"Predicting", in the absence of a well-specified trigger with historical probability data backing it, is useless in general, but seems to be the thing that attracts most people to trading.
Even once a trigger has fired, there is no way to predict the size of either the gain or the loss from the trade, although if you have a hard stop at the beginning of the trade and you're trading a liquid and orderly market, that would give you a very good idea of the largest possible loss. Even a profit target, if reached, might only represent a small fraction of the overall final potential of a trade, which is why designing an algorithm to extract as much as possible from the maximum favorable excursion is a much more robust trade management strategy than use of profit targets.
Only at that one instant when your trigger is fired are you "predicting" anything and even then, you are only predicting in line with the historical probabilities. I think any other method of "predicting" is doomed to lead to ruin eventually, yet these other methods seem to be favored by 'dilettantes'. I had to learn this lesson and it was definitely part of the tuition the market charged me for my eduction.
