They Call Him MISTER Carlsen

I am sure the Anand-Topalov match will take place. But probably the prize money will be less than forseen 1 year ago.

Politicians and other powerful Bulgarians are very keen on supporting Topalov in becoming World Champion. Topa has heavy backing in his home country, he is a ''national hero''. Eventually the match will be played in Sophia, similar to the Topalov-Kamsky match.

I guess at least Topalov won't be happy with a =9 -1 score in the end ... among other things, he cares [a lot] about his #1 spot in the (live) ratings. So he will probably put up a fight at least in his second (white) game against Carlsen!? And even as far as the other players are concerned: What's the difference in prize money between 2nd and 6th(last) place?

"Chinese don't like the US$"
Maybe, but another reason could be that three of the other Grand Slam tournaments (Corus, Linares, Bilbao final) take place in the Euro zone, and Bulgaria (MTel) also wants to join eventually.
In any case, your post confirms that "the rest of the field" still has something to play for and care about in the second half of the tournament - maybe the difference between 55K and 20K is relevant, so are ELO points to win or rather (everyone but Wang Yue) to limit the damage.

"Magnus will for sure play just as hard in the 2nd part."
Yeah, in the past he was fighting hard when he still had competition for tournament victory until the end - the most recent example is probably his last-round loss against Shirov at MTel 2009. But having a commanding lead and "nothing to fear" after the first half is another situation.
Are there recent examples of a player continuing at full speed till the end? In all cases coming to my mind, the early leader DID slow down and coast to victory:
- At the San Luis WCh, Topalov started with 6.5/7, followed by seven draws in the second half
- At MTel 2008, Ivanchuk had 5/5 followed by 3/5
- At San Sebastian, Nakamura slowed down and, in the end, had to rely on his blitz skills to secure first place.
Not criticizing anyone - maybe Nakamura because clear first would have been (even) more convincing than first on tiebreak - just compiling empirical evidence.

In Carlsen's case, it depends also on his opponents. If Leko and Topalov don't take risks with white (not necessarily the same as playing for a draw from move 1), they will be hard to beat.

"Magnus will for sure play just as hard in the 2nd part."

@Bobby F

Nope, that's not sure at all. How wise is it to fight hard in an equal position, when you are in such a lead?

This is about securing the money now, and not taking any unnecessary risks.

Two very tough black games against Leko and Topalov next. Full throttle there, depending on how that goes, it might be very wise of Carlsen to slow down against Wang Yue, who will drain lots of energy before going down anyway.

If Carlsen go on full throttle, he risk running out of steam and start blundering. If that happen, Topa is more than capable of catching up in the second leg and grabbing the first place.

It's rather likely, that Jako can get a rather short GM draw in the last round too.

This is obviously true. It's easier to have a bad run than a good run. The Elo curve should not possess symmetry. So matches, where your opponent is fixated, gives you an bonus chance of facing someone on the lam.

In fact, more than half of his "Chessmetrics Performance Rating" top 10 numbers are derived using matches (two of Fischer, 4 Championships).

Laughy that Karpov gets 2851 (#21) for losing Lyon. Garbage in, Garbage out, particularly with no error bars (don't real stats guys use these up the wazoo?)

Zarg: Yes,your description is more acccurate. I can add that foolhardy he will not play. That would for sure be, ehh, foolish.

Anyway, I still expect strong and entertaining play from Magnus the 2nd half. In 19 hours we will know...

''Anything can happen'' at this high level. Probably the end table will look different next Saturday, than it does today.

Um... are you nuts? Remember that "chuckers" like Kamsky had to BEAT Carlsen on his way to qualifying for his real match with Anand (which at the last second was changed to Topalov). So if Kamsky is a nobody then I guess you're saying so is Carlsen. Honestly the nerve of some people to call a strong GM "weak" ... where do you come off saying stuff like this? Whats your ELO again sir?

"It was like watching two guys playing Rock, Paper, Scissors both throw "rock" 29 times in a row."

Oh man classic Mig. Thanks for that, I cracked up laughing!

Anand is not foolish enough to play in Sofia. He doesn't need the money that badly either. If the match takes place, and I fear it won't, neither Bulgaria nor India will be its venue.

Nanjing 2008 may not qualify as a "comeback" (from behind), but Topalov also had a slow start (four draws) followed by a strong finish (+4 =2 in the remaining rounds). There was also MTel 2007 (2/5 followed by 3.5/5 - yes, an overall +1 score was enough to win the tournament).
With this history in mind, I would still consider him the favorite ... for second place at Nanjing 2009, and give him reasonably good odds to defend his #1 spot on the live rating list for the time being

I believe frogbert is very correct in this case, don´t let his pedant manners fool you.

d_tal

I'm sorry for the overly annoyed tone in the post you responded to - and obviously I can't know the chain of event that lead up to the Chessbase article. It's my guess - nothing more.

Regarding the analysis, I've done a lot of them already, using techniques that are available to programmers, in order to produce better and more significant data from the same original data. I've also published several results on chessgames.com.

My frustration is mostly mine - I'm "annoyed" that what I consider wrong or misleading information is being spread and touted, and I'm pesonally annoyed that I don't have enough time to convince "the right people" about the flaws I mean I'm able to point out. But that's my problem in a sense, more than anybody else's.

I also think how FIDE put together an "expert group" for discussing the K-value issue was quite nonsenscial - essentially supplementing people in the QC with the people that Chessbase brought forward in their series of articles: Macieja, Nunn and Sonas most notably.

Again, personally I wouldn't have time to be involved in any official FIDE work regarding ratings, and in particular the meetings that were held in Greek (or wherever it was this time) were completely impossible for me to participate in. And obviously there's no reason FIDE should consider me an "expert" either.

However, if they cared about getting a thorough analysis of their data, then FIDE simply should make their information available to _everyone_ - at least everyone displaying an honest interest in researching the issues in questions, whether it would be changes to the K, or analysis of inflation, or other possible modifications that might be warranted to the FIDE rating system.

The chess community has a lot of talented, educated people capable of analysing these data, but as usual FIDE chooses to operate "behind closed doors". This pervasive preference of non-transparent processes in the FIDE system will continue to annoy me, regardless of which part of their operation it's reflected in.

But my own lack of time is, again, my problem - and I shouldn't take that out on anybody else, of course.

"Most important in the given context: Different from what frogbert suggests, Jeff Sonas made the rather responsible statement that five games are not enough to make a meaningful statement."

Thomas, I agree that I wasn't clear enough in this respect - my points regarding the analysis of only 5 games were directed at Chessbase, not at Jeff Sonas.

