If this is your first visit, be sure to
check out the FAQ by clicking the
link above. You may have to register
before you can post: click the register link above to proceed. To start viewing messages,
select the forum that you want to visit from the selection below.
Policy / Politique
The fee for tournament organizers advertising on ChessTalk is $20/event or $100/yearly unlimited for the year.
Les frais d'inscription des organisateurs de tournoi sur ChessTalk sont de 20 $/événement ou de 100 $/année illimitée.
You can etransfer to Henry Lam at chesstalkforum at gmail dot com
Transfér à Henry Lam à chesstalkforum@gmail.com
Dark Knight / Le Chevalier Noir
General Guidelines
---- Nous avons besoin d'un traduction français!
Some Basics
1. Under Board "Frequently Asked Questions" (FAQs) there are 3 sections dealing with General Forum Usage, User Profile Features, and Reading and Posting Messages. These deal with everything from Avatars to Your Notifications. Most general technical questions are covered there. Here is a link to the FAQs. https://forum.chesstalk.com/help
2. Consider using the SEARCH button if you are looking for information. You may find your question has already been answered in a previous thread.
3. If you've looked for an answer to a question, and not found one, then you should consider asking your question in a new thread. For example, there have already been questions and discussion regarding: how to do chess diagrams (FENs); crosstables that line up properly; and the numerous little “glitches” that every new site will have.
4. Read pinned or sticky threads, like this one, if they look important. This applies especially to newcomers.
5. Read the thread you're posting in before you post. There are a variety of ways to look at a thread. These are covered under “Display Modes”.
6. Thread titles: please provide some details in your thread title. This is useful for a number of reasons. It helps ChessTalk members to quickly skim the threads. It prevents duplication of threads. And so on.
7. Unnecessary thread proliferation (e.g., deliberately creating a new thread that duplicates existing discussion) is discouraged. Look to see if a thread on your topic may have already been started and, if so, consider adding your contribution to the pre-existing thread. However, starting new threads to explore side-issues that are not relevant to the original subject is strongly encouraged. A single thread on the Canadian Open, with hundreds of posts on multiple sub-topics, is no better than a dozen threads on the Open covering only a few topics. Use your good judgment when starting a new thread.
8. If and/or when sub-forums are created, please make sure to create threads in the proper place.
Debate
9. Give an opinion and back it up with a reason. Throwaway comments such as "Game X pwnz because my friend and I think so!" could be considered pointless at best, and inflammatory at worst.
10. Try to give your own opinions, not simply those copied and pasted from reviews or opinions of your friends.
Unacceptable behavior and warnings
11. In registering here at ChessTalk please note that the same or similar rules apply here as applied at the previous Boardhost message board. In particular, the following content is not permitted to appear in any messages:
* Racism
* Hatred
* Harassment
* Adult content
* Obscene material
* Nudity or pornography
* Material that infringes intellectual property or other proprietary rights of any party
* Material the posting of which is tortious or violates a contractual or fiduciary obligation you or we owe to another party
* Piracy, hacking, viruses, worms, or warez
* Spam
* Any illegal content
* unapproved Commercial banner advertisements or revenue-generating links
* Any link to or any images from a site containing any material outlined in these restrictions
* Any material deemed offensive or inappropriate by the Board staff
12. Users are welcome to challenge other points of view and opinions, but should do so respectfully. Personal attacks on others will not be tolerated. Posts and threads with unacceptable content can be closed or deleted altogether. Furthermore, a range of sanctions are possible - from a simple warning to a temporary or even a permanent banning from ChessTalk.
Helping to Moderate
13. 'Report' links (an exclamation mark inside a triangle) can be found in many places throughout the board. These links allow users to alert the board staff to anything which is offensive, objectionable or illegal. Please consider using this feature if the need arises.
Advice for free
14. You should exercise the same caution with Private Messages as you would with any public posting.
15. Have fun!
(Thanks to Nigel Hanrahan for writing these up!)
Reasons to believe there's FIDE rating "inflation"?
Re: Reasons to believe there's FIDE rating "inflation"?
Fischer trusted his own analyses more than the computer’s.
From a Guardian interview with Vishy Anand in 2011:
The BBC are currently showing the documentary Bobby Fischer, Genius and Madman. You met Fischer in 2006, a couple of years before he died. What was he like?
I found him surprisingly normal. Well, at least not very tense. He seemed to be relieved to be in the company of chess players. He was calm in that sense. He was also a bit worried about people following him, so the paranoia never really went away. But I am really happy I got the chance to meet him before he died in 2008. It was weird as well because I kept having to remind myself that this was Bobby Fischer sitting in front of me!
