IMO the first thing we should consider is the time frame we're looking at because a retracement in a 60 min chart will look like a reversal in the 5 min chart or a retracement in a day chart will appear as a reversal in a 30 min chart and so on going from a higher time frame to a lower one. I believe that the key is to look at your higher time frame and find out how the price action in that time frame affects your trading time frame ( a lower one). It is rare that a trend will turn in the opposite direction ( reversal ) suddenly without showing early signs of weakness.
I think that retracements and reversals can only be distinguished after the fact. Going forward, however, I think they should both be treated the same way. Therefore, trying to distinguish between them a priori is effectively a moot point from a practical standpoint.
That would be difficult to do since an entry at the reverse would be stopped out and reversed by taking the retracement.
I think thunderdog's point is that you look for your direction. You get the movement counter trend, you put on you trade, you put in your stops and you see what happens. I not sure I endorse that point of view, but it is what I have been thinking.
I believe that would depend on how you treat the price action. In my humble view, buying straight into a dip or selling straight into a rally is tantamount to playing chicken with the gods. For all the methods and approaches that have been suggested thus far, I would wager that, for each instance that a method successfully distinguishes between a retracement and a reversal, I could find an example where it does not. What, then, does that say about the usefulness of the "distinguishing" method? This is why I am saying it cannot really be done. The fact that it is occasionally done is no more compelling to me than the outcome of a coin flip.
Would you not have to go first so that I can show you a counter example? Besides, are you telling me that for every instance of an indicator crossover or divergence or whatever that does work, that you, yourself could not find an example where it does not work? Come on, eh?
Why would I have to "go first"? I didn't make the challenge. In any case, reducing it all to whether or not some particular aspect of trade management "works" or not is oversimplification. There are no guarantees, only probabilities. Therefore, one must also develop a set of entry, management, exit, and stop rules in order to protect himself if he assesses the probabilities incorrectly. The guidance I provided earlier is just that - guidance. Getting down to creating a detailed retracement/reversal strategy is something else. If arzoo wants to do that, I'll be happy to help.