The idea of putting 5 games in a "historical context" also origins at Chessbase offices, and the criticism is hence directed at them for that, not at Sonas.

Like raindeer and Stat-boi pointed out, there are a number of issues with how performances are compared in Sonas' system. alexmagnus on chessgames.com has pointed to several other weaknesses with Sonas' "Performance Rating" definitions too.

And the entire idea of making "historical comparisons" of ratings and rating performances is basically broken, for obvious reasons. Sonas does the circus act of

1) telling people that ratings can't be compared over different eras

2) making elaborate age-lists, where players are ranked & compared between different areas

3) making historical lists like "top 100" performances over multiple areas

Which talking head are people supposed to believe in? The one that says 1) or the one that does 2 and 3?

Thanks, I had not thought about Nanjing. Sofia I didn't count since it wasn't all that special. You may count Nanjing but it was more a spectacular run than a "comeback": he was just half a point behind Aronian and Bu after those four rounds, and caught up with both in the very next round, never to let go of his lead even though Bu kept sharing it for two more rounds.

Not quite Sofia 2006, for instance, where he was at -1 after 6 rounds and counted out by many just to win the last four. If he does _that_ again he could catch Carlsen with some luck.

Carlsen is only 12 points behind Topalov in the live ratings. I can't wait for the next World Championship cycle.

"A good comparison would be Ivanchuk's 4.5/5 against Yudasin (a top-ten player) in their 1991 match. If Carlsen maintains the 90% score to the end (i.e. scoring 9/10) then I would call it the best performance ever, a little above Karpov Linares 1994 and Fischer's 1971 match against Larsen. If Carlsen "only" scores 85% out of 10 games, then I would call it an all-time top-ten performance, roughly comparable to Kasparov at Linares 1999 or Kramnik in his World Championship match against Kasparov." Jeff Sonas at Chessbase

I found this interesting.

In fairness also to Chessbase, the reason for contacting Sonas - as stated in the article - was:
"A number of readers have sent in questions or speculated on whether this is the most impressive performance ever seen in the game."

Maybe these were overly enthusiastic Carlsen fans? In any case, while some other Chessbase pieces promoting their commercial products (of course nothing a priori wrong with that) may not really be "on popular demand" but rather unsolicited 'advertorials', in this case I think those (Friedel's?) words can be taken at face value. One can still criticize Chessbase for an overly zealous header ("an interesting historical assessment of Carlsen's performance"), not quite [or not yet] consistent with Sonas' actual words.

"The near future might tell" - but I don't think Carlsen is (at least he shouldn't be) affected or "obsessed" by considerations such as: "If I score 4/5 in the second half, it will be an all-time top10 performance _sensu Sonas_, 4.5/5 would mean BEST PERFORMANCE EVER" ,:)

Concerning the Sonas lists themselves: One may disagree with his long-term comparisons, but even then it represents IMO an interesting compilation of remarkable results. Just as one example, Bobby Fischer's 6-0 wipeouts in matches were only hearsay to me (I was 4 years old in 1971 and not yet playing chess ...) - I still didn't need Sonas to get informed, but younger people might. And his homepage has the option to "zoom in" on certain decades, when results may be internally consistent and meaningfully comparable with each other.

"but sniping just because Sonas has achieved publicity seems churlish."

mishanp, whatever you or d_tal (or anybody else) might think, that's not what I'm doing. I've got all the publicity I want and more. [I could extend my site with articles about rating, I could write regular articles about ratings at Chessdom, and so on. Yesterday my site had more than 15000 hits (thanks to Carlsen, mostly) - and over 1,5 years I've had an average of nearly 4000 hits daily. I don't depend on Chessbase to get publicity.] But I'm short on time for everything I'd liked to do...

So, it's got NOTHING to do with publicity (and for the record, Chessbase published also my reaction as an "expert opinion" during the K-series of articles.) My bief is that I'd like people to be educated and knowledgable about ratings, and I hate stuff that distort a meaningful picture. When a significant player like Chessbase misguides people, like talking about the 3100+ "rating performance" of a 2500-player that has beaten 4 other equally rated players, it's frustrating, to say the least.

Unfortunately Jeff Sonas does more to deceive the general public than to enlighten them, too, in my probably not so very humble opinion. You disagree about that, and that's fine.

And FIDE doesn't know how to go about too many things, whether it's supervising their rating system or it's organizing a World Championship between the world number 1 and number 2.

"The criticism of Chessbase is if anything even more misplaced."

That depends entirely on your agenda and your point of view. To me, it's very, very meaningful to criticise Chessbase for miseducating people! :o)

I also criticise mainstream papers when they distort a scientific issue for populist reasons. Do you think a commercial player is morally free to do anything, including misguiding the public, simply because its primary mission is to entertain and/or make money?

Well, I don't. :o)

Lol

I have a lot to say about various aspects of chessmetrics, but actually in terms of data, he doesn't do too bad for "old & ancient" events.

There's no point in diving into details about Chessmetrics, but its biggest problem lies in its lack of relevant recent data, especially for young players or players not playing international events with recorded game sheets.

"And his homepage has the option to "zoom in" on certain decades, when results may be internally consistent and meaningfully comparable with each other."

Thomas, do you really think so?

For instance, all the age-lists for young players, 10 to 17-18 are basically broken, even for the last 20-25 years! You can't even compare Radjabov, Ponomariov, Karjakin and Carlsen at similar age (at least not before they're 15-16+) - simply because Sonas' data are heavily incomplete, and even his system typically counts lots of active FMs and IMs as "inactive" (and hence unrated). Hence, his ratings suffer from bad data quite consistently for any players that mostly face non-GMs.

One quick example: When Carlsen played Aeroflot open in 2004, Chessmetrics rates 6 of 9 games, even though Carlsen only faced GMs or high-rated 2500s and some 2600s in all his 9 games.

The majority of players mostly playing nationally, even at the highest level, end up being unrepresented (somewhat depending on local/national variations in getting game data published on TWIC or Chessbase) - the consequence being that the Chessmetrics numbers for anyone below 2500-2600 end up being decided as much by which results are calculated as which results a player (actually) had.

For the best players in the world, this doesn't matter too much - but for the best players in the world Chessmetrics are always outdated, currently by 4 years.

"he explains his methodology clearly"

And the methodology has been _revised_ for both major rounds of publishing, both in 2002 and in 2005 - in effect allowing Sonas to tweak his formulas for potential unwanted effects of previous choices. Or hypothetically to suit a given agenda.

