01 Anand 2793,1 -5,9 (2) 15 1969
02 Kramnik 2788,0 -11 (1) 13 1975
03 Morozevich 2774,1 +9,1 (1) 11 1977
04 Topalov 2767,6 -12,4 (1) 13 1975
05 Leko 2755,9 +2,9 (1) 13 1979
06 Shirov 2755,0 0 (0) 0 1972
07 Aronian 2752,9 +13,9 (2) 14 1982
08 Mamedyarov 2751,6 -8,4 (1) 13 1985
09 Carlsen 2749,9 +16,9 (1) 13 1990
10 Svidler 2747,0 -16 (2) 16 1976
11 Radjabov 2746,3 +11,3 (1) 13 1987
12 Ivanchuk 2742,3 -8,7 (3) 22 1969
13 Karjakin 2732,0 0 (0) 0 1990
14 Adams 2729,2 +3,2 (1) 13 1971
15 Kamsky 2726,0 0 (0) 0 1974
16 Gelfand 2723,0 -14 (1) 13 1968
17 Ponomariov 2719,0 0 (0) 0 1983
18 Grischuk 2716,3 +5,3 (1) 11 1983
19 Jakovenko 2711,2 -8,8 (1) 11 1983
20 Alekseev 2711,0 0 (0) 0 1985
21 Polgar 2708,8 +1,8 (1) 13 1976
22 Bacrot 2708,0 +8 (2) 18 1983
23 Bu 2707,9 +16,9 (2) 16 1985
24 Akopian 2700,0 0 (0) 0 1971
25 Movsesian 2698,7 +21,7 (3) 22 1978

On the other hand so promising Navara continues his 2400-like league season performance by losing to Ftacnik and goes under 2650. He should be nominated to Grand Prix thanks to tournament in Karlovy Vary. If this will not help this guy, nothing will.

exactly my thoughts.

while bu lost the blitz tie-break, he was clearly the best player in that tournament imo. naka caught him by "riding the swiss wave" as we say (at least in norway), and alas naka's performance, while still good, isn't quite comparable. i've noticed that nakamura more than once have taken a loss against some "weaker" player early on in swiss tournaments, to bounce back with a string of victories - like he did here. for tournament success it doesn't matter too much how you score your points, but for rating purposes, you usually want to play the stronger players in the tournament.

nakamura only played one or two players approximately at his own level during the tournament - and he did so in the last two rounds. after round 8, naka was 6/8, but ratingwise he was just "break even", while bu (at 7,5/8) was +23 rating points at the moment. i'm sure you know how rating is calculated, but let's just compare the opponents (their ratings) of bu and naka:

bu played 5 (6 if we let 2599 pass as) 2600-players.

naka played 2.

if we ignore the two players both of them met, rated 2569 and 2638 (bu had 2 blacks, naka 2 whites, though), it looks like if they played tournaments of completely different strength:

naka - bu
2378 - 2420
2416 - 2524
2455 - 2553
2487 - 2599
2491 - 2635
2500 - 2662
2530 - 2663
2606 - 2680

naka did well to beat petrosian (2606) with black in round 9, and efimenko (2638) with white in round 10, but in a sense naka somewhat "wasted" this opportunity to play a number of strong opponents that aren't so readily available in the us as in european tournaments. (he did so by losing to underrated zhao (2487) in round 5). naka's opposition actually looks quite similar to the kind he typically has to face in us swisses.

i'd like to see nakamura in a consistently strong field like corus b or something next year - i'd be happy if you could falsify or confirm the rumour that he turned down a group b invitation this year. if he doesn't realize it himself, then hopefully his fans might see that he needs to prove himself against something better than a 2507-average like he met in gibtelecom this year. corus b 2008 had an average rating of 2618 (category 15) - while 2507 is category 11. worlds apart, really.

movsesian is _truely_ above 2700 for the first time!

01 Anand 2793,1 -5,9 (2) 15 1969
02 Kramnik 2788,0 -11 (1) 13 1975
03 Morozevich 2774,1 +9,1 (1) 11 1977
04 Topalov 2767,6 -12,4 (1) 13 1975
05 Leko 2755,9 +2,9 (1) 13 1979
06 Shirov 2755,0 0 (0) 0 1972
07 Aronian 2752,9 +13,9 (2) 14 1982
08 Mamedyarov 2751,6 -8,4 (1) 13 1985
09 Carlsen 2749,9 +16,9 (1) 13 1990
10 Svidler 2747,0 -16 (2) 16 1976
11 Radjabov 2746,3 +11,3 (1) 13 1987
12 Ivanchuk 2742,3 -8,7 (3) 22 1969
13 Karjakin 2732,0 0 (0) 0 1990
14 Adams 2729,2 +3,2 (1) 13 1971
15 Kamsky 2726,0 0 (0) 0 1974
16 Gelfand 2723,0 -14 (1) 13 1968
17 Ponomariov 2719,0 0 (0) 0 1983
18 Grischuk 2716,3 +5,3 (1) 11 1983
19 Jakovenko 2711,2 -8,8 (1) 11 1983
20 Alekseev 2711,0 0 (0) 0 1985
21 Polgar 2708,8 +1,8 (1) 13 1976
22 Bacrot 2708,0 +8 (2) 18 1983
23 Bu 2707,9 +16,9 (2) 16 1985
24 Movsesian 2702,7 +25,7 (4) 23 1978
25 Akopian 2700,0 0 (0) 0 1971

