A good place to look is here:

https://school.stockcharts.com/doku.php?id=technical_indicators

I have gone through this list, coded them up and tested them on my own universes and with my own ranking ideas. Combine them, change the parameters. Try whatever you can to examine them for edge.

There’s a lot there and over the last few years I have come back to here, or pages similar and retested ideas I had already tested and discarded, but with a bit more experience I have found things that work which I didn’t find before.

So it’s a long process of constant research and development, often going over things a number of times at different points in your journey.

Another thing to think of when trying these ideas, is maybe running a moving average over the indicator eg MA(RSI(2),5) instead of just RSI….

Also try the setup but one or two days previous…eg Ref(3LowerLows,-1) Often a pattern works but the move is better the second day after the setup, when the stock has turned and is now going up with some momentum….that might give you a better edge.

I have this saved so I’ll post it here.  A long time ago Said Bitar wrote something about “I’m not too creative, so I start with ranking”.  That is some of the best advice I’ve gotten.

When I start system creation, I do the following:

1. make my thesis (big idea)
2. make my ranking (how to express my big idea)
3. make my stretch (how sensitive is my big idea)
4. make my filter (how much support does my idea need)
5. add indicators to my filter if needed (any edge cases to be handled, such as don’t short biotech and don’t short something that was $2 one week ago)

If #1 is not done, there is no need to proceed to #2… and so on.

1. Set your trading universe
2. Set the Moving average duration , price filter and volume filter
3. Set the ranking 
4. Set the stretch

Now you have a profitable MOC system.
The second stage will be to decrease the signal count and the randomness in the system. This can be done by adding some conditions to show short term weakness in long term or medium term uptrend.
This can be done using RSI, ADX, Close below x days, close below band (volatility band or percentage band), ROC and many others
this will help in decreasing the number of trades , drawdown, and randomness.
If the number of trades decreased a lot then increase the universe.
I prefer to add one condition but if needed I can add another one but not more than two.
After I reach results that I am happy with then I go back to the ranking and try to test other types of ranking. My default ranking is based on ROC but I will test other ideas such as ranking by price, ranking by ATR/C and whatever get’s in my head.

Another idea is to develop multiple systems that look OK not great and combine them together sometimes the result is great. But you need to make sure that the trades generated from each system are not overlapped. Personally I am not a big fan of this, I can see the benefit out of it but sometimes it gives me the feeling that I am fooling around to place as many orders as possible.