If Sonas were to publish calculations for 2005-2009, I sincerely hope that he would exclusively update the data and leave everything else the way it was in 2005. Or at the minimum, present the results both in terms of his "old model" and an "updated model".

Changing the model every time you get some new data is ok under some circumstances, but here it would rob the public of seeing what effects the Chessmetrics model anno 2005 would have on newer data. And I think that would've been a shame.

"And I think the numbers he provides are interesting, his matter-of-fact style makes a good laugh, Chessbase did well to ask him."

Heh, well. :o)

My "problem" is probably that the time I've already spent on this makes it a little hard for me to take "very lightly" on the subject.

For instance, a number of people think it makes sense to say that there has been a yearly inflation of ca. 5 points per year for the last 10 years. And that this applies to the ENTIRE rating scale.

Does anyone here feel that this is wrong or right?

Let's assume for a moment that this is true - then we can subtract around 45-50 points from every current elite player to compare them to the top players in 2000.

In July 2000 there were 5 players above 2750 - in addition to Kasparov, also Kramnik, Anand, Adams and Morozevich.

Now, if we "adjust for inflation", in 2009 - nearly 10 years later - we would have one (1) player above 2750, using the scale of 2000 - and that one player would be Topalov.

Put differently, according to such a straight-forward application of the linear inflation-adjustment, Adams and Morozevich of 2000 were higher rated (and hence, better players*) than today's Anand, Carlsen, Kramnik, Aronian, Radjabov and Leko, for starters.

How many think Adams and Morozevich in 2000 were better players than Anand, Carlsen, Kramnik, Aronian, Radjabov and Leko are today, 2009? I, for one, do not.

But this is the kind of nonsense (sorry) that follows from the "inflation" definition that Sonas presented in Chessbase not too long ago.

---

*) why else would we "adjust for inflation" if not to compare actual skills? If we're happy with ratings as a relative measure comparing active, current players, we wouldn't care a bit about a slight drift in the system, would we?

"Do you think a commercial player is morally free to do anything, including misguiding the public, simply because its primary mission is to entertain and/or make money?"

No* but the crux of the matter is that your attack on Sonas & Chessbase was weirdly excessive given this particular case. Sonas explains how the performance would be rated in his system if Carlsen could somehow finish with 9/10. It's hardly "dumb and dumbening" or misguiding/deceiving the public. You describe Sonas as choosing "more personal exposure/attention" rather than "behaving like a responsible statistician"!? It's not as if he put his site/system together over a rainy weekend!

You also attack him by describing him as first, "telling people that ratings can't be compared over different eras," but where does he actually say that? Or do you simply mean that he said that the ratings can't be compared without some adjustments/difficulties, in which case there's no contradiction.

* though I'm prepared to cut some slack & generally believe in human fallibility rather than deception when it comes to anything as slippery as statistics

"[I wrote] And his homepage has the option to "zoom in" on certain decades, when results may be internally consistent and meaningfully comparable with each other."

"[frogbert] Thomas, do you really think so?"

My statement just referred to "best tournament (and match) performances", and the 'may be' in my sentence implies that I am not completely sure. Only after your reply I realized that there is much more on Sonas' site ... . To his credit, he mentions the deficiencies and caveats you pointed out (it takes a bit of mouse-clicking, but the information is there). He also asks "everyone" to help him fill the gaps and inconsistencies. And you have to admit that his project is considerably more ambitious than the live rating list, while both are hobby projects (Jeff Sonas also mentions "a daytime job").
You may consider it "immoral" that Chessbase makes publicity for the Sonas project, but actually what is at stake? Does it cost Carlsen any tournament invitations or sponsor opportunities if Chessmetrics does not correctly reflect his playing strength at the age of 14? Altogether, along with Bartleby I think Sonas' numbers are "interesting" - even if details are wrong. But I do not put excessive significance on any ratings, be it Sonas, live or FIDE official ... .

"You also attack him by describing him as first, "telling people that ratings can't be compared over different eras," but where does he actually say that? Or do you simply mean that he said that the ratings can't be compared without some adjustments/difficulties, in which case there's no contradiction."

He does actually say that - and that "inflation adjustments" he makes doesn't change that. It's not an unreasonable attack to hold someone against what he/she has said.

"And you have to admit that his project is considerably more ambitious than the live rating list, while both are hobby projects "

You think Chessmetrics has been a hobby project that Sonas did all in his spare time, Thomas?

I think you need to click around a bit more. :o) The time I've spent on the (pure) hobby project that live ratings are, is quite irrelevant, btw.

"your attack on Sonas & Chessbase was weirdly excessive given this particular case"

Like I've explained above, the context is much bigger/broader. But I've already said I'm sorry for the grumpy tone in my first post (you got that, didn't you?) That doesn't mean that my opinions have changed.

Wow, that must be the best performance ever.

I need to correct myself a bit:

"He does actually say that - and that "inflation adjustments" he makes doesn't change that."

Actually Sonas did say that he's made ratings comparable over different eras - but it's very important to read his "clarification" of what he means by saying that:

"A rating of 2700 or 2500 should mean approximately the same thing in 2001 that it did in 1901. [...] This correction enables the comparison of ratings across eras. Of course, a rating always indicates the level of dominance of a particular player against contemporary peers; it says nothing about whether the player is stronger/weaker in their actual technical chess skill than a player far removed from them in time."

So, after Sonas has "corrected for inflation", comparing player A with a 2500 rating in 1901 with a player B in 2001, also with a 2500 rating, says NOTHING about whether player A or B is stronger, or whether they are equally strong, in terms of "actual technical chess skill".

What do we think an average person thinks when Sonas says that the ratings are comparable? But that equal ratings doesn't mean equal skills?

So, what does this "adjusted" rating tell us then?

"a rating always indicates the level of dominance of a particular player against contemporary peers"

Does anyone want to explain to me what "level of dominance" a rating of 2500 means? Without relating it to how many players are rated higher or lower than that? At any point in time?

2500-level of dominance, 2600-level of dominance

The 2001 player's skills might have been way beyond that of the 1901 player's skills, but they have been adjusted to having the same 2500 rating, so they had the same "level of dominance against their peers".

Put differently, Kramnik's 2800 rating in chessmetrics for 2000 might indicate better "technical chess skill" than Fischer's 2895 (chessmetrics anno 1971) rating - but Fischer's 2895 indicates higher "level of dominance". NOT that he was a stronger chess player in terms of "actual technical chess skill".

Now, how many do we honestly think are able to interprete Sonas' adjusted ratings that way? And again, can anyone tell me that they honestly have a feeling of what a 2500 "level of dominance" means?