Were you tempted to whip out a pocket chessboard and challenge him to a quick blitz game?
No, because he whipped out his pocket chess set first and we started to analyse some recent games I'd played.
Really?
Yes, I showed him some of my games from Wijk aan Zee and tried to share some interesting developments. He was sort of able to follow everything – he hadn't lost his sharpness for chess – but his methods were a bit dated. In that sense he had fallen behind.
How do you mean?
Well, he had some suggestions, and he was sort of in the ballpark … but when I would tell him that the computer says white is winning here, for me that was a sign to move on – but for him it was a starting point to argue with me! [Laughs]. I found it difficult to say to him 'No, no, no – these computers are really strong. You shouldn't be arguing with them!"'
Re: Reasons to believe there's FIDE rating "inflation"?
Just to collect them, here are four arguments for the conclusion that there is FIDE rating deflation. The first two (at least) have already appeared in this thread, I've just restated them and added titles.
1. Points Retired
Players tend to quit with more points than they entered with. FIDE ratings are zero-sum. Therefore there are fewer FIDE rating points per player today than there used to be. A consequence of there being fewer rating points per player in the system is that each rating point is harder to earn, so a rating of #### today is harder to achieve than the numerically identical rating years ago.
The first premise, 'players tending to quit chess with more rating points than they entered with', is an empirical observation: most players get their first rating when they are still improving, as they get better their rating goes higher (because they take points from their opponents), most players stop playing before their rating drops all the way back down to its initial level, so most players take rating points out of the system.
"zero-sum" means the sum of the changes in all the players' ratings for any game (or tournament as a whole) is zero. This happens when the points gained by one player are taken 1:1 from another, e.g. when there are no bonus points added to the system. FIDE ratings seem to be very close to zero-sum.
I don't know that the FIDE rating system is perfectly zero-sum, but I think it's reasonable to assume it is close enough. When new players get a provisional rating it might be much higher (or lower) than their opponents', without adding or taking rating points from those opponents. [there's a well-known case of a US inmate who manipulated this provisional-rating loop-hole: he ran tournaments in prison, losing to weak unrated inmates so they'd get high provisional ratings, then beat them so he'd collect their ratings, eventually crossing 2700 USCF.] Each time a newly-rated player gets a high provisional rating, and then quits tournament chess after losing rating points, rating points are added to the system, causing inflation. The question is whether the effect of these players is greater/equal to/less than those players who start with low provisional ratings and quit tournament chess after raising their rating. I see no prima facie reason to think provisional ratings are skewed more up than down, let alone enough to completely mitigate the effect of players retiring from competition with higher ratings than they began.
2. Population Increased
There are more people exposed to and playing chess than ever before. When the size of a population increases the number of members at the extremes increases, and the extremes tend to become more extreme.
The world's two most populous countries, China and India, have much more chess activity than they used to. When the population size increases the bell curve describing that population goes "up" everywhere (explaining why there are more players at 2700+ than there used to be), and the "tails" on the bell curve get more extreme (explaining why the highest rated players are rated higher than ever).
Strictly speaking, this is not an argument for or against inflation. It is an argument against the feeling that there's "something wrong" with there being so many more 2700+ players than there used to be. According to this argument, more 2700+ players is the only reasonable thing to expect when comparing larger and smaller populations.
3. Knowledge Accumulates
Chess is (in part) a knowledge game, and knowledge is cumulative. Players who use this knowledge have an advantage over players from the past who could not. Therefore modern players have a knowledge advantage over their predecessors which, all other things being equal, will produce better players.
More is known today than ever before about how to play chess and dedicated players can learn things their predecessors could not. As a simple example: today you can study all the books and magazines Fischer studied when he was "getting good" --- you can even do it without using computers, if you think that would be better--- but you can also learn from the games of Fischer, Karpov and Kasparov, which no player before Fischer could.
Not only is more known about how to play chess, but more is known about how to teach and study chess. Modern coaches and players who take advantage of this will get better results.
[as a bit of anecdotal support: I remember struggling to learn about "distant opposition" by studying Znosko-Borovsky's "How to Play the Endgame in Chess". Since then, much better ways of explaining it have been discovered: last week I read the first few pages of "Amateur to IM" by Jonathan Hawkins, and the explanation there is MUCH clearer, more succinct, and easier to learn. Students who read Hawkins have an advantage over old-timers like me: they'll learn faster and have more time to learn other stuff.]
