I'm stuck with my strategy, because the solutions mentioned above are not meeting the demands.
My problem is still that I need exactly X bars of H4 data and exactly X bars of M1 data. Unfortunately I can't PRELOAD the whole amount of these data from the computer, because:
- Huge amount of RAM, unnecessarily, because Zorro's way is loading X*60*4 M1 to generate the needed X H4 bars. So it means (X*60*4 - X) amount of unnecessary M1 data: I only need X of that.
- I want flexible code to run on multiple assets, anytime, just by pressing the Trade button. I don't want to start hunting for historical .t6 data for every time when I start trading, with 60+ assets I trade, especially for exotic ones. And theoretically it's not a problem, because my broker gives me the bars I need for my strategy. But I just don't know how can Zorro handle in that way, so that's why it's a problem.
- Plus, maybe the historical prices that I can get in .t6 to PRELOAD from an independent source has maybe not the same prices that my broker shows me in the trading session after the LookBack. So I need the data exactly from my broker, not from outer sources.
So if I would make a single-timeframe strategy, it woud be very-very easy to get these X bars of datas: my broker just serves it by BarPeriod = 1, or BarPeriod = 60*4. The problem comes in when I want to use them together, in a multi-timeframe strategy, because Zorro forces to load X*60*4 M1 data to generate the X amount of H4 data - but this amount of M1 data is not avaliable from my broker - and also the RAM problem as mentioned above.
I only want to load X amount of M1 candles and X amount of H4 candles, AND for higher bars: not calculating, but loading from the broker.
For example it would be perfect if Zorro would load the datas 2 times at start: X bars for M1, and X bars for H4.
Is it possible somehow? Do you know an other solution?
In live trading I could run two instances of Zorro, one is in BarPeriod = 1, and the other is in BarPeriod = 60*4, then make some communications between them. But of course I want to backtest the whole strategy, and I would like to not change the code for testing.
I appreciate any ideas about reaching these goals.
Thank you!