Good luck!

I assume those ratings say nothing about "actual technical chess skill" either, similar to the adjusted Chessmetrics ratings...

No, you are wrong. According to Jeff Sonas, it was 2899 by Anatoly Karpov, Linares 1994.

I thought 3191 was better than 2899.

It's based on only 4 games.

Not that it's relevant, but using Sonas' formula for performance ratings on FIDE ratings, and Chessbase (raw) performance rating (based on a formula, not FIDE's tables) the figures become these:

Average of opponents: 2416
Chessbase performance: 3191 (avg + 775)

(FIDE uses +736 for 100% even if it normally is considered to be undefined)

Chessmetrics performance:

((3191 * 4) + (2416 * 4) + (2300 * 3)) / (4+7) + 43 = 2709

And 2709 << 2899

Sonas calls this "padding" - he adds 7 hypothetical games to the performance, 4 draws against the opponents you played and 3 draws against 2300-players, but since the latter will typically drag the performance down for higher-rated players, he's kind and adds some 43 points in the end.

It's partly his way of being able to assign a performance rating to Fischer's 6/6 match victories, partly a way to say that high performances over many games are better than over few games, and partly a way to encourage activity, since the same technique is used for the ratings themselves. :o)

"Also, yes, the entire Chessmetrics website was indeed completely done in my spare time, for zero money, almost exclusively late at night because of my regular day job."

I'm sorry about that misunderstanding on my part. I probably mixed up your background with that of Glickman (creator of Glicko), who did some of his rating work as part of his thesis (or other university) work.

"Thanks to all the people who are supportive, or at least tolerant, of my work, and I also appreciate hearing from people who clearly have put a lot of thought into the topic and read my website carefully. Even if you disapprove..."

I'm neither intolerant of your work, nor do I disapprove of your website which I link to on my site, but I disagree with some of your definitions and several of your supporting arguments. And I seriously don't understand how an inflation-adjusted "2500-level of dominance" should be interpreted intuitively.

If you don't mind and would be interested, I might send you a private mail both with my own calculations, counter-arguments and improvements that could be done for the rating data.

Cheers.

Shameful sexist rating list! This is the correct eval, by the way, limiting it to the top five:

Lilit Mrktchian: 9/10
Shen Yang: 7/10
Marie Sebag: 6/10
Betul Cemre Yildiz: 6/10
Zhu Chen: 6/10

These are my top 5

1 Polgar, Judit g HUN 2687 0 1976
2 Koneru, Humpy g IND 2595 25 1987
3 Hou, Yifan g CHN 2585 9 1994
4 Zhao, Xue g CHN 2542 8 1985
5 Kosintseva, Tatiana

Bartleby, I can't watch that coverage either. It crashes my machine, too. I'm watching at TWIC, the delay is unnoticeable.

Same here, how the hell can a simple webpage use all the resources of a fairly strong pc?

"can anyone tell me that they honestly have a feeling of what a 2500 "level of dominance" means?"
Utterly dominant in my own local chess club, as well as most if not all amateur clubs ,:) . With respect to the world top at any given time: statistically no (or very little) chance to win against the very top, still an outsider against the subtop [hard or impossible to be specific, e.g. top10, top 30 or top100 - also because the total number of (professional) players wasn't constant through time].
Generally, do you consider rating inflation a myth altogether? Or do you just think that it isn't properly quantified by Jeff Sonas?

Regarding the entire Sonas project, I see nothing fundamentally wrong with 1) admitting that ratings aren't directly comparable over time, and 2) still giving it a try. A remote chess analogy: What should a player do in a position where he doesn't fully understand what's going on? Resign? Offer (or force) a draw? Or play on, knowing that his play (as well as the opponent's) is unlikely to be perfect?

"You learned some HTML and PHP and figured out how to reference a database with a few figures. Big deal. What are you, Einstein?"

Oh, that hit me really hard.

I've never, ever claimed there's anything fancy or complicated about the liverating site - the only aspect of it that users might be thankful and/or impressed by, is that I bother to update it at a rate of 250 times/days per year, and is capable of keeping an eye on most things that go on with top players around the world, continously. That's the time consuming part. And most likely the reason why nobody else did anything similar before.

That I've spent 20-30 times the time it took to implement the liverating site on various kinds of rating analysis and research, including implementing a couple of rating systems [the first simple one I made has been online since 1994 btw - I'm not sure how much CGI-programming you did back then] - that experience and accumulated knowledge is the reason why I have strong opinions on chess ratings and systems.

The liverating project is basically simple, but pretty time consuming. And it was Thomas that brought it into this discussion, not me. I said it was irrelevant.

"Generally, do you consider rating inflation a myth altogether?"

No in some sense it exists. However, it's usually hard to agree on what we mean by "inflation", and hence how it should be defined and measured.

But the important part about it, is that you CAN NOT "adjust for inflation" by applying some linear (or more fancy) "correction" to ratings, in order to compare players that were active in completely different rating pools.

That an "adjusted rating" describes some objective and meaningful "level of dominance", which is Sonas' attempt to establish something "comparable" between eras, is something I simply don't buy.

My claim is that nobody has any intuitive understanding of what a "2500-level of dominance" means. You did exactly what I said you shouldn't do - compare it to other levels:

"With respect to the world top at any given time: statistically no (or very little) chance to win against the very top, still an outsider against the subtop [hard or impossible to be specific, e.g. top10, top 30 or top100 - also because the total number of (professional) players wasn't constant through time]."

Why couldn't you say exactly the same WITHOUT any "inflation adjustment", simply by looking at the actual ratings of the 2500-players' contemporaries?

The only consequence of claiming one has made "comparable ratings" by virtue of some adjustment, is that people are deceived to think that it's now possible to compare the skills of players that basically are uncomparable via ratings - because they never played each other, or somebody else that played somebody else that played the other, within a reasonable amount of time. (If you beat someone who beat Karpov in 2009 (or 2010, if he hasn't retired by then), it doesn't say much about how you would do against Tal or Korchnoi in 1978, does it?)

Let me quote a clever guy on chessgames.com:

"Ratings only ever measure a player's results in and relative to the current pool. Rating differences - which the whole system is based on - do a decent albeit (inevitably, due to limited sample sizes, different opponents and random chance) far from perfect job measuring domination in a pool. Absolute rating numbers on the other hand aren't even meant to measure anything at all. Trying to compare ratings from different pools is the best kind of experiment there is - everybody can draw whatever conclusion suits their views best. Comparing ratings within one pool is tough enough - even genuine experts like *frogbert* have heaps of trouble even approximating just how the ratings of new and upcoming 2700 players who have gained a high rating by playing lower rated opponents (a different sub-pool!) compare to the ratings of already established super-GMs."

