Well put. Obviously there will be outliers but on average a "meaningless" online rating is a pretty accurate predictor

And one's ability in blitz chess doesn't correlate at all with one's ability at longer time controls, right?

On the chesstempo forums, the site founder has found that a player's blitz tactics performance correlates with his FIDE rating better than a players standard rating performance on the site. Basically, you see the move that you are going to play within the first 30 seconds of looking at a position a high majority of the time.

There's a lot of variance in standard ratings on chess.com because of the high cheating rate / ban rate, high abandon rate, the low # of games played / high RD, the small and isolated player pools, and all titled players starting at an over inflated 2000.

Blitz is probably the least variable of the 3 live time control groups. But even in blitz 3 minute games differ a lot from 10 minute games.

A realistic 95th percentile +- variance interval i'd expect chess.com to have is about 300. (I'm only guessing on the variance, not what X in FIDE averages on chess.com)

So hypothetically, if two players are 1700 blitz chess.com, they likely aren't gonna differ by more than 300 points in FIDE / USCF ratings between each other.

I played on Tiger Woods PGA golf yesterday and won a tournament versus Jack Nicklaus and Ben Hogan. How do you think I would do in the US Open next month?

Obviously, there may be some correlation between Chess.com and real USCF rating, but people delude themselves into thinking quick games on Chess.com are anything like playing in a real chess tournament.

ICC maybe has a bit more credibility.

Give me a relatively complete chart of the average US Open performance of all people who play Tiger Woods PGA Golf and I'd be willing to hazard a guess. And probably it would be a very accurate guess at that, as I'm pretty sure that such a chart would show the average result for 99.999% of people as "N/A", which would then be my prediction of your result. Go qualify for the Open and prove me wrong.

Why would you say ICC has more credibility than chess.com though? Both require you to make chess moves within a set time limit, they don't have fundamentally different rules for the game. So a rating at one of them measures roughly the same set of abilities that a rating at the other one measures. Those abilities are not *identical* to the ones needed to succeed, but there is a fair amount of crossover (a good chess move is still a good chess move, regardless of the conditions in which it's found) that there's some decent correlation. And we have knowledge of the OTB rating of enough of the players within both pools to have some idea where that correlation connects in, numerically.

So all that's left is sample size. And both sites have a large enough player base to give us that. Chess.com and ICC have equal "credibility" when it comes to whether or not their rating systems can be used for statistical analysis, with the end goal of predicting the OTB rating of a player with a large sample of online games. Margins of error may come out different, since they're different data sets, but that's not really a measure of "credibility".

firefight answered the question.

Very little cheating on ICC, very few players banned, and only a very small percentage of games are abandoned. (Much) bigger player pool and all ratings are earned.

I think there's a strong correlation in countless cases but I can also think of several cases where it is quite off. I thrive at deep analysis in tournament play, taking over 10 minutes for one move quite often.

Go play some tournaments

Oh man, epic Rei trolling response ITT.

Nothing rattles him more than that kind of torlling, besides maybe actually playing a game of chess.

The live ratings compared to average online blitz levels are higher, similar to online standard ratings, up to 500 points higher. It is hard to say much about 3 minutes chess at that level and it depends often on age. I used to play 1 2 and 2 6, as being perfect for me, but as little as i play i take 5 10 these days or play with no time, or it would be low quality. I could not survive w 1 2 against anyone these days but was very competitive with it ten years ago.

Being good with faster blitz just shows about the fast skills, there not necessarily being other skill. Just like in IQ tests. Most players who play faster blitz are mostly young players. They have no interest to think it out and win by skill but they play for volume. Taking those skills even to slower blitz games and they perform weaker and there are mostly two results. They get outplayed or win on time or the opponent makes a blunder. For this reason, i play blitz only vs computers as they are more mature, i.e. they usually play better moves. It correlates with multitabling vs deep one to two tabling. The latter is where my edge is. At top fide level the ratings correlate almost perfectly no matter what the speed is.

You can try it yourselves and just play moves, even useless moves vs average or weaker competition and until u squeeze yourselves in a tight position, it makes little difference as long as it is not a mistake or speed of advance being important. So, that is fast blitz and additionally they tend to come with immature moves that have some power only because it is faster blitz while in slower blitz they and less meaningful moves just give the edge to the opponent who thinks it through.

no matter the sites, if you can reach 2000 in 3 or 5 min games, it has "some" meaning...
You just need to cut off around 2-300 points usually for slower games.

Has for OP question, you are probably behind couple of millions players...

I figured only in russia and old USSR, at least 1 million can play better then 1700.
Each time i encounter one of those player online or in a coffee shop, they always know how to play to not get crushed without some efforts.

Haven't read every post in this thread, but did see another thread claiming people cheat quite often on that site, Chess.com. If that's the case, and OP does NOT cheat, then his rating on the site should actually be higher. If anything, the ratings on the site are only 'meaningless' because some of the skill levels should be higher due to the fact others are cheating.. (Unless that is somehow accounted for.)

I just missed this thread. Are some people really suggesting that an internet ranking means nothing? There are only two important abilities dont get measured in online chess that are measured in actual tournament games, ability to deliver under pressure and ability to reach a peak level in performance. The minor ability of premoving is not useful in tournament play as well. There is just no way all those factors make up 300 points of rating performance as some are suggesting.
If a margin of error of 95% confidence was higher than 150 I would be really really surprised. I mean if we play in the internet casually I win 75% of the game but if you take it super super seriously I only win 50%? No ****ing way that applies with more than 5% of the population.

I play games on chess.com with two accounts ( dont acuse me so I dont get banned) one I use for more serious play where Im rated 1640 or something and another I use in my phone when Im on my bed where I play when Im tired where Im rated 1590. Its only a 50 point difference.