Is it just me, or what?

I've had the tactics trainer set at from 800 to 1,600 for quite a while now. I've noticed that almost consistently, the 1,600 rated problems are as easy, and sometimes easier than the ones rated around 800 or 900, or so. I've even seen problems rated 1,900 as easy as 1,000. Anybody else noticed this?

It's probably because you don't get the idea of the 1900 rated problem, for example, see the alternatives that might involve traps.

 That's simply not the case.

Well, high ratings are based on people failing the puzzles right?  So it's either what philidor said or it's a government plot :s

I've seen high rated problems on chesstempo that involve taking a hanging knight on the first move... but you have to be sure you can survive the resulting attack or avoid losing tons of material or a perpetual etc. and so a lot of people fail thinking it's unsafe.

So woodshover is saying a 1900 puzzle is easier than an 1000 rated puzzle?

 That one was, but read the thread. I was talking mainly about problems in the 1500 range.

My rating on the Tactics Trainer fluctuates wildly from 1200 to 1800, and I think I know what you mean. Sometimes there are very easy problems rated at 1800 or 1900. Sometimes there are very difficult problems rated only 1200 or 1300. But I wouldn't go so far as to say most of them are like that. Quite a few problems are, for whatever reason, underrated or overrated. But most of them are accurately rated. 

 Not in my experience, when we're talking about up to 1600.

So, you're not exaggerating. You think most of the problems are misrated, not just ten or twenty percent of them. 

 Now that I have the time, I'll explain the 1934 rated puzzle.  I know of course that vety few 1900 problems are this easy. As my thread said, I was talking about problems in the 1500 to 1600 range. I just saw this one because somebpdy else commented on it, and I was notified.  Well anyway, the problem went like this: I had a bishop sitting in the lower corner of the board.  My opponent's king was sitting in the upper corner of the board. I had a rook sitting in the middle of the board between my bishop, and the opposing king. My opponent's queen was sitting a little off to the side. I easily saw that if I moved my rook straight over my opponent's queen, it would be a discovered check, and after my opponent moved his king out of check, I would be be able to take the queen. Now does that sound like it should be rated 1900 something to you? Me neither.

I'm going to address two issues: (1) User specific weakness (2) Rating calculation, which seems to be the issue here.

You may find that a low rated puzzle is so difficult, because you have weakness with the specific motif/pattern. It is good to be able to see what kind of motif/pattern that you mostly miss, to find out your weakness so you can fix it.

I'm not sure how the puzzles are rated here. If it is rated normally like human player (i.e when it is solved by low rated player the rating goes down a lot, when it is missed by high rated player the rating goes up a lot) then there will be issue regarding PROVISIONAL rating, i.e. when the puzzle itself has not been solved by enough user (new puzzles in the data bank for example), or the user itself doesn't have established rating (many players below 1500).

Knowing that the software is not free, it means the user is limited and may not be sufficient to establish a true rating, compared to CT for example. Therefore, it is probably possible to generate a program that will "measure" the puzzle rating by solving it. But it is too advanced imo, when you see how the programmers are struggling with the new database.