Sonas' "level of domination" is something of a red herring in my opinion - "Oh no, we can't say who were the strongest player, but now we can tell who was the most dominant one!"

But people run off thinking they can tell who's better of Capablanca and Kramnik; To tell who was more dominant, you don't need to make a single adjustment. Kasparov and Fischer can easily be compared to their peers with no use of "adjustments" - you can simply consider the rating difference between Fischer and his peers and Kasparov and his peers using the FIDE ratings. The meaning of a 100 point or 150 point difference has been the same, untouched, since FIDE created its system. Difference/delta X indicates a theoretical "expectancy" of score Y in a game between two players.

Hence, in a hypothetical 10 game match, you could easily describe how much Kasparov or Fischer was expected to "dominate" against

1) an average GM of their era

2) against number two, 10 or 100 at some point in time

No adjustments are needed to do that - and the "adjusted rating" doesn't give you ANY additional info regarding 1) and 2) compared to the actual, non-adjusted FIDE ratings.

In conclusion: you gain little or nothing in clarity, but adds confusion and misunderstandings to an already badly understood concept - the chess ratings.

I don't have more time for this discussion today, Thomas. But I might get back to you on it some time later, if you would be interested and still have more questions you'd like to ask. For instance I have different methods of defining and measuring inflation that I find much more meaningful (and trustworthy!) than the average of the top N players, or the rating of the Kth top player.

yeah ... was he drunk?? or was he already in severe time trouble to make such elementary mistake?

on the other game ... Carlsen the escape artist :-).
He really knows how to play endgames.

Of course. Ok, I don't know if he played "much, much worse" here, but even if your opponent plays badly you still have to play well to win. Nobody ever said Carlsen didn't in that game.

Yes, well, I don't have an informed opinion on how hard it is to hold such endgames. But certainly it's quite possible to lose...Never expected Carlsen to, though.

All those girls have the chance to achieve happiness in their lifes , unlike you .

How unnecessary. You sad little man.

I totally agree with you d_tal. I found frogbert's diatribe against Sonas to be reprehensible. In my opinion, Jeff Sonas has given professional chess very valuable tools and models to work with. I remember when the Kasparov website was active and Jeff's predictions were usually right on. If frogbert is so knowledgeable about the "right" was to do these statistics, he should start sharing it with the rest of us instead of complaining that his "schedule" doesn't allow it.

frogbert, it seems you have time to write all these anti-Sonas blogs on Mig's site. Perhaps you could submit some meaningful articles about statistics and discuss the reasons you think Sonas is misleading the chess public with his ratings statistics. It's not enough to criticize someone and leave unsaid what the basis of that criticism is all about. We've heard your criticsms and until you can supply the real reasons behind these criticisms, enough said by you already.

You are dealing with delicate self steem issues here , the guy is surrounded by GMs and VIP but still finds the time to have meaningless arguments with trolls like me.
I say he uses that time to demonstrate his thesis , or even better : to find company .
:)

Jim,

First, it's quite ridiculous to pretend that I'm anti-Sonas. There is absolutely no personal issue between me and Sonas - at least not from my side.

"It's not enough to criticize someone and leave unsaid what the basis of that criticism is all about."

I don't understand why you say that, because I think I've been quite clear. My two main criticisms against Chessmetrics (that's criticism towards a system, not a person) are these:

1) It is per definition meaningless to "adjust for inflation" due to the relative nature of the rating system. A specific rating number bears no intrinsic meaning, only the DIFFERENCES between ratings bear meaning, and only the differences within one and the same rating pool.

See my latest answer to Thomas for more details.

2) Due to bad quality of data (heavily incomplete for lower rated, modern players, it still includes a notable amount of rapid games without those being treated as such, etc.), the ratings are only trustworthy for maybe the top 4-5% of the many players in his database.

That means, players rated 2500-2600 and above. It also means that age lists even for players born in the 80s (for ages up to 16-17) can hardly be trusted to any extent, like I also explained in a previous post. That is, DESPITE that they should be based on results in the near past, like 1995-2005.

Sonas is aware of his data problem.

---

Another major point I strongly disagree with Sonas about, is how inflation should be defined, and why. That's the point I have not had time to describe in much detail here, since my own work on the subject isn't complete yet.

"until you can supply the real reasons behind these criticisms, enough said by you already."

Like I said, I have indeed given the "real reasons" for the stuff that is easily presentable at this point. Completing my work on estimating rate of inflation at various times, and explaining what _I_ consider to be the actual workings of the player/ratings "ecosystem", is a much bigger commitment than writing this or that post on the daildirt or elsewhere.

Finally, I will repeat that I have absolutely nothing against Jeff Sonas as a person or a statistician. I think it's great with people that work with numbers in relation to chess, and when Sonas made it clear that his work with Chessmetrics was not in any way related to any Masters or PhD thesis, which had used to be my impression unfortunately, my admiration for the time he's put into this only went up.

I still find it quite unfortunate that, according to my observations, the impression Chessmetrics has left on chess people, is that rating inflation can be "adjusted away" to leave "comparable ratings". This has, IMO, made too many Fischer fans go off with statements like "if Fischer had been active today, he'd be rated 2900", meaning that he'd been stronger than everybody else.

Jim, if you don't understand what I'm saying by now, the chance is that you won't be much more receptive to a more worked through and academically oriented article I might write when I find time.

"In my opinion, Jeff Sonas has given professional chess very valuable tools and models to work with."

Without taking a stand on whether or not I agree with that, maybe you could tell us which tools and models you are thinking of, and then explain in which way they are valuable for professional chess.

I understand that you want to "support" Sonas, even if he isn't under attack (perceived consequences of his rating system and results are what I'm "attacking") - but let's try to stick to the subject and have a debate on the issue, not about persons.

[I've already said twice that the tone of my _initial_ post wasn't appropriate, I hope you picked that up by now.]

Fischer's IQ is unknown. (The reports of some Brooklyn school teacher need not be given credence.)

Kasparov was tested somewhat more accurately (although I think John Nunn objected to the methodology) by Der Spiegel. The result was in 130s.

""Sonas, given the choice between more personal exposure/attention and behaving like a responsible statistician, chooses the former and throws some numbers toghether."-(i.e. Sonas is an "irresponsible" statistican)"

Jeez, how many times does frogbert have to say he was sorry about the harsh tone in his first post until people are able to move on?

