My predictions:
1. Carlsen 9 points
2. Kramnik 8 points
3. Svidler 8 points
4. Aronian 7.5 points
5. Radjabov 7.5 points
6. Grischuk 6 points
7. Gelfand 5.5 points
8. Ivanchuk 4.5 points

Carlsen plays Aronian and Kramnik in the first couple of rounds (with black and white respectively)...hoping for some great chess.

Why for put Svidler top Aronian?

T-13 til blast off!

This is one of the better chess website designs I've seen, sadly

So is it being streamed anywhere?

My simulated results are still "unofficial" at this point, until Chessbase posts profiles of Carlsen, Kramnik, and Gelfand for me to adjust their draw rates and white/black differentials from, but I started doing some preliminary analysis and putting together odds on a wide variety of prop bets.

I have calculated odds for the tournament winner, for the number of decisive games in the event, the percentage of decisive games that are won by white versus black, the number of players in first place after the final round (ie is there a decisive winner or is there a tie, and if a tie, how many people sharing first), plus an over/under points scored for each player.

I also have calculated out some props for the first round, including odds on each game (of white win, black win, or draw), and the number of decisive games, plus a parlay of those two concepts with each player's odds of being the sole leader at the end of the round.

My question, then, is are enough people interested in seeing these numbers that I should run a prop betting contest for this tournament? Presumably in a separate thread, post round-by-round odds on various things like that, give everyone some amount of play money to start with, and see who can run it up the highest by the end of the event?

I'd enjoy seeing your odds/math but am not too interested in betting.

Not everything can go according to elos.

Oh man I got all excited thinking the first games were today only to see it's an opening ceremony. Blah.

Ditto.

Sorry, I should have said in my post that the 14th wasn't start of play.

Through this thread we got the links to offical site http://www.worldchess.com/candidates/index.html and site for live streaming http://london2013.fide.com/en/compon...137-live-video and roundtower has edited them into the op so people can find them.

can't wait.

T-11 til blastoff, real blastoff this time.

And from the opening ceremony:



Magnus sporting the 5-oclock tumble in the mud.



Somehow this is very close to what I think of when I hear the name Dr. Jekyll.

So long as Mr. Hyde shows up to the board, he can be victorious!

All right, with 6 of 8 profiles published on Chessbase, I have my player tendencies as good as I can get them - the only data I'm missing is games between Carlsen and Gelfand, which will probably have minimal impact if any. So I simulated 200,000 runs of the tournament today. Here are the odds for the event, based on those results:

TOURNAMENT WINNER:
Magnus Carlsen 63.6% (1.57 to 1)
Levon Aronian 14.1% (7.09 to 1)
Vladimir Kramnik 9.5% (10.58 to 1)
Teimour Radjabov 7.2% (13.82 to 1)
Alexander Grischuk 2.2% (45.24 to 1)
Vassily Ivanchuk 1.9% (51.95 to 1)
Peter Svidler 0.8% (125.91 to 1)
Boris Gelfand 0.7% (144.76 to 1)

EXPECTED POINTS SCORED:
Magnus Carlsen 8.85
Levon Aronian 7.53
Vladimir Kramnik 7.36
Teimour Radjabov 7.19
Alexander Grischuk 6.56
Vassily Ivanchuk 6.39
Peter Svidler 6.12
Boris Gelfand 5.99

NUMBER OF GAMES DRAWN (out of 56 total games, excluding tie-breaks):
Average of 38.2

PERCENTAGE OF DECISIVE GAMES WON BY BLACK:
Average of 34.2%

AVERAGE TOURNAMENT FOR THE WHITE PIECES: +11.7 =38.2 -6.1

NUMBER OF PLAYERS IN (OR TIED FOR) FIRST PLACE AFTER RD 14:
1 (Sole winner) 81.7%
2-way tie 15.0%
3-way tie 2.8%
4-way tie 0.5%
5-way tie 0.1%
6-way tie 1 in 33,333 chance
7-way tie LOL not happening
8-way tie 1 in 200,000 chance

And then the odds for specific results of round 1:

ARONIAN (W) vs. CARLSEN (B):
Aronian Wins 11.1%
Carlsen Wins 19.4%
Draw 69.5%

GELFAND (W) vs. RADJABOV (B):
Gelfand Wins 12.0%
Radjabov Wins 20.5%
Draw 67.5%

IVANCHUK (W) vs. GRISCHUK (B):
Ivanchuk Wins 21.1%
Grischuk Wins 11.4%
Draw 67.5%