Again, this one is not an argument about inflation, but is an argument against the feeling that motivates people to suggest there is inflation: the feeling that there's "something wrong" with there being so many more Super-GMs than there used to be.
4. Technology Improved
Computer engines, databases, tablebases, the internet.... today's players have easier and more complete access to superior chess information than ever before... [we've all heard this one]
FWIW, I think this is essentially the same as the "Knowledge Argument" above, since knowledge is a tool (an immaterial one) and technology is (in part) embodied knowledge.
Just to collect them, here are four arguments for the conclusion that there is FIDE rating deflation. The first two (at least) have already appeared in this thread, I've just restated them and added titles.
Honestly, I don't get it. Can we at least acknowledge the facts? The data clearly indicates that there is significant inflation in the FIDE ratings. Average rating went up by more than 100 points since 1985:
#1: The rating floor (which used to be at 2200) sort of mitigates this problem and may actually cause inflation (discussed above).
#2: Yes, more players would mean more players at the edge of the rating pool, but we need data (membership trends) to support that.
#3 and 4: As you said, this has nothing to do with inflation/deflation. Of course modern players are better than their predecessors, but it doesn't mean that it's harder or easier to be a good player. It's what you do and how hard you train with the current tools.
Now, this feeling that FIDE rating are deflated may be due to the fact that CFC ratings went through some deflation, but that's a quite different system (i.e. high turnover of junior members).
2. Population Increased
The world's two most populous countries, China and India, have much more chess activity than they used to.
You might be surprised but China does not have many FIDE rated players If you'll compare to its population it will be laughable. I was surprised too :D
Honestly, I don't get it. Can we at least acknowledge the facts? The data clearly indicates that there is significant inflation in the FIDE ratings. Average rating went up by more than 100 points since 1985:
No, apparently you don't get it.
The a rise in the average (mean, actually) rating does not necessarily imply inflation. If the average strength rises accordingly then there is no inflation.
Evidence that the actual strength of chess players has increased over the years is now available from computer analysis. Given that the ELO system is inherently deflationary then in order for you to provide convincing evidence of inflation you have to show that the average strength of players over the years has not increased for your claim to be credible. You have provided none.
1. Points Retired
Players tend to quit with more points than they entered with.
Hi John,
You make excellent points. It is refreshing to see a thoughtful assessment of the rating deflation/inflation question. Well done.
I am particularly interested in your point 1, Points Retired. Most players leave the rating pool with more points than when they enter. To argue otherwise would suggest that we get weaker as we play more. (well, maybe that's true in my case,:( but I would not extrapolate that to everyone.) Even for myself, I hope to someday leave the pool with more points than I brought. ;)
I am confronted with the issue of rating deflation every 6 weeks when I update the junior ratings at the Mississauga Junior Chess Club. I use several techniques to inflate the rating pool. Since it is a junior club, turnover is higher and faster than you would see in either CFC or FIDE. Consequently, the natural deflation is on steroids and easier to see. For those who enjoy mathematics, here is my proof.
Background
I updated the junior ratings this past sunday. Both the updated (Feb 14) and previous (Nov 29) rating lists are on our website if anyone wants to verify the data. We have a steady flow of new kids and they are dropped from the list after an absence of 20 weeks. Thus the rating pool is current and relevant. I have maintained this rating system for several years now, following the same basic formula as everyone else. To keep it simple, I assign new players a rating based on their performance in their first tournament. To correct for obvious errors, I take great liberty in granting bonus points where appropriate.
My goal is to keep the Mississauga system on par with CFC ratings. We have lots and lots of beginners. Our rating pool has maintained an average rating of between 450 and 500, which IMHO, is accurate. Up until recently, that assessment was mostly on faith. However, we now have some supporting data. Our recent YCC Qualifier (and last year) was rated by all 3 rating systems: CFC Quick, CMA, and Mississauga. Most of the kids at our Qualifier already had multiple ratings. A comparison of the average of the 36 Mississauga kids post qualifier ratings are: Mississauga 715, CFC Quick 773, CMA 708. Close enough in my estimation.
Calculations
Nov 29 rating pool - 104,577 (225 kids, average 465)
New kids - 3,526 (15 kids, average 235)
Subtotal - 108,103 (240 kids, average 450)
Participation points - 1644 (2 points per game)
Maintain rating floor - 361 (bring all ratings to minimum of 200)
Bonus points - 334 ( trade secret)
YCC Qualifier - 499 (trade secret)
Subtotal - 110,941 (240 kids, average 462)
Kids leaving - 4,687 (12 kids, average 391)
Feb 14 rating pool - 106,254 (228 kids, average 466)
In summary, the starting average rating was 465.
