How is this possible?

Thanks for the info, guys.

One thing about the tactics training that I've noticed... Such puzzles usually don't come up in a real game. (Sometimes, though...yes.)

Saying that, I strive to use positional play/strategy to set up the tactical. One good way is setting up for a discovered attack.

And, although I've gotten up to around 1800 on the rated tactics...I usually get shot down a couple of hundred points and have to build back up again.

I find the tactical training great...to an extent. But it does not help develop positional skills, including openings.

Always something new to learn. That's the great thing about chess. lol

I think it isn`t a good idea to compare ratings like that. (My ones: Online - 2000, 960 - 1814, Live Standard - 2035, Live Blitz - 1702, Live Bullet - 1350, Tactics - 1651 [5 attempts, passed 100%]).

You have to find the best move to succeed when it comes to tactical problems which is something you don`t need when you play a game. Look at the position below (I know it is very simple). If it were a tactical problem there would be only one way to pass (mate in 1) but if I missed the mate in a real game and played one of the other two lines I would still win, i.e. my rating would increase.

Good point.

I know that, for example, when tactical training...one may win the Q but still fail the puzzle, even though winning the Q would probably result in winning anyway.

Also, you may get the right answer but take an extra 20 seconds and get heavily penalized.

None of this answers, in my mind, how someone in the 800 tactical puzzle range, or the 1000 blitz range can be a 1700+ in online. It simply defies my comprehension.

So...I just "suck it up" and move on. lol

And I still don't understand why most online ratings seem to be higher than blitz ratings, and not just here (assuming of course that all the ratings have the same "spread," ie, as wide a spectrum).  Yes, maybe some players take full advantage of all the options under a 3 day/move time control, but that would mean that some other players would suffer from not taking advantage, and thus it would (or should) all even out.

I think along your lines.

But, in the past when I've raised thoughts such as yours, the response is something like: "it's all a matter of the particular pool of players".

Doesn't quite make sense to me, either. But...shrug.

I've found a similar thing with the tactical trainer. I'm not a slow player. In fact, my preference is rapid. Yet, I've sometimes found that I could get the right tactical answer at, for example, half the time required on some of the more difficult puzzles. Yet on some of the lower rated ones, it might take me twice the allocated time.

I have a theory, but I'm not allowed to talk about it here. It has to do with the shape of the ratings distributions and why online chess has a different shaped distribution curve than otb.

Hm, "I have a theory, but I'm not allowed to talk about it here"...could that maybe be translated as:  "Everybody's cheatin'!"? (heehee!).

Why aren't you allowed to talk about it here? because you saw other peoples "Chess is solved with my simple system that I can't tell you about except for some charts that mean nothing and instructions that say nothing" threads and didn't want to get trolled?

I'm still going with the "Everybody's cheatin'" theory.

You're just saying that because it isn't poker.

I think you got it, there's simply not the same spread.  I'd guess that blitz is much more popular in general online, and that titled players shy away from CC (or long live games) due to ever present computer opposition.  So just speculation, but I'd guess blitz pools here and everywhere are wider and stronger on average.

Okay, wouldn't the greater numbers tend to inflate ratings?  Not that I'm exactly a math whiz...so go easy on me lest my big fabric head explodes.

Look at mine, their both around 1600, but my tactics, and standard are low, because I don't use them often.

IMO, the rating distribution curves should all have roughly the same shape.

http://www.mark-weeks.com/aboutcom/aa03a25.htm

This OTB distribution is skewed left, while most of the chess.com distributions are skewed right.

Availability of allowed resources could have some impact on this. Hard to say how much.

Barriers to entry for non-allowed resources are typically higher in fast games than in cc.

Greater numbers would tend to give a wider gap between highest and lowest rated, while a higher average strength would make higher ratings harder to come by.  So you're right that if a pool didn't have many members, its top rating probably wouldn't be that high by comparison.  The part about the pool being bigger in blitz was just guessing on my part, the real point for lower-by-comparison blitz rating IMO would be the average strength being higher.  If blitz is more popular among stronger players, then this would explain why other sites have this too.

Inflated only really happens when you mean they have ratings higher than what they should for some reason outside of what the math can account for... like lots of new players joining, losing, then quitting.  Or maybe if isolated sections of the pool rarely or never interacted.

I know some woman on here who started out rated lower than me, whose is now rated at about 1700. Her tactics rating is still a bit over 900 though.

more people play blitz-bullet so the rating pool is much more dynamic and reactive (i would also wonder what the formula is for each)

really fast time controls especially at the lower rating pools and 1 min people can play for different skill sets like fast mouse and time not really playing chess so hacks can win games. 

Live chess has a great skew because the same group of hacks get crushed in longer games and then quit while more serious players stick around at the top. 

Just a guess. I have enjoyed the CC games more now that i have waded through the hacks at the lower level

I thought most people knew that glicko (the rating system that is used here) deflates over time.  In the Elo rating system, rating points are conserved: ie player A wins 12 points, player B loses 12 points.  With glicko, this is not the case.  Player A can win 12 points and Player B can lose 15 pts.  This has to do with the players having different RDs (In Elo, everyone has the same RD: 0).

In blitz and bullet chess, many more games are played (because they don't take as long to complete) therefore the system loses points faster.  This makes the ratings lower over time.

Please do not confuse "I have a higher rating in x category" as being the same thing as "I am better in x category."  Ratings are only relative to each other (in the same category/player pool) and not relative to an absolute scale.  In some systems, a grandmaster could be rated 1800, and in others 3200.  This isn't a "flaw" in either system.  Expecting the same rating in two different pools is just misunderstanding how ratings work and their design scope.

But, the thing does come down to tactics.

How can someone be terrible at tactics but an ace at anything else?

(Of course, I am excluding speed chess where time is often more important than tactics. Quick garbage moves may actually be a winning factor in bullet.)

The tactics here are timed, so that doesn't really matter.  If you're very low on TT but high in bullet or blitz then I would find that suspicious.  But CC you have tons of time and an analysis board to assist in sorting though tactics.

I remember playing this kid at a club one night who seemed to see in 15 seconds all the tactics I could find in 2 minutes... but that was all he could see apparently, and I only lost once to him (and I always had the better positions).

Older guy at my club I always win with a tactic in the middlegame... but I played him on a site like this for a few months and he founds some great tactics.  I'd even say he was tactically better than me in CC.  I think it's because he's retired and analyzed all day :)