Why do tactics ratings fluctuate so much on this site?

Anybody know the reason why the same problem can see rating upswings and downswings in the space of a few months? Is the site changing its algorythm or something? It seems that of late the ratings are higher than a few years ago.

I find it difficult to track my progress with these constant fluctuations? Is my 1860 now really worth more than my 1560 last year?

Mine is lower than it was a few years ago. But, yes, the site occasionally tweaks the algorithm.  The last change I know about was the guaranteed 1 point if you completely solve, regardless of time taken.

True, but I didn't notice such an upsurge in rating then.

I agree that is a good change. But oddly enough I have noticed my rating drop well over a hundred points in a short amount of time ever since I started taking more time to solve the tactics problem. So either I'm getting dumber or there were some other changes also made to the problems.

I don't do nearly as many tactics as I should but I'm only off about 100 points off my high since my last reset, a year and a half ago. About 200 from my all time high.

Adding tactics to the pool can impact ratings while, until the volatility of those decreases. Can't say if that happened or what. But, if you are taking more time now, you may be passing more but getting less for the ones you solve. A few misses can put a downward pressure on ratings.

Same thing can happen with studying chess in general. Rating/results can drop while you try to integrate new knowledge and processes into your play.

I just try to compare my rating to that of players who I consider good and who do tactics trainer problems regularly.

I compare my rating to other tactics sites. Only on this site am I so volatile (over 1800 in November, then under 1650, and 1900 today).

But it is especially the problems' ratings volatility which I find disquieting, looking into the comments it seems many problems are now several hundred points higher than in 2014 or 2015, but nonetheless lower than back in 2010.

The way this site handles rating for the tactics has a lot to do with why. I haven't looked at other sites, recently, but I'm guessing most of them have a set rating for the tactics and here they fluctuate based on player passing and failing them.

Probably isn't a huge driver of swings, but surely must impact things.

I sometimes question the method of determining a puzzle's rating.  Periodically, I find myself trying to solve a complex puzzle with a 20-30% pass rate and a rating of 1400-1500.  These problems have had at least 10,000 attempts, so it's not that the problems are new and have a large ratings fluctuation.

The best one that I saw was a problem that had a 15% pass rate and a 1700+ rating three years ago.  It still has a 15% pass rate, but the rating has dropped to 1450.

I've come to the conclusion that the rating is meaningless.  I just try to solve the puzzle and make my own judgment whether TT is helping me to improve.

I've just checked my recent TT history.  If I miss a problem, my rating drops 9 or 10 points.  The largest increase that I have gotten for a 100% score is 8.  The really strange thing is that I sometimes get a larger rating increase for a lower rated (relative to my rating) puzzle  than I do for a higher rated problem (100% score in all comparisons).

Another strange thing:  When my rating dropped 10 points, the puzzle's rating was higher than mine.  When my rating dropped 9 points, the puzzle's rating was lower than mine.  Logic says that when I lose to a higher rated opponent (in this case, the puzzle) my rating should drop less than when I lose to a lower rated opponent.

And today some guy has achieved a 6872 rating out of the blue. Oh well, I guess I just shouldn't use the rating as a benchmark for progress anymore.