IQ is a far shakier construct than Sonas' statistics.

If you read the entirety of my post before commenting you would see that it was in explanation not admonition, and I preempted precisely what you just said.
Besides, I was also referring to his later posts. Like I say, please read and understand before you pass comment, tnx.

I did read it before commenting. I've re-read it and my point about not being able to move on clearly stands. But thanks for your concern.

Is it, perhaps, a question of which is worse?

Contrast:

"A specific rating number bears no intrinsic meaning, only the DIFFERENCES between ratings bear meaning, and only the differences within one and the same rating pool."

with

"the impression Chessmetrics has left on chess people, is that rating inflation can be "adjusted away" to leave "comparable ratings"."

I'm not a statistician, but I'll accept the first statement -- that ratings from different eras are NOT comparable. At least, I think that's what you are saying. And Chessmetrics of course starts with the same basic premise; it's just a question of whether that's possible (in which case, is the Chessmetrics approach "good enough") or a fool's errand?

Your argument seems to be that it's better to just accept that the ratings are not comparable rather than giving people a false assurance that they CAN be compared by adjustment.

I might tend to agree, but is that would people would really do, take an objective stance and just accept that ratings from different eras cannot be compared? If so, they will ignore Chessmetrics or just treat as something amusing but meaningless. They won't even try to compare, say, Fischer to Topalov or Capablanca to Carlsen in any sort of quantifiable "who is/was best" way.

But I suspect a lot of people are not that objective and are unable to avoid the human desire to compare and contrast players from different eras, and would either: (a) assume that unadjusted ratings from different eras ARE comparable; or (b) assume they aren't and come up with their own method of adjusting. In which case, which will be the greatest error: (a) assuming unadjusted ratings are comparable; (b) assuming Chessmetrics ratings are comparable; or (c) assuming that their own method of adjusting results in ratings that are comparable?

Pointing out the flaws in the Chessmetrics approach will not necessarily lead people to just drop all three of those options and retreat to the idea that the ratings can't be compared at all. More likely, they will drop (b) and choose either (a) or (c).

But then, I'm a cynic when it comes to the ability of people to be rational.

"a more worked through and academically oriented article I might write when I find time"

Unfortunately, not what most people are looking for. :-)

"Like I say, I don't care about it myself, but I suspect that this is the source of whatever flack you are getting."

Probably true, but still irrational. Then, people aren't rational all the time, and I'm as good as an example as anybody else, because the first post I wrote was too coloured by a reignited annoyance that, albeit real, it wasn't rational to allow to colour my writing.

But like acirce points out, I've made that clear a couple of times already (3 I think), and my honest and true stand is that I care much more for the subject/issue than any perceived personal issues, issues I don't (intend to) have with Jeff Sonas. I certainly don't think his intentions are to mislead anyone, but I fear that's the consequence nevertheless. And I think that is unfortunate.

The "irresponsible statistician" *implied* notion came from the initial post that I've repeatedly admitted I should've worded differently, something I said in my response to d_tal. I admitted that _prior_ to mishanp's reaction, and _prior_ to Jim's reaction. If more people want to say the same thing, they're really beating a dead horse...

"Unfortunately, not what most people are looking for. :-)"

In the US, I would guess most people are looking for the final proof that Fischer was the Greatest Ever, so you're probably right. ;o)

(Sorry that I will have to get back to you with anyting looking like a serious response, it's already 3 AM here, and I have to get my son up and to school in about 4 hours ...)

Yes, Frogbert, I replied to your postings BEFORE reading in full your later comments in which you provided more detail as to your reasons. In the contrary, based on those detailed reasons, I would be very interested in a more academically oriented essay which would spell out in detail your approach, or, at the very least, you should share your findings with Jeff Sonas. Perhaps a collaboration between the two of you would provide us with a more accurate approach. Of course, this is dependent on FIDE to provide ALL of its data to the general public, or at least to guys like Sonas and you who have a deep and abiding interest in these things. So, please accept my apologies; I didn't mean any disrespect.

"The "irresponsible statistician" *implied* notion came from the initial post that I've repeatedly admitted I should've worded differently, something I said in my response to d_tal. I admitted that _prior_ to mishanp's reaction, and _prior_ to Jim's reaction. If more people want to say the same thing, they're really beating a dead horse..."

Frogbert, I think you're in denial. Again here you highlight *implied* as if you didn't very clearly claim that Sonas was an irresponsible statistician. And it's not just a matter of the first post, as chesshire cat rightly pointed out. Most of your posts relating to Sonas are dripping with a mixture of disgust and mockery, even when you seemingly aim at a fair or detached summary of his point of view.

The most telling for me is this attack:

"Sonas does the circus act of
1) telling people that ratings can't be compared over different eras
2) making elaborate age-lists, where players are ranked & compared between different areas
3) making historical lists like "top 100" performances over multiple areas
Which talking head are people supposed to believe in? The one that says 1) or the one that does 2 and 3?"

When I challenged you to point out Sonas saying 1) you initially said that he did and that it's fine to attack someone based on what they said. Then later you issue a "clarification" that Sonas said the opposite - you couldn't even bring yourself to post an apology, despite having *implied* that he was an opportunistic hypocrite.

They both kept playing quickly after 21.Rec1, so Topalov's try (not a novelty, btw) must have been analysed by Carlsen too. Good preparation by Carlsen, possibly a slightly naive attempt by Topalov.

I find Radja a lot more boring and predictable than Leko , i don´t know why exactly but i´m tired of waiting for him to jump into a more Morozevich-type of attitude.

Although I had that confused - he plays the Dragon, but the Sveshnikov more often..

Radjabov's White openings aren't necessarily tame per se, of course it's what he does with them. He'd like to get more than he does, no doubt, but he doesn't.

Perfectly true about the Sveshnikov. Though not totally analogous the Petroff is somewhat similar in that there are plenty of sharp lines but that it has been so thoroughly analyzed that it's still to be considered quite solid in practical terms on top level. And I do believe Carlsen used the Svesh mainly as a drawing weapon - why not?

"Then later you issue a "clarification" that Sonas said the opposite - you couldn't even bring yourself to post an apology, despite having *implied* that he was an opportunistic hypocrite."

mishanp, you are the only one who seem to hold a personal grudge here, and I see no point in following you into that area. "Opportunistic hypocrite" are your words and your words only - I would never say stuff like that about Sonas.

