How long before fide realizes we need candidates matches?

I think most prediction based on probability and ratings would likely yield the wrong result.

Here is why. I have to pick a popular sport. Say NBA, NFL, NHL or Major League Baseball. Can you predict who will be the champion of those sports for the year 2013/2014? I guess you only have less than 25% chance of naming a club that would become the champion of those sports.

I think the grand prix series of tournaments is one of the best systems FIDE has come up with in a long time. It's extremely rigorous, as results over a series of tournaments matter, but each individual tournament also seems to quite palatable to sponsors. It also manages to include a very large portion of the world's top players.

I would like to see much more weight given to the grand prix. Take the top 2 finishers of the grand prix, and have them play in a match of at least 8 games to decide the challenger to the champion. The top placed grand prix finisher should have draw odds, so never is a rapid game used to decide!

In many of these sports they have a system of deciding the champion whereby the champion has some legitimate claim at being the best.  In chess when you decide who gets to the final by a single tournament the system doesn't do that.  Therefore it devalues the world championship title.     

I don't think how we choose who gets to play a world championship match is irrelevant.  

Currently if we assume the players are as good as their rating would indicate, then the best player more likely than not will not play in the world championship.  Even when that rating is quite a bit above his peers as Carlsen's is.  

Carlsen making it is good for chess in the short run but it unfortanetly masks the fact that the qualifier is very flawed.   

If they did a series of candidates matches (like I and others propose) then If Carlsen was as good as his rating (and I think he is) he more likely than not would have made it to the finals.    

I still prefer tournament qualifiers, because everyone gets to play everyone else, and because issues to do with player pairings and seeding are avoided.

"Randomness" of single tournament results is not an issue over a series of tournaments. For example, I find it hard to believe Carlsen would not show his superiority over the series of 6 grand prix tournaments. And if he doesn't, then he doesn't deserve to be the challenger.

Yes, and ?

You miss the point. That one player is still more likely than others to win it.

If rating were the ultimate qualifier, why hold a world championship at all : after all, just say the one with the highest rating on date XXX is WC for that year.

1) Thats true of even Kirsan's 64 player knock out format.  Are you going to defend that?    A good format is one where the best player is most likely going to win.

2) I don't think a rating is the best method.  My statement was to assume the players are as good at their rating.  I'm not so convinced as others of the infallibility of fides current rating system - especially at the top levels.    The problem with the current qualifier format is the best players will likely not end up in the World championship final match.    This is statistically demonstrated by the article I linked.  

You cannot have it both ways.

If you insist the best player should be likely to win, either you consider that ratings are a more or less accurate measurement of skill and there is no reason you couldn't just pick the highest-rated player, or you admit the necessity of a tournament dedicated to choosing the WC and then there is no reason why the highest-rated should go straight up to the final.

The problem with 64-player knockout is not an intrinsic limitation to putting a lot of players, it's the fact that you need each of them to play only a few games at least in the first rounds. If it was physically and financially possible to make a 100-player, all rounds tournaments (ie each player plays 198 games), I would see no problem with it.

A tournament is a must. Rating can't be used to measure one's performance effectively.

A 64 players tournament is too much. A 32 players is optimal with the difference of 1st and 32th ranking to be at most 150 points. We really don't hope that a 2600 will beat a 2800 player.

It seems to me that the ratings are fairly accurate.  The best chess player is likely one of the players rated in the top 8.  But they are not so accurate where they can really sort out the top 8.  16 game Matches would be better than ratings for that.

Your point in the earlier post was that as long as the best player is more likely than the other players to win the system is ok.  That a criteria is in fact met by the 64 player ko tournaments.

I think a good system should more likely than not be won by the best player.  That is very different than your very weak criteria.  However it would be met by a series of 16 game matches.     That is why a match system would lend legitimacy to the title.

This is the crucial point where I disagree.

I do not see how you "know" that ratings give an accurate guess that the "best" player is in top 8, but maybe not in top 1, or 4, and that he is not in the range 9-16, or 17-1000 for that matter.

It's a plausible assumption, but there are other plausible assumptions that lead to more reasonable formats (relatively to time, fundraising and rating manipulation issues).

Honestly I don't even think a candidates qualifier is necessary. Chess is going somewhat in the direction of big stadium sports in America with super organization and the abyssmal formula approach. What do I mean by this? Well in math there is only one output for every input, so the answer is objective and absolute. With games like chess you can't do that, it would be better to have the old way but fide regulated where the players arrange their own matches. It worked for a hundred years but after the russians and their bullshit no one wanted to do it anymore, now we have this stupid system which is ridiculously beauracratic and mechanical. Every two years there is the potential to be a new champion, it doesn't matter if there is anyone worthy at all, there has to be a new challenger every 2 years. This confusing and annoying system leads to percieved chaos so that we get articles like this one. I don't understand why fide has to cause 100 years of chess tradition to be called into question when the could just use a system where the players arrange their own challengers as regulated by fide. Thing is, although the system is strange, we have to remember that fide rose to "power" because alekhine died without a challenger. FIDE was ofc powerless because the russians were considered the true best chess players and so elected the man widely considered the true king at the time to this day. That's key. How are we ever going to get another player like lasker who can hold the title 27 years if we don't use a longetive system. I could talk so much more about this but im done for now, just think on what i've said.

I actually agree with your overall point.   I do not necessarilly think nor can I prove that the best player in the world is rated in the top 8.  I do think its more likely true than not, but to get into specifics I would need to do more analysis along the lines that jeff sonas does in his chessmetrics site.

I would not be against having the matches start at 16 players.  Or maybe having the bottom 8 of the top 15 play a seperate series of matches and the winner could then play matches with the top 7.   There could be some variations on this.  Maybe have the bottom 8 of 14 play and the top 2 join the top 6 rated players in the major candidates matches.  

I think there is room alternatives.  But the main point is that matches have a significance beyond the rating chart.  That is why matches tend to hold far more interest of the public than tournaments.

See this chessbase article on the match:

http://www.chessbase.com/post/chennai-g1-anand-holds-carlsen-to-16-move-draw

The interest in the first game was beyond belief, and even the most optimistic chess fans and pundits had not properly anticipated it. Site after site went down due to overload. Our very own web client collapsed under the strain (though Playchess itself worked fine), and it was far from the only one. Even the official site’s video transmission, whether low resolution or high, kept freezing, and an hour into the game, the appearance of the video player changed its face, suggesting they had resorted to a backup to handle the load.

That just doesn't happen with tournaments.

Sorry about that hack editing job.  The forum seems to want to keep the chessbase quote in gray and that made it very hard to read unless I upped the font size.