FIDE Rating inflation/deflation

And what unreasonable people think and what they argue in forums is just something that happens. I admire Morphy more than I admire Magnus, but that doesn’t mean that I think that Morphy would do anything other than perhaps get a game in here or there… maybe… probably not.

Nikki uninteresting depends on your perspective. I played chess on and off, mostly off, since the late 70’s.

When I started playing my rating jumped almost instantly into the 1600’s, and there it has stayed.

So it is an interesting question to me. Am I better, worse, or the same as I was when I was a teenager?

Not quite. Inflation or deflation will tell you nothing about your absolute skill because of the reasons that I have outlined. If you really wanted to know if you are better now than you were then, feed your games from then into Stockfish and count the number or errors. All you really know with elo is that you are in the same percentile as you were then.

Now, what we DO know is that players are better today, so if you've maintained the same ELO, you're probably (probably!) better today. But you don't need inflation/deflation to know whether or not that's the case.

Your post confuses me, because yes if I maneuver the same rating, and there was deflation, then I have improved to the extent there was deflation.

Stockfish comparison is interesting too, but it has its own problems. You will make more errors if you face more difficult opponents and game situations…. just look at the error counts of your games on this website.

I have had games with a 20% accuracy, and games with a 98% accuracy. I wasn’t drunk!

Well, plots seems clearly that ELO is decreasig, however I thinlk that it is not conclusive.
More study is needed for a more accurate concusion
The númbers of players also have increased. And while it increases ELO  curve get down.

https://en.chessbase.com/post/the-elo-ratings-inflation-or-deflation

i disagree

Inflation is good for us!!

But that doesn’t matter, right? As you agreed, elo tells you nothing about absolute skill. In my experience, and i am younger than you are, players today at the same rating know loads more theory, and have a better understanding of principles and things like endgame technique than even ten years ago. They do more puzzles, watch more videos, and play several times more games a weeks because it’s so easy now to do so. Inflation or deflation doesn’t measure this improvement in absolute skill, so it won’t tell you anything about whether or not you’re better or worse. Again, you can’t compare elo amongst groups. You just can’t. It’s not the same group.

What has happened to the average rating is not meaningful, as it has dropped due to lowered minimum rating. But the rating of a player with a given error rate, like 12 or 13 centipawns, has dropped at least 100 elo over the last twenty years. I call that deflation, you can argue semantics but what it means is that you had to improve your level of play by 100 elo in that twenty years just to maintain the same rating. I don't consider this to be desirable, it turns players off from playing, but that is the actual situation. I think the same thing has happened to USCF ratings, since the gap from FIDE has not grown, it seems to have shrunk. 

HUHHH its GM Larry Kuafman shogi and chess player!!!

The underrated young take rating from the overrated old...

Centipawn loss is a pretty blunt tool.

A player like Tal who loved to create chaos would have much higher CP loss numbers for both himself and his opponent, but we don't have many people like him at the very top these days.  So one could argue that what they are observing is possibly at least partly due to a change in stylistic preferences rather than simply due to a change in strength of play.

However, by broadening their analysis to include players at a wide variety of levels, they have significantly reduced the effect of that.  At lower levels, you have many more players and therefore a much wider variety of preferred styles of play.

One other point: even if we have rating deflation, we indisputably have title inflation.  Not in the sense that titled players are not as strong as they used to be, but in the sense that more people have done so - partly due to more people making the attempt, and partly due to the fact that there are many more study tools available.  If a GM title used to mean that you were probably in the top 100 players in the world, that is no longer the case.  A modern GM is probably stronger than most of the top 100 players in the world from 1980, but he might not actually be in or even particularly close to the top 100 today.

So the GM title is still meaningful in identifying you as a very strong player above a particular level, but is no longer as meaningful in identifying the best of the best of the best.

@MGleason, the author points that they are talking about broad trends not individuals. When looking at the general population, centipawn loss is a very good way to measure relative quality of play. The complexity issue will even out in a large population. 

Every generation has its Tal-like players and its Capablanca types.

The author’s point makes sense to me.

Regarding title inflation: FIDE has set up a system where at least theoretically the next generation of GMs is stronger than the previous generation. That’s what norms plus ratings is supposed to create.

We know that there are norm “factories” which corrupt the original intent.

But I don’t think there was ever supposed to be a limit to the number of GMs. Max Euwe lamented when he was President of FIDE that he no longer knew personally every GM in the world as he had done when he was World Champion. But he accepted that the GMs had earned their title.

There is always talk of creating a new title, Super Grandmaster. But we already have at least two ways to delineate the very best players.

We can look at the player’s rating. Players over 2700 are generally well known and it is a special club of players.

Another way to separate out the very best is to see who has been a Candidate for the World Championship. That is a very select group that generally has the very best players in the World.

 

We don’t need to rethink the GM title. But we may want to consider codifying who is in the highest echelon of elite players.

Yeah, if they had focused on top GMs, stylistic differences would be a potentially plausible explanation for the trend.  But a large sample of players at lower levels makes that explanation unlikely.