And Sonas does not say "the opposite" of what I *meant to express* when I wrote my point 1):

"1) telling people that ratings can't be compared over different eras"

I wrote this without looking up anything, based on my memory. When I did look it up (again), I realized that I hadn't been precise. You say that I issued a "clarification", written with quotes.

I did certainly not call it a "clarification" - I said I had to CORRECT MYSELF a bit. I didn't have to tell you that I hadn't been precise - but when I saw the error I'd made (minor, from my point of view), I told you and everyone immediately.

"1) telling people that ratings can't be compared over different eras"

What was I trying to express about Sonas' statements with this? That ALSO HE says that ratings, EVEN AFTER ADJUSTMENTS, can't be used to compare players from different eras to tell who was the STRONGER.

A more complete and precise version of 1) would be:

"1) telling people that ratings can't be compared over different eras, to tell who the stronger players were/are (even after adjustments)"

That's what he says, but he doesn't fully take the consequence of it. If one takes the consequence of this observation, then both

2) age lists (for younger players) and
3) historical performance ratings

do NOT make much sense, in my very clear opinion.

Hence, whether you understand why I think so or not, there IS indeed a clear self-contradiction between saying that even adjusted ratings can't be compared across eras (to tell who the stronger player is), and doing 2 and 3.

The self-contradiction does not go away because I quoted Sonas imprecisely the first time (although in an easier-to-read way). And it would be VERY NICE if you would give me a tiny little benefit of the doubt here, instead of touting that I'm "in denial" and similar nonsense. I've been completely honest all along, I don't have photographic memory, but you should realize that there are few people around that have read Sonas' articles on Chessbase and on his own site more thoroughly than I have.

Jeff Sonas himself obviously understood. One final comment about this:

"here you highlight *implied* as if you didn't very clearly claim that Sonas was an irresponsible statistician."

If we for a moment ignore that I've already apologized for the wording in my initial post (and that you keep beating a stone cold horse here), I think you should consider the difference between stating that someone IS something (on a permanent basis), and saying that some behaviour in ONE SPECIFIC situation is irresponsible.

You can add that example to the other examples where you rephrase my statements to the worst ones possible given what I _actually_ did say. The amount of honor you ascribe to me, despite my efforts of setting things straight, appear a much graver personal attack than anything that can be made up that I did to Sonas. Did anyone give you a moral right to describe me as a vicious person in denial?

Please at least give it a thought. And I won't answer another round of personal attacks from you, so you have an excellent chance to fire whatever you have left to finish your picture of me as a lousy individual.

Well I'd be that someone, and I think they have a point. The porridge I am now eating is much healthier than the caviar and champagne I'd gotten used to.

"At an early stage, he criticized Jeff Sonas for changing his formulae through time. Later, he implicitly writes "there are better methods to correct for rating inflation, mine for example" "

Thomas, two misunderstandings here, one slight and one clear:

1) Regarding Sonas updating/improving his models, I said that was fine, as long as the public (or I) still got to see how his previous model would work for new data. And in a scientific setting one would typically explain the rationale for changing a model, clearly pointing out what was "wrong" with the previous one, and how the new one is better. I certainly didn't say that he wasn't "allowed" to improve his models.

2) The big misunderstanding: I never said and never wanted to "correct for inflation" in a better way. My view is that this simply can't be done in any meaningful way. What I actually said, was that I had a better and more meaningful way of _defining_ and _describing_ inflation, which might allow me to better understand how it works and describe its levels at various points in time.

If one understood it better, one could possibly make countermeasures in a meaningful way, not in arbitrary ways. At no point would I try to retroactively "adjust" or "correct" for inflation.

Thomas, note that even with a system that didn't experience any inflation, comparing ratings from different eras does STILL not tell us anything useful about who the stronger players were/are. The motives for eliminating inflation are different and NOT related to enabling player comparisons between eras.

I'll try to keep this response short without too much quotation - the posts are already unmanageably long.

I'll accept that Sonas didn't say the opposite of "what you meant to express" - but he did say the opposite of what you wrote. And taking his actual statements there's no easy rhetorical attack on him for playing "circus tricks" or being an inconsistent "talking head". You painted him as hypocritical when he was actually consistent on his terms (I realise you disagree).

Again, it's not a question of disagreeing with your arguments (though I don't find "dominance" a useless concept for comparing players and in any case ratings never directly show chess skill, over any period of time) - I just object to the tone. I seriously believe you don't quite realise what you're doing (certainly if you believe only your initial post was problematical), but then that's only a minor character flaw. I find your site a very useful resource and don't have an axe to grind.

p.s. sorry about "clarification", that was a misquotation - I see you talked about Sonas' "clarification" after having said you needed to correct your post.

"...just another troll like Manu or Luke..."

d_tal, we already know that you aren't very nice.

You are just angry all the time for some reason.

Shame that Evry Grand Roque couldn't get a bigger team. In Europe Echecs, the manager (Baumhauer) said they spent 100000 euros for the budget for victory in the Top 16, but no Nakamura, Svidler, Vachier, this time around. Or even Yusupov. In the previous, Feller was 10/11 before, but on 5th board or lower, but he'll have to grow his britches to play board 1 here.

Ugh. If you've been working on that little ditty for a special occasion, you wasted your time. Not reflective of your usual creativeness.

Round 2 has Yerevan versus Moscow as a big matchup.
Aronian-Gelfand, Akopian-Karjakin, Sargissian-Caruana, etc.

My guess is that Urals (Grischuk, Shirov, Malakhov, Rublevsky, Dreev, Bareev, and #3 seed) will win. Yerevan, is All Armenian, but I think their 5th/6th boards are too weak. Plus Aronian is doubling as captain and player (bad in cricket, as Svidler could say, extra shoulder weights).

Svidler-Nakamura (Petersburg versus Husek Wien) also a second rounder to watch. Don't think they will quick-draw it!

You're clearly not a Gunsmoke fan.

I have to agree with mishanp as well. Incidentally for someone who complains about lack of time Frogbert seemed to spend (waste?) a lot of time on lengthy posts on this thread. I enjoy/find useful both LiveRating and Jeff Sonas stuff......

Baden-Baden is missing their top four boards: Carlsen is busy elsewhere, so is probably Anand (preparing for his match against Topalov). But Svidler (Sankt-Petersburg) and Shirov (Ural) are playing for Russian teams, so does Caruana (Moscow) - also on Baden-Baden's Bundesliga lineup (board 12!!? as they rank players strictly by rating).

Is this a question of national ties - or for Caruana, feeling obliged to his Russian coaches? Is there more money in Russian club chess? Or do they take the European Club Cup more seriously?

