Keano wrote on 04/25/16 at 11:31:10:

This already happens in the qualification cycle in much worse ways, though?

One of the players in the candidates is a wildcard selected by the sponsor. You get selected based on which country you are born in and who throws out the money - actually, if some billionnaire played chess at 2750+ level and sponsored every candidates tournament, he could potentially buy himself into the cycle every time?!

Two of the players in the candidates are selected based on average ELO over the course of the year, going down to decipoints.
This is bad enough considering the natural inaccuracies of the ELO chart, but worse if you consider that it is quite possible to lose rating - Grischuk started the year at 2810 rating; if he had REALLY wanted to qualify for the Candidates, he could've just not played for the entire year except getting 0.5 points out of the first match in the World Cup and he would've been a shoo-in.
Consider the possibility that someone loses their rating-qualification for the Candidates because they simply didn't get kicked out of the World Cup early enough (with a first round loss going 0.5-1.5 you'd lose significantly less rating as a 2800+ player than if you went 1-1 all the way up to the semifinals and passed every round through the playoffs) and the entire system just look ludicrous.

Both of these things look like far bigger issues to me than the given tiebreak used within the tournament - when half the Candidates are already "unfairly selected", what does the tiebreak used really matter anymore?

---

Blitz Playoffs are entirely No-Go for me personally -
It's just too different games compared to classical; and deciding the challenger for the classical world championship via blitz games certainly has a worse ring to me than "most wins in classical".
Of course, the same applies to the WC match itself - does anyone really want to see a Carlsen-Caruana match that went 6-6 in classical get decided by Carlsen winning 5-0 in some Blitz playoff? I would certainly prefer a coin toss over that.

Rapid may be alright.. but again, I dislike separate playoff matches. Something along the lines of playing a separate Rapid tournament prior to the Candidates themselves that will determine your tiebreak score for later (So in the event of a tie, the one who placed higher in the rapid wins the tiebreak) is an idea IMO, but that again leaves many questions and lengthens the entire event.

The current tiebreak is doing its job (to whatever extend) while making the length of the event largely clear for organizers and players alike. It's.. workable, I guess. As long as the tiebreak is known in advance and doesn't get changed over the course of the qualification, that's pretty good for FIDE already :x

Just of note: I would disagree that you can judge the excitement of a game by its end result at all - many a great attacking game ends with the attacker settling for a perpetual; and I doubt the "general public" gets too fired up by 95 moves of shuffling pieces back and forth until a drawn R+B vs R+N ending transforms into a won one.
If you wanted to make spectators happy, how about a "Most sacrified pieces" tiebreak? 

Overall, you'll never make everybody happy.
My tiebreak idea for the WCh (15 rounds, challenger has white in one additional game, in the event of a tie the WCh retains his title) will probably get called downright insane by players and forumusers alike, but shrug.

Stigma wrote on 04/26/16 at 01:18:40:


No. Cronyism, greed and stupidity are operating, here as elsewhere in the world, and those of independent mind must not act as if the forms those currently assume are inexorable.

I agree that the world championship and qualification for it should not be tampered with lightly. But I get the impression that for some people the sanctity of the cycle trumps all other concerns. Hypothetically, if we could make the cycle perfectly fair at the cost of losing 3/4 of the audience, who would find the chess played as a consequence too boring, should we do it?

I don't think so. We will likely never see candidates tourmanents with four games between each player or matches like the endless first Karpov-Kasparov WCh match again. The world is turning faster, and chess must adapt to it. The world championship cycle has the biggest media potential of all tournaments, and for that reason it can't afford to completely ignore what the mass audience wants to see.

Besides, is a qualification decided by a tiebreak really a bigger spot on the pure cycle than the split into two cycles in the 90s, Shirov's qualifying for a match for the world championship but still not getting to play, or (under the current system) someone potentially qualifying via a wildcard because the candidates tournament is played in his home country, or he is on good terms with the sponsor? I don't see it.

But I agree playoffs are better if possible, provided they don't degenerate into blitz or armageddon randomness. For a tiebreak, if you have to use one for whatever reason, most wins is one of the best, if not the best, apart from head-to-head result. I certainly prefer it to a roulette wheel or a coin. 

P.S. I also have to point out that even if Grischuk ultimately failed to get to the match in 2012, he came extremly close, and it would be hard to argue that he would have done better with another strategy. Andreikin managed to get all the way to the 2013 World Cup final (and hence a spot in the 2014 Candidates tournament) with essentially the same strategy. Apparently both he and Grischuk knew they were comparatively better at rapid than at standard time controls. But yes, this tells us more about the weaknesses of mini matches, and such a drawing strategy would not work well in a candidates tournament follow by playoffs. There's a very high probability that someone will end up with more than 50%.

To clarify my earlier comment, because it seems like there are two separate discussions going on here.

One discussion is whether in general it is better to have tie-breaks or play-offs. I can certainly see the case, particularly in Open tournaments where tie-breaks are pretty much essential.

However, this is a Candidates 2016 thread, and I think I made it clear, and btw I dont see the term "ludicrous" as being too strong at all, I would state it that way to put it very mildly, that selecting the challenger to play a match for the World championship based on a mathematical tie-break formula is simply not on. It goes against the whole tradition of the event going back to Steinitz, besides being simply unfair on the player who tied for first place but is denied a shot at the World ch title in such a crude way.

