Rating Comparisons among Tactics Programs and Servers

I am trying to get a sense of how much ratings of players vary among some of the online tactical trainers and software.

Here are mine (and I'd like to hear yours, so I have some data to go off of):

chess.com's Tactics Trainer: 1801 (peak 1801)

chess.emrald.net's Chess Tactics Server: 1528 (peak 1550)

CT-ART: 2150 (peak 2150)

chesstempo.com 1910 (peak 1955)

I haven't started using chessimo yet.  Feel free to mention other ratings you have on other sites or software.  I imagine the biggest difference between these is the degree to which timing plays a role rating and the difference in problem ratings for those that do not use time constraints.

I'd be interested to know how tactics ratings match up to OTB USCF or FIDE ratings.

They don't.  However, under the rating of 1800 or 1900, they there is a strong correlation between tactical ability and USCF rating.  Nonetheless, I know many players who are brutes with tactical puzzles, but can't set up much more than a basic one- or two-move tactic.  In fact, I know someone on here with an insane TT rating, and he's absolute garbage at the game, probably because he's been through most of the problems and remembers the exact solution rather than recognizing the pattern.  That's much harder to do on Chess Tactics Server and chesstempo. 

I think what's the case is that someone who is strong OTB will have a good tactics rating, but a good tactics rating doesn't necessarily mean you're strong OTB.

However, IMHO, if you're crap at tactics, you're just not going to be a strong OTB player, no matter how much strategy you know.

At any rate, here are a couple of data points for consideration:

My son:

chess.com Tactics Trainer: 2300+

chesskid.com tactics: 2900+

USCF:  1900+

An adult master we know:

chess.com Tactics Trainer: ~2550

USCF:  ~2300

It's probably the case that young players are much stronger tactically than stategically while adults tend to be more balanced.

i was 1700 blitz and 2000-2200 tactics and now i am 1900 blitz and still 2000-2200 tactics

Do you have tactics ratings on other servers?

not really as i havent been elsewhere in years

Okay.  I was kind of interested in whether getting better at one tactics server correlated with getting better at all of them.  You know, a lot of peope do the chess.com Tactics Trainer and get bettter at it by memorizing the problems.  That doesn't mean they are legitimately getting better.

Chess.com tactics: ~2250

Chesstempo.com tactics: ~1950 (2000+ if I would actually be consistent)

USCF ~ 1975

Chess.com tactic 1500-1550

Chess365 tactics  1650-1700

USCF                    977

chess.com's Tactics Trainer: 1551 (peak 1609)

chesstempo blitz: 1571 (peak 1576)

chesstempo standard: 1739 (peak 1742)

Chess.com tactics: 2226 (peak: 2390)

Chesstempo blitz: 1869 (active peak: 1952)

Chesstempo standard: 2133 (active peak: 2133)

(OTB: ~1800 KNSB)

Pretty interesting numbers.  I wasn't even aware that 365chess.com had tactics problems.  Thanks.

It's kind of amazing the swingy-ness of my blitz and timed tactics ratings.

My CT-ART rating has gone over 2300 and my standard chesstempo have gone over 2000.  I added one more point to my Tactics Trainer rating (new peak: 1802), but I think my short-time-constraint tactics ratings have stayed the same or dipped because I am only doing slow chess (and absolutely no blitz), in prep for an OTB tournament.

Yes, and I definitely do tactics for that reason, too.  For that purpose, I normally use Chess Tactics Server on chessemrald.net, because each tactic flashes before your eyes and you have 3-7 seconds to get it right, otherwise it is wrong and you lose points.  Getting 1500 tactics done in a day is pretty difficult unless are doing them that quickly.  It took me two (plus) days to do that on Tactics Trainer.  I like to vary the speed.  I thinkseeing tons of patterns quickly is important in a similar way to reading through "reams" of master games, as Dan Heisman and Jeremy Silman advocate.

Nah, just trying to get these patterns down so well that I can seem them without calculating.  Trying to get better.  I'm not sure how many I could do if I spent a whole day on just tactics.  A few days ago, I did 750 on Tactics Trainer in (supposedly) 5 hours.  I think it took a little longer than that, maybe 7 hours.  On chess.emrald.net, I can do 750 in about 3 hours.  TT mixes quality and quantity, but it should be better if it accepted all winning moves for puzzles, not just the winningest moves.

Dang! That's impressive Milliern! I do 5 a day here (it's all I can do), and 5-15 over at CT, depending on how many I get wrong and have to study. Sometimes I find doing that many tedious. I can't imagine going through 1500.

We'll see if it produces results.  I try to mix up slow (untimed) with loose time constraints (e.g., chess.com's TT) and tight time-constraints (e.g., chess.emrald.net).  My slow tactics get better and better, but I am not sure my blitz is getting much better.  It just takes a second or two too long to find them.  Kids absorb these patterns so easily, but I think an adult can still absorb them through hellish hard work.

Yeah, it's super hard.  I think I recently peaked somewhere in the 1605-1620 range, but my rating is no 1495 or so.

I think two things change how quickly your rating changes: 1. if you go on a losing or winning streak, "winning" being solving succesive problems correctly and in less time than they have pre-established for the problem; and 2. if you haven't been on the server for a while.

If you pay close attention, the problems are really good, in that there is almost never a confusing second option that looks good; so you either see the tactic or you don't.  Part of my problem on chess.com's TT is taking a Q when there's mate, or some such nonsense. 

I stopped using chess.com's. The problems seem to be tuned so much that you solve a problem relatively quickly and you get 1 point, then you solve another problem but you take 2 seconds longer and you lose 3 points. If there are so few problems that they can be memorized (as you suggest), that might explain it. I much prefer chesstempo.

The memorization angle that you mention is interesting, especially that you have a friend who can solve problems quickly, presumable due to memorization, but his tactical ability doesn't tranfser to playing ability. It's interesting because memorization is exactly the method recommended by GM Ziatdinov, just get exposure to many patterns that you can recognize quickly. He literally says to look at the answer and move on if you don't solve it quickly.