Hi Jamal. I wrote a long reply earlier but it disappeared into the ether like a message in a bottle. Who knows who is reading it now. Anyway, here is the gist of what I said.

If your only goal is to improve your TT ranking then you may need to gamble a bit. If after a few seconds you are pretty sure you have the solution then go for it and try to score lots of points. (But don’t blame me if you get a big minus instead.) Especially above about 1200 level problems there will often be a catch. E.g. You see a straightforward knight fork immediately but there’s actually a tricky mate in three there also. So if you go for the knight fork immediately you lose loads of points. But if you are going to gamble, bear in mind the following.

• From memory, I believe you get an 80% score if you solve within the average time (the halfway number on the time bar). So is it really worth gambling in that first period? If you are sure you are right then go for it, but for the sake of a couple of points it may be worth spending an additional few seconds checking that you haven’t missed anything.
• Once you pass the bottom number on the time bar (twice the average time) then it’s hardly worth gambling either, because I believe you will get a 20% score for getting it right, however long it takes you. For me that means the difference between about -7 or -8 and something around -14 to -17. So I tend to treat it that after the time bar has run out then the pressure is off and I can take my time to try to limit my losses.
• So that leaves the bit between the average time and twice the average time where you are losing points quite fast, and that’s where you have to decide whether to gamble on an educated guess or persevere in trying to get the puzzle right. I used to do more of the former but as the tactics I take get more involved I tend to go more for the latter.

Now I said at the top that all this applies if “your only target is to improve your TT rating.” I would just say that you ought to think about other things also, and especially you should watch your pass rate. As an illustration, my brother also does TT and his rating is very similar to mine, but his pass rate is around 80% and mine is around 63%. This is because he does not look at the time at all; he just keeps going till he solves the thing. Whereas I will sometimes gamble when the clock is ticking down. But as I say, I’ve been taking longer and longer because, to be honest, I would rather have his stats than mine. As a result my pass rate is creeping up and my rating hasn’t suffered (except on days when I’m rubbish!)

The only other thing I’d say is that my rating went up at least 100 points when I became a paying member and started doing 25 tactics a day instead of 3. And it has stayed there. That was a year ago and for most of that year my rating has been between 1200 and 1400, whereas before that it was usually around 1050 to 1200. I did once hit 1300 on 3 tactics per day, but then over the next 6 weeks I crashed to 1050 again (that’s a big fall on just 3 a day). So that’s when I made the decision to take out a gold membership. For me it’s been well worth it.

Good luck!

"this rewards those who just try out the most "inspiring" move without calculating. And that's not how it works on a real game." yes plutonia but only in blitz and lightning and i don't think you will find many brilliant games in those categories. i have seen bad chess including missed mates even in in blitz at very high ratings. what's the point of playing bad chess just to play quickly? 

my goal is to break through 1200 in the tactics trainer. i am very thankful for josechu's advice and will use it in my quest.

Plutonia - I think an alternative TT scoring system that awards zero for a slow but correct answer should also not give plus points for a partially correct (i.e. incorrect) answer.

I agree that TT should reward speed; but it really ticks me off that you lose points if you're too slow. You should just get 0, not having point detracted.

Don't forget that the amount of time it gives you to solve it with a plus score depends on what other people have done before you; this rewards those who just try out the most "inspiring" move without calculating. And that's not how it works on a real game.

The more you do, the faster you get. That's what I've found. It's all about pattern reinforcement, going over and over the same ground. The first time you try to walk through an area of wilderness it's hard and it's slow, because you have to hack your way through with machetes etc... But the more you walk along the same path, the more it gets trodden down and the easier it gets. Then you make another new path for a different route across the wilderness. Then another and another. After a while you can travel all of the routes you need quickly and easily. But every now and then you discover a need for another new route and you have to clear that, then walk it often to keep it open, along with all of the other paths you made. The more time you spend on it, the more ground you clear and the easier it becomes to move around your bit of wilderness. Eventually you realise that you've reached your level. You just haven't got the resources to clear any more ground and keep it open. So you think, this is as good as it gets. Then some guy comes along with a jet aircraft travelling at twice the speed of sound, and you realise that your achievements were insignificant compared to what some people can do with apparently no effort!