I've read too many books and articles, and heard too many anecdotes that make some technical indicator sound fantastic. But until you backtest it (and then forward test it) against a large amount of data, and make sure you're not curve-fitting, I don't think you can have much confidence in it actually working in the real world in the long run (though you may get lucky and have it work for a brief period of time).
For backtesting, the best tool I've found is AmiBroker[2]. It's orders of magnitude faster than any other backtesting tool that I've found, and lets me backtest mountains of data against any indicator I can dream up within seconds or minutes, and then go on to the next one. It's pretty awesome (though kind of quirky in that its programming language is array-based). Anyway, highly recommended.