I drove the data from BTCUSD into the tester, the trick is that I first downloaded them from investing.com, but it turned out that they did not match with TradingView and bitfinex, at least at the beginning of 2017. I had to run a script and parse from TW.
Then I drove and a mistake, P / L calculations are not adequate. Bug The fact was that the asset for the USD pair had a slightly different logic for taking margins, so I rewrote the tester.
Well, at the end of the launch, the results ... no better thanyesterday on ethbtc) and when I ran it with the pens and built it visually, it turned out that I myself didn’t though, got to that period where it all worked well. Glory to the historical sweep) True, I still haven’t checked this pair with pens for all the correct operation.
Positions on real closed everything, I saw strategy further.
In PND I will be engaged in correction of logic of movementthe reverse point, I’ll look whether it will become better, then the optimization of the parameters of the ATR size and its period, based on these 2 pairs I will single out the universal average for both, and then I will take up the logic of recruiting and dropping positions. The latter should play a key role. And, well, before that, I also verify the correctness of the algorithm's operation, whether it checked everything for me.
It would be interesting to know in general from whom how the trend reverse works and by what logic, consult plz
I will be glad to pluses and subscription!
I also keep a diary on developments, deals and steps to build my hedge fund here - t.me/drsombre