Rating Inflation or not?

Hi, my Rapid rating has recently been increasing but I don't feel like I'm any better.

I understand vaguely that GLICKO system puts us somewhere on a normal distribution. So whilst it is true that I am better than 70% of the players on Lichess, when last year I was only better than 50% of the players. I wonder now has my chess objectively improved over this period?

If there has been an increase in users of Lichess, it's likely they are lower skilled (because they are new), which would correspondingly push my ELO up without my skill level actually increasing. I'm intrigued if this is the case?

I'm not complaining this is the nature of ELO systems, I'm just wondering if my theory is possible, or; as I hope; that my theory is debunked by your goodselfves and I can be content that my chess skill has indeed objectively improved.

My experience is Lichess rating sucks as a measurement of skill. Maybe if you play a lot of games it starts going to the right direction but otherwise it's just noise.

@vvaaaqqqqvvaqvvvvaqv said in #2:
> My experience is Lichess rating sucks as a measurement of skill. Maybe if you play a lot of games it starts going to the right direction but otherwise it's just noise.

How does one measure skill then?

@ywb2 said in #3:
> How does one measure skill then?

Not everything is measurable. It's like a quantum state.

rather, ratings are eroding (although slower for the elite who avoid lower rated players) as a well-prepared lower rated player is more than a match for a coaster. Hence chess960, with its ridiculously low ratings (while bullet players are called 3000 in spite of often missing simple tactics)

@ywb2 said in #3:
> How does one measure skill then?

I used the same chess playing programme for years, so that I could compare if I was able to win most games on a particular level. But it was a very slow improvement, as computer style is not what you play and practice normally against a human player. It is not the best measurement in the world, but it gives some stable information on how you perform against that software.

@pointlesswindows said in #6:
> I used the same chess playing programme for years, so that I could compare if I was able to win most games on a particular level. But it was a very slow improvement, as computer style is not what you play and practice normally against a human player. It is not the best measurement in the world, but it gives some stable information on how you perform against that software.

That's what I'm wondering, is there any way to objectively measure my improvement (if any) over the years, i've played thousands of games on Lichess and they're all recorded...

>>is there any way to objectively measure my improvement

you could compare your "accuracy", this would show how your games compare with The Perfect Game. Nowadays e.g. with computer analysis we can't really compare top players playing closer to perfection to top players from previous eras. Chess standards have objectively improved, so just maintaining your rating normally shows progress.