Adding new kids, average drops 15 points.
Various adjustments, average up 12 points.
Kids leaving, average up 4 points.
New average rating is 466.
John, as you can see, the turnover of players has caused the average to go down 11 points.
With all my "inflationary techniques", I have brought the average up 12 points.
IMHO, this proves the point. :D
However, no doubt 3% of climate scientists will disagree.
You might be surprised but China does not have many FIDE rated players If you'll compare to its population it will be laughable.
Maybe not so surprising. You don't need a FIDE rating to find out how good you are, so why pay the FIDE service charges? So another way to look at it is that it would be embarrassing waste of money if China paid FIDE rating fees for all their players, and not just their strong ones.
It is laughable, and maybe this time the joke's on FIDE. :p
There's a funny anecdote on Kasparov's new Chessbase DVD about how this used to happen in the Soviet Union, when players could get lots of strong competition but none of it was FIDE rated. Kasparov was sent to his first international tournament because the SU chess authorities thought it was a junior event, and not the strong GM-level RR it actually was. He entered with no FIDE rating, won big, and earned what should have been a provisional rating higher than Fischer's 2785. Kasparov says that since FIDE didn't want to give him such a high rating they went back to look at some of his past non-FIDE rated tournaments, and decided to rate some of them so his provisional rating would be lower!
Re: deflation:
if China (and India) don't pay for the majority of their players to get FIDE ratings but only their strongest players, this would not cause the bell curve of FIDE ratings to go "up" everywhere, but only up at the top/right end, which would be consistent with the observed increase in very strong/high-rated players.
Re: Reasons to believe there's FIDE rating "inflation"?
Why do some think the purpose of the rating system is to measure past players with present players? I thought the purpose was to measure current players against each other. Using some of the logic expressed in this thread, one could extrapolate and come to the conclusion that half the world's tournament players in the year 2300 will be better than Alekhine. If true, that wouldn't make them GMs. It wouldn't even make them Masters.
"Tom is a well known racist, and like most of them he won't admit it, possibly even to himself." - Ed Seedhouse, October 4, 2020.
The a rise in the average (mean, actually) rating does not necessarily imply inflation. If the average strength rises accordingly then there is no inflation.
He he... sorry if my previous message was not clear. Here's a little more precision on what part I don't get. ;)
1-To me, inflation (in the context of this discussion) is just the phenomenon of ratings going up with time - i.e. exactly what we observe with FIDE ratings.
2- As I wrote previously, I agree that the average player is better now
However, rating points have to come from somewhere.
Overly simplified example: 10 guys train and play intensively against each other for 1 year. The average level of skill will go up, but not the average rating.
So, basically, I agree with the elements: yes FIDE ratings are going up and yes, players are getting better with time... but I'm not convinced that the two are intrinsically related.
If 'better players' = 'increase in total amount of rating points in the pool', then somebody must be able to explain where these rating points are coming from??
Honestly, I don't get it. Can we at least acknowledge the facts? The data clearly indicates that there is significant inflation in the FIDE ratings. Average rating went up by more than 100 points since 1985: http://www.chessbase.com/newsdetail.asp?newsid=5608
I reread the Sonas article in your link, but AFAIK it does not show that there is inflation of 100 points since 1985, let alone that the "average rating went up by more than 100 points since 1985" (my emphasis).
Sonas believes there is inflation, and he carefully distinguishes two senses of the term "inflation":
a) "that a player with rating X today is not as objectively strong as a player with rating X was in the past"
b) "indicate that a previously elite club, such as all players rated 2700+, or all players with the grandmaster title, has become much less exclusive."
There is no doubt that there is inflation in the second sense above, no one is arguing about that. Sonas 's article is trying to show that there is inflation in the first sense: that rating points are easier to get than they used to be.
He believes there is inflation, but AFAIK his article gives no evidence for this claim. The closest thing he offers is summarized his graph "Inflation of entire rating pool, since 1975". The lines on the graph do show an increase of about 100 points of the ratings of the players ranked 100th, 500th, and 1000th. He also graphs the ratings for the 5000th and 10000th ranked players, though that data is (probably) less reliable, since FIDE had only recently started rating/ranking players that low.
The problem with using this as evidence of inflation is that it is all data from the right hand side of the bell curve.