SVIDLER (W) vs. KRAMNIK (B):
Svidler Wins 11.6%
Kramnik Wins 13.9%
Draw 74.5%

NUMBER OF DECISIVE GAMES:
4 0.8%
3 7.7%
2 26.8%
1 41.1%
0 23.6%

SOLE LEADER AFTER RD 1:
Carlsen 6.7%
Aronian 3.8%
Kramnik 4.4%
Radjabov 7.3%
Grischuk 4.1%
Ivanchuk 7.5%
Svidler 3.8%
Gelfand 4.2%
2 or more players tied 58.3%

great post

-my lowest simulation had carlsen at 68%. but i was boosting carlsens rating in every simulation due to it steadily increasing. leaving it the same would have been in the mid 60s. so im on board with those %

-most of the other numbers look kinda accurate, with the exception of black wins. i think these formulas dont take into account how often black is playing for a draw and not gonna win at the super GM level. i would be shocked if over 1/3 the wins were by black and if carlsen is more likely to beat aronian with black than he is to lose

It's all pure ELO. The white pieces are worth an average of roughly 40 ELO points (I have reams of evidence to support this). Since Carlsen is rated MORE than 40 ELO points ahead of everyone else in the field, he will of course be favored to score >50% with black, even against his highest rated opponent (Kramnik or Aronian, depending on whether you use live ratings or the March rating list). You can't have an expected score higher than 50% unless you're more likely to win than lose, so Q.E.D.

As for the draws, historically when these eight players compete against each other at classical time controls, only Carlsen and Kramnik draw more than 70% of their games with the black pieces. Now, it's entirely possible, even likely, that in this format and with a World Title Shot at stake, the field will collectively be more tentative than normal, and draws will be more common. I would bet the "over" on total draws for the event, if I were putting money down. However when you model the simulation off of historical draw rates between these players, this is what you get. I definitely do believe that of the decisive games, black is a (slight, but definitive) favorite to garner 33% or more of those wins.

I'm really interested to see how this goes. If this were another knockout I'd 100% agree with you as losing a single game is practically the same as losing them all so players naturally play not to lose. But in this event the only way to win is to accumulate the most points. I expect this to lead to some exciting play as, especially in games involving the lower half of the table, the players are absolutely going to be looking to pick up points. And given that this is a candidates event the difference between 1st and 2nd is in many ways greater than the difference between 2nd and 8th.

Thanks for including the links! I can't wait either, I'm actually gonna wake up at 5:30am to watch the 1st round from the beginning

It will definitely be fascinating to see how this turns out. My prediction of more draws than "normal" is psychological, not strategic. I absolutely agree with you that strategically, the best tactic is to play a bit aggressively, especially against the lower rated players in the field (or against anyone who shows signs of being "out of form") - though of course there aren't exactly any patzers or pushovers in this field. Shocking, I know, someone on 2+2 touting aggression

However I don't have faith in these players, as a group, to play optimally in that regard. In particular, I expect the leaders to play cautiously, and try to "protect their lead", and not "fall behind". And in round 1, everyone's a leader until they start losing, or until someone else starts winning. I expect relatively drawish opening rounds, essentially a feeling out period, with the prepared novelties being held back for later "must win" rounds. I hope I'm wrong, though.

Great analysis BJJ. Excited for the tournament, slightly afraid there will be a lot of short draws (except for Magnus, who doesn't roll like that)

I can't wait for it to start. I will try to watch as many rounds as possible but that's hard during work...

I am hoping for Ivanchuk to win it all. I think he has deserved a shot at the title. As that probably won't happen, I also support Kramnik.

GO CHUCKY!

Is the 40 point rating boost for White based on analysis across all ratings or just among the players in this field?

I think it's probably more than 40 points at the very highest levels, while close to insignificant at ratings below ~2000.

Either way, it probably doesn't make a big difference to your model since all players play each other with both colours. The only effect would be on the volatility of the results.

Specifically the players in this field, and 40 is where most but not all of the players are at. Against the other 7 candidates, Radjabov has almost the same score with white as with black, I only have his "white advantage" at 7 points, while Kramnik does the vast majority of his damage with white, rating a white advantage of 56 (and a black draw rate of 78%). Those are the only two players not in the range of about 35-42, though, and 40 +/- 1 is the most common. In any given game, I average the white advantage for each player involved, so in both Kramnik/Radjabov games, white will get a 31.5 boost. In a game between Kramnik and Gelfand (who was at exactly 40, I believe), that boost would be 48 instead.

Those numbers are my best recollection, I don't have the spreadsheet in front of me as I type this post, but the concept is accurate either way. All the numbers come from the results described in the Chessbase series of profiles of the Candidates.

Whoever did the webdesign for the official site (http://candidates2013.worldchess.com) needs to be brought up on charges. You couldn't find an uglier board/piece set combination for the live game viewer if you tried.