One thing that is never mentioned when talking about divergence is that divergence is only obvious after the fact. If a market is coming down to re-test a price you are nearly always going to get divergences because of the way the indicators are calculated. Now sometimes the market will bounce and you'll say look divergence worked. But other times price will continue down and the divergence will not have worked.
Look at any market in a persistent trend like the Yen you will see divergences all over the place and none of them worked.
Stockcharts has a good description of all the common indicators and how they are calculated
Edit: I notice that Las has made almost the same point. However what I would say is that after a double bottom/top you will always get oscillator/momentum divergences so it does not really confirm anything. It is a redundant.
Look at any market in a persistent trend like the Yen you will see divergences all over the place and none of them worked.
Stockcharts has a good description of all the common indicators and how they are calculated
Edit: I notice that Las has made almost the same point. However what I would say is that after a double bottom/top you will always get oscillator/momentum divergences so it does not really confirm anything. It is a redundant.