I agree that "fewest losses" is hypothetical and unlikely to be picked up--nor should it be adopted, in my mind.  I'd rather just see a playoff.

I would like to agree that the "most decisive games" tiebreak at least contributed to exciting chess by affecting players' opening preparation, but I simply can't, unless we get some interviews from top players stating so.  For instance, the Ruy Lopez showed up 11 times, 7 of which were Berlins and 2 of which could have been a Marshall (which is just a drawing line at the top).  The last two featured 6.d3 and likely would have been headed to Marshall territory otherwise. 

So on cursory glance, I don't think that the tiebreaks changed anyone's preparation or their overall mindset of "win with White, draw with Black." 

Also, I just noticed something that I'd overlooked about the tiebreak system used.  Here it is:

1. The results of the games between the players involved in the tie.

If they are still tied:

2. The total number of wins in the tournament of every player involved in the tie.

If they are still tied:

3. Sonneborn – Berger System.

4. If there is no clear winner with the above 3 criteria, there will be a special competition between the players who still remain tied after using the 3rd criteria (Sonneborn – Berger): after a new drawing of colors, each tied player will play two (2) tie-break games with the other tied opponent(s). The games shall be played using the electronic clock starting with 25 minutes for each player with an increment of 10 seconds after each move.

I can't believe it!  Perhaps they figured that remaining tied after the first 3 conditions was a mathematical impossibility, but it looks like they had a provision for a playoff.  Why not move the playoff to the #1 or #2 spot?   Ridiculous not to, in my mind. 

ErictheRed wrote on 04/23/16 at 16:34:03:

Subjective or not, can there be any doubt whatsoever that a majority of the interested audience prefers ambitious play and some wins now and then to lots of safe, risk-averse chess? And the tiebreak you end up with is 100% objective: it is known to the players beforehand and consists in a simple count of 1s and 0s.

ErictheRed wrote on 04/23/16 at 16:34:03:

Both this question and the objectivity issue above should be put equally to all the alternative tiebreak systems. My argument was never that more wins is fantastic, but that it could be the least bad of many problematic alternatives. I am starting to wonder whether the real disagreement here is that I find most of the other tiebreaks at least as bad as more wins, while you and others think some of them are quite good?

Head-to-head result is obviously a good tiebreak, and seems to be used as a first tiebreak ahead of more wins in these tournaments. But using rating (as in your more recent post) sounds very strange and unfair to me. Sure, some players were more certain of getting to the candidates tournament than others, but once they are there and the clocks start, they should have an equal starting point. This seems so self-evident to me that I find it hard to argue for it. For one thing, it has overwhelmingly been the case throughout the history of tournament chess, as in most sports leagues.

Playoffs are great, and I'm not against them if they're economically and logistically possible. But once you get to stuff like blitz and armageddon, you are really playing a different game, and this is unfair to players who are much better at long time controls. There is also something absurd about deciding the most serious tournaments there are with an "unserious" game like blitz. So if possible, I would prefer to go no lower than rapid. Playoffs are also what led to the very negative (for the image of the game) strategy used by Grischuk and others, mentioned by katar earlier.

Sonnenborn-Berger has a long tradition, but organizers seem always on the lookout for alternatives. It rewards unstable players, or those who do better against stronger opponents but slip up against weaker (or out of form) opponents. I don't think this is any better than "more wins", though just in the case of a Candidates tournament, I suppose you could argue that players who perform better against stronger opponents might therefore also do well against the best of the best, i.e. the World Champion.

So to summarize the case for more wins as a second tiebreak, as requested:

- It's objective, easy to calculate and no less fair than the other known tiebreaks (apart from head-to-head result and playoffs, at least down to rapid)

- It rewards ambitious play

- It may also actually lead to more ambitious play and/or more wins and losses (awaits research), which is an important secondary goal for top tournaments once fairness is assured

- It avoids some economic and logistical problems with playoffs

Disadvantages:

- It may be unfair to solid players who tend to draw a lot of games

- It's vulnerable to collusion by players agreeing to play two decisive games between themselves. Though I don't think this is a big problem in the current elite.

LeeRoth wrote on 04/22/16 at 16:56:28:


I'm actually curious about this: could it be possible that a playoff would generate more revenue?  Certainly a World Series that goes to 7 games generates more revenue than one that goes to 5 (from advertising, etc).  The economics of chess are different I realize, but then I can't really see how tacking on 1-2 extra days to an event could cost THAT much more.

The problem with the most-wins system is that it fundamentally misunderstands the nature of the world championship and indeed chess itself. Chess is a war game, and in any war, the object is not to lose; victory is nice, but not essential. Fewest wins would be a considerably better tie-break.

Most wins is designed to promote interesting chess. That's fine in exhibition events like Linares, but it's anathema in a serious tournament. The world championship is trying to find the best player. Rewarding a particular style of play skews that objective and is fundamentally wrong (and typical of the cretins chess is presently run by).

A play-off is essential. I really don't know why this is so hard to organise. If not that, literally any tie-break system would be better than most wins. Fewest wins would be better, or tossing a coin, for instance.