Remove outliers from a backtest

When I run a backtest on a code, sometimes I'll see trades that deviate quite a bit from the norm. These outlier trades are not likely to repeat and they can distort my backtest results. To eliminate the outliers, I can simply delete those particular symbols from my watchlist and rerun the backtest. Alternatively, I can go to the quote editor and delete the range of dates for each particular outlier trade for each symbol.

Is there another way to do that, that anyone is aware of?

@Marcel, if you put them in a watchlist and select such watchlist in the "Exclude" tab of the filter settings dialog, they should no longer appear in your results.

immagine

2 Likes

Just be cautious with your assumption that "These outlier trades are not likely to repeat and they can distort my backtest results." It's one thing to create a new rule that helps eliminate outlier trades, like "100-day Historical Volatility < 50". It's quite another to cherry-pick specific stocks or trades to eliminate using the benefit of hindsight.

6 Likes