I've heard it all by now, super inflated, a little inflated,
pretty much equal, .. . Who can give me some arguments
behind their thoughts/opinion on this matter? One observation I have, is, how come most ratings of FM, IM, .. rated players are, in general, very comparable to their Fide - rating.
The ratings are definitely inflated. You have players such as
@BahadirOzen,
@BeepBeepImAJeep,
@JusticeBot, and many more that have ratings similar to super GMs. Even my rating cannot be compared to my actual one. I am a 1807 in real life while on here I am a 2125. Definitely inflated.
I know its a dry answer, but probably it is best to look into "better than x% of players" figure and consider the nature of the players in the pool, which is one of the biggest chess sites on the net, but a little poor on the heavy-weight players side, like titled
players.
Other than that, I think any rating system is logical only in itself,
and not comparable to others, since we can shift the whole pool's rating by the same amount with no effect and ratings will
change without change in quality of play, if a specific range of players were introduced to rating system's pool of players.
The ratings here (as on any other chess site) are only relevant for comparing players to each other within this site. It is not possible to have them reflect the player's would be FIDE strength.
Can the perceived inflation simply be due to the fact that as Lichess becomes more popular, the proportion of less experienced players increase? The first people who knew about Lichess and created accounts there must have been more experienced players on average than the new subscribers today. That would certainly increase the percentile of a given player, but probably also the rating. If so it's not really inflation, it's that people of a given strength really become better in relative terms within the pool.
A very good paper "A Comprehensive Guide to Chess Rating" can be found here.
http://www.glicko.net/research/acjpaper.pdfInflation might partially come from android app where one's rating is about 100/200 pts lower due to losses on time that would not occur on a PC with a mouse.
That's what I experience personnaly.
#1 This "controversy" is anecdotal unless you establish some sort of baseline, such as:
* Objectively define how strong 1000, 1500, 2000, etc. are -- nobody has done this
* Compare ratings against previous ratings (measure inflation) -- nobody has done this
* Compare ratings across pools -- an immensely laborious and complicated task with questionable results
Unless someone bothers to establish a baseline, no "corrective" action is possible.
u must know that starting rating on site is 1500 not average of 9 opponent +- your score
