Everyone's approach is going to be different and there have been literally dozens of threads wondering the same thing. Some take an educated guess and some don't. There are days when I get stubborn and stay on the same problem for a very long time rather than guess and feel happy at the end that at least I got my 20%. Other days I just give it my best shot without seeing the problem all the way through. I recently erased my 21,000 problem history because I was up in the 1800's and I'm just not that good. Every week or two now I reset to force myself to climb the rating ladder by reviewing more basic problems. Everyone has to figure out their own way of dealing with the TT. OOPS. Just realized your question was less theoretical than about the numbers. If the time runs out, then the best you can get is 20% for getting it right and the usual 0% for a wrong answer. So yes if it is a multipart question and you get like the first part right and not the second, you will score higher for guessing within the time allotted than you would if the time runs out and you get the whole answer correct.
Tactics trainer question
Thanks.
I think I'll try thinking the problems through, however long it takes. Whether that produces more points or not I think I'll feel more comfortable grinding through the problem in my slow sort of way.
My answer from another thread:
"Honestly, with the way the TT here is set up, I think it lends itself best to a brute force approach. i.e. attempting as many as possible rather than trying for 100% accuracy.
Play by the clock and if you don't see the idea pretty quickly, make what seems like the best move to you.
The ones I get right it's because I see the pattern almost immediately, and it's just a case of checking for any defensive resources the opponent might have and making sure I have the right move order (if two moves look like they may be interchangeable, make the effort to find why one may fail).
If you do fail, play over the moves and check that you understand it, and why your move doesn't work, if necessary looking at the analysis and source, the user comments or as a last resort an engine. It's amazing how much you just pick up the patterns you get wrong, my progress there has been mainly trial and error.
Use books to practice tactics you don't already know, use TT to reinforce the ones you do. Not everyone agrees with this, a lot say ignore the clock but I think you need a mixture of long study of tactics to learn new concepts and quick tactics just to reinforce, which I think the system here is better suited to."
http://www.chess.com/forum/view/community/how-to-use-tactics-trainer-effectively
I spent 2 minutes on one to-day and dropped 12 points. Average time taken one minute apparently with just under half getting it right.
It felt better taking the time to analyse it out. I am not sure that being one minute faster to work out a tactic would often matter in traditional time limit games, especially if you are like me and make sure you don't have to play a time scramble. Of course it is critical to five minute chess (or fasyter) but as fast play does not suit me I doubt I'll ever feel the need to train for the skills needed.
Yeah, but it means you don't get to see as many patterns in the same length of time. Plus you will be stuck on easy problems because you're getting docked points for taking so long. I don't see the sense in using this tactics trainer if you plan on ignoring the timer, just use chesstempo.
Conscious of the clock ticking, if I can't see the answer I usually have a stab at the answer figuring that I might as well give myself a chance of getting it right before using so much time that I lose points anyway.
But is that poor strategy? If I keep thinking until I crack the answer will I lose as many or more points because of being slow as I would for getting it wrong or will I lose, at most, the same number (or less) than it costs to get it wrong?