.

Clearly no sensible 'win' incentive could work in certain situations. Two examples from opposite ends of the problem. Last year in the GB v China match, Jonathan Rowson took heavy defeats in the first three rounds. He became disspirited both personally, and for his team. He simply had to end the sequence of defeat. No one raised an eyebrow when he quickly halved out in his remaining games. Jonathan didn't lack fight; he lacked form. Second: this year in the final round just gone, Jan Werle sees his closest rival draw. A draw now nets him a prestigious title and tournament win, not to mention plenty of cash. So an uncontested draw, it is. Between these two extremes, some 'tipping point' exists. But quite where is hard to establish.

For that reason, I've had no draw limitations in any tournament I've promoted. I prefer to trust the great players to look to their responsibilities and look after the game. It is after all their livelihood.
Fair enough - mostly - although I seem to remember the same Professor Robertson arguing, in re: the GB v China match, that in order to produce fighting chess it was merely necessarily to ask the players nicely. (I paraphrase, but I do not think I distort.) This isn't really the case, and if it were, it is unlikely that there would be a debate in the first place.

As there are a number of ways of addressing the problem - including deciding that it is not so great a problem that it needs addressing - and as all of these may be useful in certain circumstances but not in others, what I imagine will happen is that they will all be tried, at different times, by different organisers. The outcomes will be evaluated and discussed, in different ways and from their different perspectives, by the different groups involved, being professional players (of different standards) and organisers and sponsors and writers and chess fans both serious and casual. And at the end of the day, they will come to different conclusions. But hopefully something will be learned.

In this piece, however, I've tried to avoid coming to any conclusions at all. Which has not been hard, since there are none that I have come to. It's just a discussion. Over to you.


Anonymous said...

Has the "3 points for a win" rule been used anywhere other than Bilbao?

In the final round at Liverpool, Adams knew a win would guarantee him at least 2nd, and he duly drew in 13 moves before a single piece had been taken, against a player 143 points his inferior, in an opening where he is considered a world authority. If the Bilbao rules had been in place however, a draw could have seen him beaten to 2nd place by the people half a point behind instead, giving him a bigger incentive to play for the win.

Adam B.

Anonymous said...

If you use 3 points for a win in a swiss, you shouldn't do the pairings the traditional way since 1 win 1 loss no longer equals 2 draws. So the last round pairings could have been different.

Apart from the colour imbalance issue , another reason why 3 1 0 isn't logical for a swiss is ratings imbalance. For example why should two draws in 50 moves against near IM opposition score worse than a win in 10 moves against a tournament back marker?


ejh said...

I can't see what the strength of the opposition has to do with it - surely a draw against a 2450 scores less than a win against a 1600 as it is?

Anonymous said...

2 draws against 2450 would score less than one loss against 2450 and one win against 1600 in a 3 1 0 scoring system.


ejh said...

Yeah, but all that really says is that a win is worth more than two draws. It's anomalous, but the same is true in football. A draw against Arsenal and a draw against Stoke* is worth less than beating Stoke and losing to Arsenal.

[* if any Stoke fans reading will forgive me using their club to stand in for the 1600. As I support probably the worst fully-professional club in English football, no offence is intended.]

Anonymous said...

In a swiss tournament with a large number of entries and a wide spread of standards, the analogy is rather more that if you draw with Arsenal in the first round, then you get Stoke in the next. If you lose, you get a pub team.


Anonymous said...

Just an idea.

Why count draws?

A win counts as one.
A loss or draw counts 0.

Only playing for a win would make sense , hence hard fought games going all the way.

From a rating point of view a draw counts as void game.

ejh said...

Well, county cricket used to do this (until they realised it was absurd) but it really does neglect the fact that a draw doesn't just happen because people can't be btoerhed to play for a win. And somebody trying to hold a draw - as in cricket - is a really big part of the game, it produces much tension and there is much skill on both sides of that contest. Offer no reward for success in that endeavour and you simply distort the game, I think.

Anonymous said...

Why not have 3 points for a win,i for a draw, 0 for a loss and interpreting this into the Grading system may be add 75 points to opponents Grade for a win, minus 25 for a draw and minus 50 for a loss.

There are also plenty of draws at club level with players offering a draw because they are scared of losing, bored etc.

Anonymous said...

The previous comment should read plus 25 points for a draw.

Whatever, I'm anonymous.

Anonymous said...

Seems a good idea, but surely

plus 75 for a win
plus 0 for a draw
minus 50 for a loss


Anonymous said...

+75 for a win, -50 for a loss? Well, you'd cure the deflation problem in the grading system, anyway.

3 points for a win? Well, I'm not a big fan, myself. (Although I grant the rule would favour me.)

Anonymous said...

I can't believe anyone is seriously considering 3-1-0 for a bog-standard Swiss with no spectators. Who cares about quick draws when nobody's watching?

Alu Chin