When Kiev took 3rd on tiebreak last year, it was the top news story for Ukraine sports. The Part D Section V of the FIDE Handbook concerns the rules. Every federation gets two teams if it has a (national) Team Championship, or one without. There are no rules about forming teams, other than (1.2): "The teams shall be composed of six players all of whom must be members of the club and were entitled to play for the club in the national team championship of the federation... However, players who are foreigners to the federation, must have played at least two games in this championship."

But then, you could also change your federation at will presumably, so it's largely meaningless, so they extend "Foreigners" also to mean permanent residence elsewhere, but you can still tilt the system. "If a player has been playing national team competitions in more than one federations he has to decide, for which team he will play."

Unfortunately, chess is not UEFA, and so there's no news about the intrigues of player acquisition. As with Caruana, last time around Linex-Magic (Spain) put him on the list, but he didn't play the requisite 2 games. It was gossipy all-around, with Shirov being replaced (by Rublevsky, at first, I think). Eventually, he was vetoed (though one of the organizers OKed it prematurely) by a Captain's Vote.

In fact, Caruana was 5/6 (top performance of 2896 with 3rd board) for Moscow64 in the April Russia Club Cup, but mysteriously? didn't play in the last round, so maybe it's logical he plays for them now.

Strange that Vachier-Lagrave is not around. He's be logical for Economist. Guess Mulheim (Bundesliga) didn't make it, and he played for them last go-around.

Seems they are also flexible concerning the number of teams per federation. Russia has five teams (of course this makes sense, and all of them are in the top8 by average rating). Germany has four teams ... this would also make sense as the Bundesliga is the strongest league (or second to the Russian league). However, the participating teams are Baden-Baden (#1 of the competition), Bremen (#2), Solingen (#6) and Berlin (#13! for them it is sort of a tradition, using the Club Cup as a warmup for the coming season). And ALL of these teams lack (to a slightly different extent) their strongest players. So it rather seems like "whoever wants to play is allowed to play". Mulheim was #5 last season, but apparently not interested or unable to cover the expenses.

Nicely spotted Boz.

But then by similar reasoning ordinary ratings are also of limited use for comparing players who've come up against a different rank of opposition, yet Frogbert is still "reaping the benefit of media exposure" by publishing the numbers. Not to mention encouraging chess fans to become more and more transfixed by daily rating changes that of course mean close to nothing when it comes to assessing the chess strength of the players involved. Or indeed to mentioning tournament performance ratings when it suits: http://reports.chessdom.com/measure-up/final-july-arting-list-estimate

I'm not seriously making those criticisms, by the way - both sites are entertaining and add something of interest for chess fans. Of course Frogbert is on much more solid ground in just updating the (also flawed) FIDE ratings, but I don't see any great evil in Sonas' site. We all just need to realise that however much statistics help chess strength is still a highly subjective concept.

And again - I don't think anyone objects to Frogbert presenting reasoned arguments against Sonas' system, just to the unnecessary tone with which they were expressed.

"Svidler-Nakamura (Petersburg versus Husek Wien) also a second rounder to watch. Don't think they will quick-draw it!"

Nope - Svidler won as usual.

Very nice B+B vs. B+N ending by Svidler.

¨Otherwise you will be mistaken for just another troll like Manu or Luke, who always address the man, rather than the argument.¨

You were schooled in the other thread ,learn to deal with that .
;)

Has Wang said he is systematically playing for draws even with White? If not, why do you think he is?

Live ratings is becoming the most expensive free-service in the chess world .

Oh, because it's obvious. Funny because that's not my impression, rather that he doesn't have good enough openings, but since it's obvious I guess there is nothing more to say.

That's an easy prediction: with Carlsen's current two point lead, three draws in the final rounds are enough to win the tournament. 1/3 should still be enough, unless Topalov (or, for the sake of completeness, Wang Yue wins his last three games).

We could already move on to other issues (I give my approximate calculations):
Can Carlsen cross 2800? "Yes he can", 2.5/3 should do the job.
Can he become #1 on the live (and 11/2009 FIDE) rating list? This seems to require 3/3 AND Topalov losing some more points in his remaining games.

Yet I don't think Carlsen himself should be(come) carried away by such considerations ... .

Yeah, but it's wild theory ... . All three games look promising: Radjabov has black today and went for a dragon fight. Leko against Topalov plays the same variation that yielded him a "moral victory"(!? ,:) ) against Carlsen. It took him six moves to get his knights on the usual squares c3 and f3 ... without being worse as far as I can tell.

Short off-topic question: Where can I find ALL the games from the European Club Cup? Only the top three matches are covered live, regarding Svidler-Nakamura I could only find the result.

And in round 3 Svidler beat Ivanchuk with Black. Very nice to see.

"This _is_ the Svidler thread, right?"

Wrong. It's the acirce monologue.

That is probably correct, thanks.

But, and this is an extremely important question, why is Пётр so often transcribed as Peter? What's wrong with Pyotr? Are we going to call Tal "Michael" as well?

Peter/Pyotr and Michael/Mikhail

Don't know why but both FIDE and Wikipedia go with Peter and Mikhail.

How about Mick/Mickey (O') Tal? That way he could have an Irish connection.

Svidler doesn't seem to mind! http://www.psvidler.net/

And if we really want to get into it then shouldn't Tal really be either Tal' or Tals :)

If you have to "blame" one of the players (not sure why), at least choose individual tournaments for that. In any event draw against Gelfand is a good result for Svidler, who is normally more likely to lose than to win such a game with Black.

It was a missed opportunity for Svidler to push his TPR above 3000 ,:) . No, seriously, I was looking forward to watching a game between two top GM's, the opening looked interesting and then it was suddenly over - no big deal though, the event has many interesting games and 36 were transmitted live. As you (and I) pointed out it's a team event so quick draws on the top 3 boards, having the lower ones fight it out, may have been part of the match strategies - it worked out well for Svidler's team in the end.

On Sigi Strut's summary: It's not over yet for Yerevan, losing a close match to a nominally higher-rated team can happen (it also happened to the Armenian Olympiad team). Skopje - well it's not really a Mazedonian team ... of course the same holds true for, e.g., Baden-Baden: only the Russian/ex-Soviet and less so the Israelian teams can compete on the top boards with only local players. And mismatched pairings will continue in the next rounds: Whenever a top team loses a match they are likely to get a 'beatable' opponent in the next round. For example, tomorrow we will see Moscow (#3) - Solingen (#31 with six German players).

I feel bad for you, wallowing in another monologue. So, at least I said something. Hope it cheers you up.