Why is there rating deflation on Lichess? • page 1/2 • General Chess Discussion • lichess.org

Curious about the phenomenon of rating deflation on Lichess.

Looking at the the @maia1, @maia5 and @maia9 bots as a "standard candle" benchmark, they have all lost 100+ rating points in the last 3 years.

I assume (happy to be corrected) that ratings have therefore deflated on Lichess by a similar amount.

Why?

Curious about the phenomenon of rating deflation on Lichess. Looking at the the @maia1, @maia5 and @maia9 bots as a "standard candle" benchmark, they have all lost 100+ rating points in the last 3 years. I assume (happy to be corrected) that ratings have therefore deflated on Lichess by a similar amount. Why?

manmanmanmanmanman

#2

hmm found a long thread here on the same topic, the trend seems to continue since this was posted:

https://www.reddit.com/r/chess/comments/w46kht/my_unscientific_case_for_rating_deflation_in/

hmm found a long thread here on the same topic, the trend seems to continue since this was posted: https://www.reddit.com/r/chess/comments/w46kht/my_unscientific_case_for_rating_deflation_in/

CM Sarg0n

edited

#3

The inflation just has got reversed.

manmanmanmanmanman

#4

when was the inflation, pre 2021?

mkubecek

#5

With the ratings often fluctuating by as much as 200 within few months, how did you come to the conclusion that the ratings dropped by 100 in 3 years? Did you use some form of smoothening/averaging the values or did you just pick a random day 3 years ago?

manmanmanmanmanman

#6

just eyeballed the graphs, looks fairly obvious? maybe that's too naive...

Alientcp

#7

@manmanmanmanmanman said in #1:

Why?
Not familiar with the case, but overall, rating is inflating, not deflating.

In any case, I suppose there is the same niche of people playing against those bots. The bots have maintained the same strength and were beating new players, but the players got stronger and keep playing the same bots, eventually winning and taking their points.

Maybe thats what happened.

@manmanmanmanmanman said in #1: > Why? Not familiar with the case, but overall, rating is inflating, not deflating. In any case, I suppose there is the same niche of people playing against those bots. The bots have maintained the same strength and were beating new players, but the players got stronger and keep playing the same bots, eventually winning and taking their points. Maybe thats what happened.

manmanmanmanmanman

#8

what's your evidence for saying rating is inflating?

there must be some data somewhere in the Lichess archives which would prove the point one or the other. Not sure how to access it though.

what's your evidence for saying rating is inflating? there must be some data somewhere in the Lichess archives which would prove the point one or the other. Not sure how to access it though.

Cedur216

#9

Not sure why these bots are benchmarks.

Katoh1

#10

@Cedur216 said in #9:

Not sure why these bots are benchmarks.
Well my guess would be because compare to humans their strength probably remains the same over time.

@Cedur216 said in #9: > Not sure why these bots are benchmarks. Well my guess would be because compare to humans their strength probably remains the same over time.