I didn't notice this at first. When I looked at it I thought the 10000th ranked player must be in the lower half of the ratings bell curve. But as you can see from his next chart, "Distribution of FIDE Ratings Pool across time (total players)", the year he wrote that article there were over 60,000 players with FIDE ratings. To show that there is rating inflation -- rather than just rating increase among the top players -- Sonas would have to either:
a) chart the ratings of sample players rated in the lower-half of the FIDE bell-curve (e.g. players ranked 35000 and below); or
b) sum the ratings of all the active players and divide by the number of players (i.e. calculate the mean as we all learned to do in school).
Sonas is a competent statistician, and if he'd wanted to argue that "average ratings" have gone up he could have done either of these. He didn't.
My guess is that Sonas didn't do this because he (like me) is more interested in the pointy right end tail of the bell curve than the pudgy middle.
That's fine -- people can be interested in whatever they want -- but if he (or anyone quoting his article) is going to claim that there is systematic inflation he ought to be looking at the entire rating system. At the very least, Sonas should relabel his chart from "Inflation of entire [sic] rating pool, since 1975" to "Increased ratings in top quartile, since 1975".
The significance of using only data from the right hand side of the bell curve is that what he cites as evidence of inflation is exactly what you would expect when the population (of rated chess players) increases: with an increase in the size a population described by a bell-shaped curve, each ranking position in the top-half of the pool must correspond to a higher rating.
Sonas tries to get around this by saying "I don't think we are adding in players anymore at the right edge; I think we are adding in players at the left edge, via inclusion of new provisional players or via the reduction of the rating floor."
[It's question-begging for him to say "I don't think we are adding in players anymore at the right edge", but let that pass.]
I'll admit that most of the increase in population has happened at the left end of the bell curve (esp. when FIDE started keeping ratings below 2200), but not all of it. First, almost all the top players were once low-rated players. Second, the additional cost of FIDE rated events makes it economically rational for clubs and federations to hold FIDE rated events only for their strongest and most dedicated players; so FIDE ratings would disproportionately represent stronger players at the "right edge" of the curve.
Bonus Troubling Note
Even a brilliantly complete statistical analysis of FIDE rating data will not be able to explain "inflation" because FIDE ratings are not a random sample of all tournament chess players, they are significantly self-selected.
They increase the cost of playing in a tournament because players have to play FIDE-rating fees and pay for a FIDE-certified arbiter/TD. This creates a disincentive to run and pay to enter FIDE-rated events, and this will tend to select players for FIDE events who are either:
a) indifferent to the extra cost,
b) want a FIDE rating no matter how low it is, or
c) think they are strong enough to have a chance to get either FIDE titles or at least a 'brag-worthy' FIDE rating.
My guess is that groups a and b above are so small that most TDs and clubs do not hold FIDE rated events except for their best and most dedicated players. As a result, FIDE ratings represent a disproportionate number of the stronger tournament players in the world.
Why do some think the purpose of the rating system is to measure past players with present players?
I hope I didn't give that impression in my post. Of course it isn't.
But still, Arpad Elo did express concern that ratings should ideally consistently reflect true playing strength over time, and suggested some methods to address this. The Elo system, in and of itself, does not address this.
I think there is a general feeling amongst players that an 1800 rating should say something about the quality of one's play, and that is a normal human desire. The Elo system, in and of itself, cannot do this. It cannot even guarantee that an 1800 player in Canada today will be of the same objective ability as an 1800 player in Brazil today.
1-To me, inflation (in the context of this discussion) is just the phenomenon of ratings going up with time - i.e. exactly what we observe with FIDE ratings.
Well, that isn't what Arpad Elo meant by the word. And it isn't, to my mind, even a coherent idea.
So, basically, I agree with the elements: yes FIDE ratings are going up and yes, players are getting better with time... but I'm not convinced that the two are intrinsically related.
Yes, but the only actual evidence given in this thread is that the rating a player achieves for a given level of actual chess playing quality, is going DOWN over time, not up.
You are right in that there is no actual inherent relationship between Elo ratings and the actual quality of play. But today we have computers which, while not perfect, can give a pretty objective measurement of the actual level of play in a given game.
If 'better players' = 'increase in total amount of rating points in the pool', then somebody must be able to explain where these rating points are coming from??
That would be an inescapable consequence of the playing strength of players entering the rating pool increasing over time. So long as more new and stronger players are entering than leaving, the average rating will go up.
Tweet
It will take more time. Though it is possible to see that the 'mass center' moved to lower ratings in 2013.
If you'll compare to its population it will be laughable. I was surprised too :D
Comment