The Chess.com "UNOFFICIAL" Ratings Survey

Ok so I have quite a few additional ratings histories analyzed and I have some solid results to share...

the first important thing is to understand this chart

What this is saying is What the first comment on this thread alludes to...

gaining Rating is harder when your rated higher.... and that if we place everybodies rating difference across, how strong they were.  we see that there is a general downwards trend- with stronger players less likely to post rating increase.

we also see that this effect is particularly evident with a 3 month interval. no single individual stronger than 1380 showed a ratings increase in my survey- even though I looked a people up to 1510.

Next chart...

ok, so here's my current "model" (given the population) of what rating improvement is likely.   Assuming you were at 1000 today, and started playing with average rating improvement- your rating would be climb to a little of 1300 and stagnate.

assuming you were one of those guys that had uncommon rating improvement.   your rating would do much better- getting to 1300 in 3 months and not leveling off until well above 1600 about a year later.

if you had uncommonly low rating- you would really never increase past 1000 even with a full year of blitz playing.

We could specify as lines away from the black line of the probability of attaining ever-higher rating gains...

ok, so what are my conclusions...

• First sample size was still small (only about 60 people) and the studied only looked at people up to 1500.   to do this right, Rating should be propotional to the chess.com database - and describe people at higher ratings.
• Ratings increases are generally small and gradual. 50 rating per month or less.. BUT if you look at peoples ratings histories, you often see will highs and lows in rating.   By looking at a lot of people you average out the erratic nature of people curves... but this is important- because...
• if for example you find yourself a 100 rating up in 2 weeks you probably Haven't experience a real ratings increase of 100 rating.-Instead your merely on a winning streak.  Keep Playing.. and several months later- your "real" ratings increase will be more evident.
• Whats REALLY needed is to have more people hand in their surveys.  I think age, OTB, and study are 'factors' that make it More likely for you to experience the Higher rating increases.  but to what extent I do not know. 

I've thought about studying the peak ratings- I have real ratings for it. but I don't know.  I think it might be important just to increase the depth of the database.  more people, more described ratings. and hopefully more surveys.   thank you to everyone who has participated.   

Ratings, for me, just help me track my progress.

Any new developments in this survey?

about 1680 uscf, 1550 standard chess.com, 1400 chess.com blitz

I wouldn't trust chess players to accurately self-report their study habits.

actually SK.  when I was busy doing the study- very few responded at all.

I would gauge less then 1:20 answered me back, and most claimed to be a casual blitz player.

I think , as a percentage of the chess player, improving players that claim to be studying are a very small fraction of the people on chess.com.

to kk10.

no not really.  it seems that its pretty cool and exceptional to improve past 1200- and no matter how high you are, you are more likely to be losing rating at higher elo ratings.

thats what I discovered.

I study everyday.