Rating Inflation

@petri999 explains it correctly. When more weak than strong players enter the pool, rating inflation occurs. Think about a player entering the pool who quickly drops from 1500 to 500. After his first couple of losses, his opponents get more points than they should. Exactly until his rating stabilized around 500. The reasons given for why more weak players entered the pool (covid and queens gambit) are also plausible.

I have been following my percentiles for a while and my very rough guess would be that in order to be in the same percentile my (rapid) rating would have to be ca. 50-100 points better than 6 month ago.

In my previous comment, the cheating example was just an example of a source systematic 'error', not a implication of its change direction. I agree that it would deflate the pool as described by @petri999 and @Argammon . I should have made that more clear in my post.

Interestingly, rating inflation does not show when you consider most bots. For example,

lichess.org/@/maia5/perf/rapid

I don't know much about how these bots work, so perhaps somebody can explain why the bot's rating does not increase over time.

I don't really know

@Argammon rapid is not popular so how about blitz bots? al maia5 was introduces last spring when the major inflation was just taking place so obviously it cannot have gained rating points as it was one of the new comers

@petri999

Maya's blitz rating improved a bit over the last 6 month but not over the last 3 month, see

lichess.org/@/maia5/perf/blitz

Seems like this indicates there was no rating inflation over the last three month, or I am missing something.

As to your other comment, I do not see why it should matter rapid is less popular. There are still enough players, in particular beginners, playing it.

I don’t know if this has been mentioned 10 times already:

65 < 70

Since the stat you are referring to shows you the inverse (you are better than x% of players) it means 1700 is now „worse“. Or your talking about something different. In this case can you please clarify?

I was talking about this:

lichess.org/@/egeus/perf/rapid

Yes that is true. rating inflation happened last spring. Small deviation are normal even for computer player.

I do not know but I presume most new players enter blitz pool and hence inflation should generally be there if anywhere.

I don't see any inflation

Rating inflation refers to people's rating inflating, that is going up over time, without any change in their ability.

It's not related to percentiles.

If the average salary for the employees in a company is a 100K and they hire a bunch of lower paid people and now 100K is above average that doesn't mean anyone's salary has inflated.

Argammon, the reason that the BOTs don't see rating inflation is because the idea is fictitious.