Forex Trading 2018- TA/FA- Setups and discussion, page-4220

Hi Mason

Welcome to the thread. Nice looking equity curve you've got on that ichimoku/tdi strategy - the return is very good especially the small number of trades - obviously the signal doesn't occur very often.

I think TraderNin have a very good response about backtesting. I will add a few comments:

How large should the backtesting sample be? Generally speaking I think the larger sample the better but as you say it depends on nunber of data points as well. Personally I try and obtain a minimum of (approximately) 100 data points per market for in-sample testing and 100 for out of sample testing, so if using 12 markets then I'd want a total of 2400 data points covered big both sample periods. I think "the number" should be whatever is enough to give you confidence that the testing is solid enough for you to be willing to invest money in the strategy.

For my 1st strategy - my H12 reversal strategy - I used 5 years of data covering both sample periods. For the next 5 EA's I used 3 years of data (18 months in sample, 18 months out of sample). They all met the minimum number of data points required but I must admit I feel more confident in the first EA than the others, probably because the stats for that one were all done by spreadsheet than in the MT4 tester so I understand it much better (so far it's also performing the best).

Final comment - probably more important than the volume of data is the process - if you follow the in sample and out of sample testing and then demo testing process properly, and if the numbers stack up after following that process, you should be very confident in your strategy before you apply $$$'s to it.

Cheers, Sharks