I wonder if anyone else has had this feeling in Tactics Trainer...

One important thing is you have to train recognizing patterns, with a tactic book for example : fork, skewer, mating net, discovery attack, X-ray, pins...

I only set Tatics Trainer to rated (timed) when I want to get an idea of how much I have improved. So, I will do 25-50 rated puzzles and then STOP, record my rating, and then RESET my tactics trainer history. Then I will go back to doing unrated puzzles and solve them at 100%, no mistakes, taking as much time as needed per puzzle.

I will spend 99% of my tactics training sessions on unrated puzzles, both here and at Chess Tempo, and ONLY do rated puzzles every other month, once, as a test, to check my progress.

So, I suggest seeing the unrated puzzles as practice, and the rated puzzles as the test. In school, didn't you do far more homework (practice) than quizes and tests? See tactics traing as going back to school.

Last, I have improved a lot more in tactics by doing it the old fashioned way, without a computer. I go through ONE puzzle book until it's all worn out, and the basic patterns are burned deeply into my brain, and then I pick up another tactics puzzle book. Rinse and repeat. I don't think there is any good substitute for high quality tactics puzzle books. Dan Heisman talks about the problem with doing 950-1200 rated puzzles on chess web sites. It's not as efficient as getting a few really good tactics books that cover the basic patterns we NEED to know.

They could be vague and just give it a theme. They don't have to specify queen hanging or queen sac, they could just say sacrifice, mating puzzle or hanging piece as it applys per puzzle.

Another common confusing situation is when you think you see a tactic to capture a piece and you don't see a forced mate, many people play for the material, but the puzzle could be looking to limit the king's escape ability, setting the stage to try forcing mate there after you established the mating net.

What if your queen sac was a 5 move victory and you could also fork and when a rook or queen ? If it is a puzzle looking for a fork and capture, you might get 30 seconds, but if you are calculating what looks like a 5 move checkmate with a sac involved, you most likely won't figure it out in 30 seconds. So if you think it looks like a sac mate theme, but it isn't and they are looking for the fork and capture, you are screwed.

Forgive me for throwing snark into this; but is not your score getting higher a sign of improvement and the score dropping just the opposite?

I snapped off 4 tactics problems after starting the thread (last night and a few minutes ago) which got me back to around 1240. Then I missed two, back around 1220. 

Or maybe it wasn't actually a sign of improvement, just luck. I'm not sure.

I dunno, the more I get wrong, the higher my rating goes 

I would like to learn those patterns so that I am improving my play against soundly played chess, instead of figuring out replies to bad hope chess. Do you know where I can get a glimpse of those needed patterns ?

BTW, I do the same thing that you have done to some degree with timed and untimed. I think it puts the emphasis on learning instead of attaining a rating. Also I wonder how often players that complain about their progress, actually keep retrying the puzzles they failed at until they figure them out and why the ideas work, or if they move on to the next in disgust, missing an awesome learning opportunity ?

Oh I certainly get disgusted, but only move on without retrying it if I accidently hit the button. I will admit I do them multiple times, and sometimes get them wrong on second, third, and fourth tries. I do use them to learn. Or at least I'd like to think I am.

I've read some of the comments here and moved to TT and turned off the time. I also felt so daring as to move the ratings up way beyond me. This is under the assumption I should learn faster with more advanced problems but not burdened with time constraints. 

I now do my TT occasionally......it drives me nuts too....but I have to admit they help me

twice I got all correct and my rating decreased

The thing that perturbs me about how TT puzzles are rated is using the players' ratings as a measuring stick. If you consider the following idea, I think you will begin to see what I mean.

One day I started doing random puzzles in unrated mode for the first time. I didn't realize how much the rating fluctuated per puzzle in this mode. I noticed that unfamiliarity seemed to have more to do with my ability than the ratings given the puzzles. 

I was failing many puzzles but getting most of them on the 2nd or 3rd try. When I looked at the ratings for them I noticed a few 3 move puzzles in particular I got in two tries were , one in the 1300's and I solved another in the 1900's .

The crazy thing is my rating is usually around the low 1500"s and goes up or down a bit. This and the next idea I am about to present, have me questioning the validity of the rating system.

For example, I think a 2 move puzzle at a given rating is more likely to get solved faster and more accurately, regulary than a 5 move puzzle of the same "given rating " under the current system. In theory, if they were of equal difficulty, the solve rate/average score, by the same rating group should be virtually identical. Yet, I know for a fact from my results, I fail far more often and get lower scores on 5 move puzzle than on 2 move puzzles of the same rating.

I think the problem lies in that, each play at a given rating has different weaknesses for each different catergory.In spite of realizing the scope of the problem, it still leaves me at a loss to either find a way to tweak the current system or provide a better alternative.

Your thoughts ? I may make a whole thread dedicated to this idea...

The best piece of evidence I see that testifies against your idea is that you have to remember the previous move as you calculate the next one and you have to also consider all of the possible consequenses of each move down the line in order to see if a move is safe or desirable. Memory is everything in calculation. Without it, you have nothing but a blank stare...

That's just a factor of the proportional rather than right/wrong method of scoring. You're more likely to fail and get an increase in rating on a 10 move puzzle, especially if the last move is hard because you can still score 90%. Some stalemate problems are like this, a lot of natural moves and people didn't realise they were playing for stalemate so it has a relatively low rating and low pass rate. Whereas a 2 mover the best you can score if you fail is 50%.

Is it also possible that the different genres of tactics can sneak up on a person and make them lose points because they've become acustom to certain positions in TT? 

Those are really good points that would certainly have to factor into any other scoring system that could possibly be made. Also I noticed, thought a 7 move puzzle could still yield you 85% if you got 6 moves right, but that is also conditional on how fast you do it. The way the timer can affect the score, as well as, the percentage based on number of correct moves, and the likelyhood of getting all of the moves correct going down by percentage, as moves are added, most likely, all would have to be factored in. The moves could be obvious, but often, atleast one move is no give me in a long sequence.

Maybe I should have tried harder in math when I was younger. Trigonometry just doesn't cut the mustard to build and solve algorithms very well...

Another way of looking at this would be that a problem with 3 moves @1700 could have the same difficulty in terms of yeilding an equal increase in aptitude rating, as a 5 move 1550 problem...

Your Rating: 1184

Today's Attempts: 22 (Average Score: 43%)

Not doing so well today either. 

I dropped from an alltime high 1656 down to 1494 in a few days...I tend to stay in the low 1500's...it all seems to be relative to my mindset and the puzzles I get to work with, it can make a huge difference. I never realized how much my aptitude could vary before looking at all of this...or better yet, the rating of that aptitude, who knows what it really is, especially that it varies...

I just wiffed on a wide open queen. Tricky Tactics Trainer I'd call it, you go through so many so quick (even in losing points and dropping the rating of problems) you find something so pure and your lessons of missing stuff in the past gets you secodn guessing. 

Strike three, you're out!

Your rating dosen't matter in TT, some advice is to go over the problems you missed and attempt them again and get it right, then they will stick in your brain.

FYI all I don't retry problems when I fail them... I play through the solution. The point for me is to learn the pattern, not to keep retrying until I "get it right".