Hello all,
This is a backtest from a modified version of an algo I found on this thread . After reading through it seems like I found a pretty common problem here on Q, being that the Backtester inflates your returns for low float/illiquid stocks. To avoid this, you'll see that I increased the min target price to $5 and max to $100 and run the backtest. For data integrity I've ran both the backtest and papertrade with $2000 because that is the amount I'd allocate to this.
The below backtest was run using the dates that I've papertraded with it so far and found that I've gotten much different results here than in PT. My PT results were the following through those dates:
Returns: .72% (.0072)
Alpha: .58
Beta: .82
Sharpe: .93
Sortino: .51
Volatility: .21
Main question: Do you think I'll be able to regain the (extremely promising) backtested results in Papertrading/Livetrading if I play with max/min candidate pricing again or is it just a classic case of overfitting?
Follow up: How do I sift through and see what the problem is so I can tighten up the way this algorithm trades? I suspect it's an issue with identifying appropriate crossovers and getting the fills in time.
