In my studies I found they offer no edge and I did the research with 6 different oscillators between various timeframes over thousands of samples.
And then Russell spikes 8% higher, ES rallies 81 pts, NQ up 6-7%...The cute part was the break below 2025 ES during the first hour of trading 2 weeks back...Never looked back.
I been using indicators along with Charts long time, what I have found going from the 80s to 2016 is night and day. Way too many who designed indicators had no clear concept of what they exactly had when they made their indicators. I understand RSI by far the best in it's waggles(movements), most traders compare the highs/lows of the price with the indicator, but they should be comparing the closes. I had someone design an RSI for Tradestation then Ninja once where when RSI was above 50, it was an RSI based on highs of price and when it dropped below 50, smoothed over RSI based on lows of price, so now I would be able to compare price lows to other price lows and that took me to about 2010. Thereafter for scalping ES, I dissected price more. I wanted to get my scalping down to like an hour, take lite coat of Elliott wave of 3 moves trend/2 counter-trend, so I been using 7-10 points as Swing Zone of where in first hour of trading, ES will likely reverse. So when I went back and included this in testing divergences of RSI(6) worked out better where divergence with trend under 7 points and counter-trend 7-10 points, I know what the RSI will show on one minute timeframes and no longer even have the indicator, when pivot highs are broken by couple ticks twice, that shows me two divergences as one divergence is not very reliable. So often times where something does not work for me, it is cause there is not enough information, or doing what is not in the book. I have never been able to make CCI work, and only been able to get Stochastics work in very weird way to make more consistent. But others swear by both of them. Nice trade FreakofNature