You already know that strategy for range trading is not profit for trending trading.
So why not let test your idea on range, run all setting, find all the best settings, then let them compete in a survival race, to choose the best of best.
My trading model:
Period #1: 12 to 48, step = 6
Indicator #1 value: - 1.0 to 1.0, step = 0.2
Indicator #2 value: - 1.0 to 1.0, step = 0.2
Period #2: 12 to 48, step = 6
Indicator #3 value: - 1.0 to 1.0, step = 0.2
So i run around 35,000 tests on a chosen range, 7 ranging periods, then run all setting again all data, trending and ranging to find the best of them. There is a ranging period, the tests ran for 5 minutes on my X200.
On USDJPY, run tests with MA period from 12 to 48 (step = 6), RSI period = 12 to 48, value from 40 to 90 (step = 5), to find setting work best for selling.
It is around 540 settings for a ranging period. Then let them run against all data to find which one survive.
The awesome thing is you only point and click, dont' have to code, choose symbol you want, period you want, indicator you want. Then send those settings and ask a coder to program an EA for you.
Quickly find an which idea work, which doesn't.
Select a period to train,
when you train all period, then let them race, show you the results.
(I view daily chart then train on M5 data)
The program is for my trading model, but since i already have working code base, and there is a ta-lib, implement this idea will be not a problem. I will work on it on my spare time.
If you guys are interested, enough interest, i will code and release it for FREE.
If not, i will continue to work on my profitable trading model.
See you guys later!