Cool yes I agree, 50-100 on forward should be about right.
I actually think a forward test is the only test!
With backtesting the introduced bias is so significant that most people just get into a cycle of backtest-optimise-backtest-optimise-forward test-start again
I’m just cutting out the looping part and forward testing on demo.
To be honest I’ve been watching the system for months now on charts and following the hypothetical entries exits etc so am very confident that it works, hence bothering to code it.
I have purposely coded it with flexibility on the most influential parameters whilst making sure to avoid over optimisation.
For example, I had a choice of trying to fine tune entries (but lose many profitable ones) or accept more entries (but get into more losing ones)
After looking at this I just realised that “fine tuning” entries was a psychological issue as opposed to a system issue so to overcome this simply reduced risk per trade so that each loss didn’t really scare me.
As a famous trend following trader (Ed Seykota) said “Risk no more than you can afford to lose and also risk enough so that a win is meaningful”. This underscores my approach to position sizing.
I’m still weighing up pyramiding positions as they become profitable...