I agree. But the odds can be heightened if it’s utilized as a complimentary addition to something else that has been proven to offer you a slight edge when initiating positions.
This guy submitted a very nice post offering one such example of using divergence.
oscillator divergence, but the trick is to use it across multiple time frames, and use a mix of regular and hidden divergence according to market behavior in each timeframe. If you’re looking for a reversal in a particular timeframe, use regular divergence, if your looking for trend continuation on shallower pullbacks, use hidden divergence.
If your using fibs, find a timeframe that’s trending, lets assume the 30 min chart shows a good trend with classic higher high higher low price action. Determine your fib levels on that chart, and then look for hidden divergence in the oscillators as price reaches those levels. That’s your signal that the trend may resume. On deeper pullbacks, you might even get regular divergence.
Then look on a faster timeframe, say 5 minute, and look for regular divergence in the oscillators, that’s the signal that trend might reverse. If your late entering, and the reversal is already in place, drop to the 1 minute chart, which should now be trending, and look for the hidden divergence re-entry into the new trend.
I don’t really use TA these days, but if I did, I’d be looking at trend structure (higher highs, lower lows etc) support and resistance (including fibs), price deviation from short term averages, oscillator divergence, with candle patterns as the final trigger.
If I was using fibs, I’d also be looking at using extensions too in order to get an idea where swings might terminate, when you get a confluence of a retrace level in your trend timeframe, and an extension from faster timeframes these tend to give high probability zones, and if these occur at levels tested by price previously, so much the better.