Originally posted by Reverb
When you're clearly losing a game, do you sometimes gamble with a move that's clearly not the "best" but has some small chance of trapping your opponent? Any tips on when to try this? Or should I always make the objectively best move?
It all depends on what you mean by clearly losing, on the type of position, and also on your own skills (tactical, positional) and those of your opponent.
If you're facing immediate catastrophe, then you have nothing to lose, and everything to gain, by complicating the situation by attempting a cheap shot (which only becomes a "swindle" if it succeeds); that is, by attempting some tactical trap which, with best play, your opponent can avoid while maintaining his advantage.
On the other hand, if you're in a situation in which your opponent may, with perfect play, grind you down in 20 moves, and you don't see any forced lines offering tactics (and you should always look for such lines), then you should probably play moves which improve your position the most and are the toughest for your opponent to meet -- which is to say the most sound. Taking your time here is important.
Here's the contrast: in the first instance, your opponent must be able to see through a single tactical position; whereas in the second, your opponent must be able to consistently play strong moves, even as you are doing so. So, in the first instance, he needs good tactical insight into a one-shot; but in the second he needs to be consistently tough, move after move after move.
Note also that in the second instance, while you should also be looking for forced tactical lines, you should also be looking for ways to get a draw should he err. As a player seeking a draw you also have certain advantages (e.g., you can try reversible moves that he may know how to meet, but if he insists on meeting them in the best way, you can force a threefold repetition of position). Perpetual check is merely a specific case of the threefold repetition rule. And don't forget that in certain instances stalemate can be achieved by forcing tactical lines. Finally, endgame knowledge is always good in a tough situation that your opponent can win on over the long term, but only if he plays perfectly. In such situations, don't just make one last desperate cheap shot, hoping he won't see it; at every move, try to improve your position and play the one move you would least like to see if you were him. By requiring superior play from him, move after move after move, you give him plenty of opportunities to make subtle mistakes (more likely to be overlooked perhaps than your desperado tactic).
Of course, sometimes lines simply aren't clear. Then the decision whether or not to make a speculative sacrifice resolves to your own preferences. I suspect that, for many of the better players at least, such moves are made more often from a position of strength, because they seem good though they can't calculate to the end, rather than from a position of weakness due to desperation.