puzzle rush score (5 min) vs OTB rating strength ?

I have a 22 and my uscf is 900.

I have no idea why I have a FIDE rating of 1300 and have maximum score of 32 in puzzle rush. Not going to complain though.

Exactly. Once I get one wrong I lose my way since I'm still focused on the one I got wrong, which has now disappeared from view

Same.

20 - 1500

15 - 1200

10 - 900

5 - 600

These are my estimations based on most frequent high score. So, the idea is that a puzzle rush high score is useful to estimate playing strength if that high score has been achieved multiple times... say 5-10 times.

I typically get 18-22. What is my rating with ur scale?

yeah, if you achieved 22 high score 10 times, then your rating estimate would be around 1550-1600. 18 high score would be 1300 rating. 

ive gotten 22 like 3 times or so

Another method would be to take the average of a player's top 21 high scores and then use the above estimate sheet to arrive at rating estimate.

At most, this gets you an estimate of online strength, unless you are talking about just the very top players where you can get an actual OTB FIDE rating.

Even for online strength, the number is pretty meaningless unless you give a measure of the standard deviation or some other statistical measure of how predictive your results are.

well, to get that you need a sample of players. So, let players post in this thread, their avg of top 21 puzzle rush high scores and let's see if it correlates their playing strength. 

And I don't see any difference between online playing strength and OTB strength unless those players are using some extra assistance during online play. (or they are not used to 3D board).

Online strength and OTB strength correlates only weakly. Different rating pools, different playing conditions, different levels of how serious they are taking the games, etc.

Puzzle rush once/day and unlimited attempts are completely different animals.

I think that the correlation is fairly linear:

600 elo -> 10

900 elo -> 15

1200 elo -> 20

1500 elo -> 25

1800 elo -> 30

2100 elo -> 35

2400 elo -> 40

2700 elo -> 45

3000 elo -> 50

These are according to Chess.com blitz ratings, roughly. Obviously they are just averages.

I think Puzzle Rush more accurately reflects Bullet rating rather than Blitz rating due to the time pressure. I normally get 30-34 but I also mouse-slip a lot.

Great example of the point I was making about playing unlimited attempts vs. once/day.  Your average score is 15.68, but your top score is 34, in over 2000 attempts.  Your average time spent per move is maybe 5 seconds...which makes sense, because you can just keep trying over and over so why not guess faster?

My top score is 21, my average score is 14.30, and my average time/move is is 13-15 seconds, which undoubtedly reflects my nature, but also reflects that I am only able to try once/day.  My longest streak is 20 correct in a row, one less than my highest score.  But I have 63 attempts, and only once per day, so there's no way to get into a flow.

Clearly you are better than I (which is why it's not a perfect example), but the high of 34 vs. 21 is somewhat deceptive.  The 15.68 vs. 14.30 is more accurate, but also not correct.  In terms of OTB ratings, I agree that clearly puzzle rush correlates much better with bullet and blitz ratings, but I think it would prove a fairly poor predictor of OTB ratings in standard time controls.

P.S. I don't really care about my ratings on chess.com, but I don't want puzzle rush to become some e-peen measuring stick in general.  We already have enough of that on the forums .

Sorry to interrupt this great and very useful thread, but @ndb2010: who is this?

My two cents: puzzle rush average over the past 10 attempts would be a much better predictor of playing strength than one's personal best. 

HEY! My high is also 27!!! I'm as good as a 2047!!!!

Guys, old-fashioned Tactics are better at estimating your strength. I f you get a 2000-rated tactic right, it proves you are as good as a 2000 in Tactics.