Carlsen, Caruana & Yifan: a ratings analysis

I don't believe that weaker players are underrated and stronger overrated, the top players in general do not lose rating points in opens, the last event of that sort some of them played was Giri and Kramnik in Qatar, where both gained quite a few Elo points.

Carlsen's rating has been stagnating because he simply didn't perform much above 2860 in tournaments. However, he can get +2,2 for beating no. 100 on rating list, rated solid 200 points below him. 

Could you please explain yourself, what makes you think they are 'clearly not as bad as 2500s'?

Seriously. 

Okay, done. To remove the effect of ratings inflation, I took the value of 50 rating points per eight years as given in the article to which I linked above, and used it as a linear offset based on a birth year of 1987 (i.e. Nakamura, Humpy, and Dzagnidze retain their original ratings and the younger players' ratings are lowered accordingly). The resulting graph doesn't look much different since most of the investigated players are quite close together in age, but there are three noticeable effects:

1) Wei Yi is no longer clearly better than Magnus Carlsen; the two are now pretty much evenly matched.

2) Hou Yifan is no longer clearly better than the rest of the women; despite her early start, she seems to end up evenly matched with Humpy Koneru. Neither of them has reached the level of Judit Polgar (~2700).

3) I only now realized that Humpy is the first name.

As is stated in the linked article, comparing absolute rating values may not be as useful as comparing differences; but most of the features discovered so far don't seem affected by the speed of inflation in such a short time period.

fabelhaft, they played one event. And gained a little. They would play another 2 and gained a little. One would be average and one catastrophic. And we have a sum that is -10 for example.

NewArdweaden, Carlsen performed above 2860 in "closed" tournaments. As you mentioned, he gets 2 pts for winning a 2650. I'm pretty sure he would go down after 5 "open" tournament. The overall score may be the same, but rating change would not be.

Not having an argument(?) for the 2nd one. Just intuition. Some men collegues that are 2500-2600, are pretty 'bad' sometimes.

Merely speculations then. 

Yeah, or "artificially" maintaining 50+ lead. Blah.

"fabelhaft, they played one event. And gained a little. They would play another 2 and gained a little"

That depends on what is much in one quite short event, Giri gained more than ten points.

Mimogrede NA, si videl, da so Tisaj in Markošek ter tudi ta drug par remizirali. Če zmagam, bom imel solo lead s 100%, hah. 

I set up these two artificial tournament in which there are 20 rounds.

One has average rating 2700, the other one 2500. To have 2860 performance, you need to score 14 and 19/20 respectively.

If we assume there all player have the same rating (highly unlikely) the player would lose -2 in 2700 tournament and gain +10 in 2500 tournament. And I doubt the difference between Carlsen and ordinary GM is artificial.

Kdaj boš pa igral?  Lep bi bil tudi solo lead pri 9. rundi 

Yeah, but I highly doubt Carlsen (or anyone else) would score 19/20. :)

Nimam pojma, upam samo, da ne med počitnicami. Če ne se bom moral doma spomniti en beden izgovor, zakaj bi šel v Ljubljano, ki pa bo težko sprejet. Pff. Najbolje, da napišem Francu, da ne morem. :D Aja, kolokvij matematiko sem pisal 43 od sto. :)

P.s: Upam, da tega moj nasprotnik ne bere, če ne sem naj*. :D
Če ne'j pa bulš, de u naričju pišem, čeprow'b se dalu tut tu razwezlet.  

This is true, though.

Jebeš, bo druɦič bulš. 

Pa majm špileje, lan je blu sednɦa maje, litas je pa šestɦa (pa še ti bi moɦu u tark petɦa špilet, jelde?), tak de bi moɦlu bit. Zmisln se ki, reč de imate ki za šula nardit al niki. Ti wje pa ja pustil jet.

Se pa prejmu maju u'uriedu, le tel'h n'par dni ka'u uastalu pret prejmu majem u'lih ma bel tak tak. Porkalapipa, ka sn že djau, de im'ma fraj, edinu, če se'm spounu, de sa kešne uaje alpa nek šit.

"Porkalapipa"

Aja, se pačitnce sa druj tidn . Dijobono porkamadona, se 'uš ze zmislu niki!

Believing the sexes are naturally inclined to different types of intelligence is not sexist, though of course it is not necessarily true (it may be but there is no conclusive evidence either way). Women and men do tend to respond differently to certain stimuli, the difficult part is determining with certainty what is nature and what is nurture. Personally, I believe men and women are fundamentally different in a few ways (I recognize that nurture plays a greater role in who we become, but that doesn't completely take nurture out of the picture) -- but I do not believe either sex is "better" than the other, nor do I deny opportunity to anyone based upon their sex -- so, I am not sexist.

This is one of those self-fulfilling prophecy sorts of experiments. You expect a difference and then go looking for it. 

There's no proof that ratings inflation has happened just because people's ratings are going up. Rating gain can be tied entirely to playing improvement. The 100th best female today is rated 2367. In 2010, the 100th best female was rated 2366. In 2000, the 20th ranked girl was rated 2314. In 2015 the 20th ranked girl is rated 2308. 

The fact that there are more players competing today than fifteen years ago also means that the competition is tougher, and the level of play has improved. 

Adjusting for "inflation" when you don't even know whether inflation has occurred, let alone the rate of that inflation, is just one problem with this investigation.

Well, inflation seems to be happening. Just a couple of weeks ago I edited an old Wikipedia article on ELO rating with information from 2004 - it said that there were only 27 or so GMs with rating 2600+ in the world, with Kasparov being 2851. Nowadays, there are 43 rated 2700+.

The fact that there are more does not mean the rating system is inflated. It could easily mean that the number of top players has increased. 

Two simple questions: what is the rate of inflation, and what is causing it?

http://en.chessbase.com/post/rating-inflation-its-causes-and-poible-cures

The last graph, which is the most important one, is unfortunately not very nicely done.

I'm not planning to explain it myself, It doesn't have a direct link to actual issue that is being discussed.

Sure, women and men are different. Each individual is different from one another also I think reasonable people don't look at the statistics and assume that males have a special gift for playing chess or driving or logic, etc., over women. Women can excel beyond any man with our minds. It comes down to the individual, not the sex, in my view:)