A short talk (YouTube video) by Adrian Reid about backtesting trading strategies to understand system performance, identify risks, and iteratively refine rules and filters. Reid describes using historical data to find outliers and problematic trades (e.g., volatile, gappy stocks) and adding filters to remove them, improving and smoothing returns over time.
This is where the back testing of your strategies comes in. So backtesting means you take the rules, you put them into the software, you apply the rules to historical data and you see, you know, all of the historical trades the system would have taken. And you can look at the individual trade, you can look at the overall portfolio performance. And when we do that, what we're primarily doing is seeking to understand how the system performs, why it works and where the risks are. So we're looking, when we do the back test, we're looking for outliers, good and bad. We're looking for situations, market conditions, where the system falls apart or doesn't work. We're looking for individual trades that took large losses, for example, and trying to find ways to avoid them in our system by adding rules or filters and so on. For example, you know, I, very early on I had a trend following system. And the trend following system is one that makes that buy stocks that are going up very clearly going up in a strong trend. You buy it expecting the trend to continue. Eventually the trend stops and it turns around and starts going down. At that point you sell it, you make money in the middle. Now I had a trend following system that was making money, but I kept having these trades that I was uncomfortable about. And I realized that I was uncomfortable because the trade, the stocks were quite volatile, the stocks were very gappy, they jumped around from day to day. And I looked at the, the charts of these losses that I was taking and realized that if I filter them out, maybe I could improve the performance of my system. So in the, by backtesting these ideas to eliminate these individual risks, what I found was you can very much improve the performance of your systems. So I added some rules to eliminate stocks that are highly volatile. It improve the performance of the system, smooth the returns and increase the returns. So we're looking for those sorts of issues all of the time. And gradually refining and improving our portfolio of strategies.
Sat.
