Measuring the Edge Of Your Trading System: Is It A Winner?



By Dylan Johnson

Often I hear many traders talk about making gains on certain trades in net profit terms such as "I made $10,000 on a swing long trade on XYZ." or they also will say "Made 5% on XYZ for a day trade." The biggest problem is that average traders lack or do not understand the importance of analyzing performance metrics for your system or means of trading. In Quantitative Systematic trading, the most important analysis of your process is the risk and performance metrics. Over the years there has been a range of conflicting opinions on what is most important, I happen to believe the three most important metrics are P-Fac, Win-Ratio & Maximum Draw Down, but for now, we will focus on the importance of the Win-Ratio.

The Win-Ratio calculation: #Winning Trades / #Losing Trades

The Win-Ratio only focuses on the probability of your signal, system, or process's success after the trade has been executed then closed. If your trade after the said signal has a positive return value it's considered a win, if negative a loss. Win Ratio basically acts as the method to calculate the positive or negative probability of your trading or system. Probability measures and evaluates the following: Risk, Luck, Likelihood, Chance, Uncertainty & Randomness. If these concepts are important to your trading then you must consider probability, since it is the only way to factor in these concepts and deal with them properly, and evaluating your system's performance.   

The Win-Ratio is important because it shows the percentage probability of your signal being correct, and naturally showing you the edge you have in being right and having a profitable trade on said signal & position. The higher the Win-Ratio the more probable your systems market-timing ability is as well as your strategy's robustness. The key to analyzing the Win-Ratio metric is not just the "higher the better" however how consistent the Win-Ratio is. This is where large samples of data come in. To eliminate bad data and misleading performance metrics, you have to test the systems signals on a large sample of data. To have statistically valid data from sampling to derive a reliable performance metric the rule is never less than 100 trades, and preferably 1000 trades or more, and of course the more trades the better when calculating the Win-Ratio.

So for example, if you have only 10 closed trades and 8 of them were winners and 2 of the losers, your Win-Ratio would be 80%. However, this is misleading and false since the sample size of the trades is too small to be consistent. The key is going back as far as possible with back-tested data or historical trades the further back you can go the better & more reliable the Win-Ratio is (some say further than 10 years is too much due to market regime change but that's a different debate which can be easily solved by comparing the 10 year to the X period longer and measure the difference) so in short the higher the number of trades and longer periods of time tested, the more consistent the Win-Ratio. The more consistent the signal, the more robust the signal/system becomes.  

In conclusion, the Win-Ratio is a fantastic performance metric to analyze your system, signal, or trading with.

Dylan Johnson 23 is the Hedge Fund Manager at Sardonyx Capital, a fully systematic quantitative hedge fund launching January 1st, 2015, the President & Founder of Obsidian Analytics, and creator of the Obsidian Mechanical Automated Trading System. 

Comments

Popular posts from this blog

How The Inevitable Lucid Motors Bounce Will Play Out

Greenwich LifeSciences: The Next 10x Bagger?

FuboTV's Bullish Game Plan