Same backtest setup - 400 crypto symbols, 1 hour timeframe, Last 10 years of data
First strategy
Backtest time - 22 sec
Second strategy
Backtest time - 21 min
First strategy
Backtest time - 22 sec
Second strategy
Backtest time - 21 min
Rename
I'm certainly not seeing that, here is a video of my experiencing, using the Dow 30 in both backtests,
https://drive.google.com/file/d/1nWUL-bn3uGgVl1sZXmc37UGt_UDRpvsO/view?usp=sharing
Are you sure you had all of the data completely up to date before the runs in order to make a fair test?
https://drive.google.com/file/d/1nWUL-bn3uGgVl1sZXmc37UGt_UDRpvsO/view?usp=sharing
Are you sure you had all of the data completely up to date before the runs in order to make a fair test?
Maybe 30 Symbols and dayly scale is not enough to produce huge 21 minute "Running Backtest..." lag. What I think is happening that in my first case RSI(Close,20) for BTC is calculated once and than cashed version used for each symbol in data set. In second case (whitch uses SymbolInd) indicator RSI(Close,20) for BTC is calculated for each of my 400 symbols in dataset at start of backtest and produces huge lag at start (possibly it even tryes to update the data for BTCUSDT 400 times).
QUOTE:
Are you sure you had all of the data completely up to date before the runs in order to make a fair test?
100% percent sure
It's not cached though, if you open the C# Code for the Strategy you'll see it's created each time via a "new" call. But I'll dig in and see if there's anything we can do to optimize the SymbolInd!
I think I found the cause and will work on some optimization for SymbolInd.
QUOTE:
I think I found the cause and will work on some optimization for SymbolInd.
Thank you very mutch! Looking forward to updates!
Your Response
Post
Edit Post
Login is required