i certainly do think the fide ratings are capable of predicting results, even if the accuracy might be even better with other systems (sonas claims that his formulas are better than fide's, but even that claim is debatable, for reasons i'm not willing to discuss at the moment :o)

however, i don't think it's within my capacity to maintain a list of running last-n games performance ratings. of course it could be done for only a limited number of players - say the ones currently in my top list - but to be complete and really list the players with the best recent run of tournament games, the list would need to include a rather notable number of 2600-players too, capable of the occasional good result. or at least it would be necessary to calculate their performances in the latest n games to really know.

also, you would need to decide what to do with league games. for instance, currently i don't register the dates of each and every game played, not even fide cares about that, i think (only about start and end date of the event to be rated, as far as i know). also, it could be debated whether it was games played during the last 3 months, or the latest 30 games, or what other possible measure one should use - like you probable are aware of, the tournament activity varies a lot for each player over the year, and from player to player in terms of total games played, and so on.

for most players in my system, i've got results from roughly 1,67 rating periods at the moment - the last quarter of 2007 (sep,oct,nov) and 2/3 of the first quarter of 2008 (dec,jan), but in terms of games, this varies a lot, from 18 (anand) to 72 (wang). most players have played more than 30 games, though, but some of the strongest players are among those who don't play too much (anand 18 games, kramnik/leko/gelfand 22 games, polgar 23 games, topalov 27 games...)

also, stefan fischl already does this on a yearly basis, see http://members.aon.at/sfischl/cl200... or http://members.aon.at/sfischl/stat.... for more of his performance statistics. he's made a few graphs for latest 12-month performance for a few selected players, too. but he does no "weighting" based on number of games - personally, i think it might be "easier" to have a 2800+ performance over a 20-25 games than over 50 games, for instance. especially when you know your opponents ahead of time, and can prepare for all of them (unlike in swiss and team events).

sonas has incorporated an element game-count weighting in his performance rating, but only on a per-event basis. if this is intuitively or statistically motivated i don't know (or don't currently remember). anyway, the biggest problem with chessmetrics is probably his very skewed and incomplete data, not mainly his methods.

[kasparov typically prepared well and played few games for all the 5-6 latest years of his active career (except maybe in 2001):

2000: 35 games
2001: 48 games
2002: 16 games
2003: 21 games
2004: 24 games
2005: 28 games

according to fischl's lists for 2006 and 2007, these are the numbers:

2007:
anand 55
kramnik 44
topalov 51
leko 53
gelfand 49
radjabov 48

2006:
anand 37
kramnik 27
topalov 58
leko 43
gelfand 62 (due to candidates :o)
radjabov 49

for reference, here are the numbers for some other (current) top players, who are less picky about which tournaments they play (those above seems to almost exclusively play supergms, matches and the occasional team event :o)

2007:
morozevich 74
ivanchuk 81
mamedyarov 61
svidler 91
aronian 71
shirov 96
karjakin 84
carlsen 100

I cant really tell anyone how to spend their vacation time however if i (as a dedicated chessplayer) was to visit Moscow i would try to find out where the players met so that "i could get my game on": at least for a few hours. :)

<
2000: 35 games
2001: 48 games
2002: 16 games
2003: 21 games
2004: 24 games
2005: 28 games
>

was based on when games were rated, not played. going by fischl's stats, this is the comparable one to those numbers above - a list based on when games were actually played:

2005: 12 (kramnik 49, svidler 97, moro 52)
2004: 34 (kramnik 52, svidler 61, moro 40)
2003: 18 (kramnik 38, svidler 80, moro 47)
2002: 21 (kramnik 0, svidler 32, moro 47)
2001: 37 (kramnik 37, svidler 55, moro 46)
2000: 49 (kramnik 46, svidler 63, moro 54)

it does look like there are more tournaments now than in 2000-2002. the higher number for kasparov (and kramnik) in 2000, was of course due to their 15-game match. other than that, kasparov typically made 2-3 appearances in classical chess in his last 5-6 active years:

2005: linares
2004: rus superfinal, ecc, petrosian mem, linares
2003: ecc, linares
2002: olympiad, linares
2001: corus, linares, astana
2000: wc match, sarajevo, frankfurt, corus, linares (wow!)