What Risk Measures Don’t Tell Us

System Trading

Aug 23

In a recent post, I shared that risk measures are only useful provided all else is kept equal– which is practically never the case.

It is one reason some naive traders may overestimate their performance. The fact an independent trader might make 50% in a year does not mean that they are trading better than a hedge fund that only returned 30%.

And, that is true even if some risk-adjusted measures are also better unless all else is kept equal.

Why? Because all else might not be equal. The hedge fund might be position sized such as to limit maximum loss to less than 1% per day. On the other hand, it is unlikely true for the independent trader. The hedge fund might have concentration risk per instrument, per strategy, per strategy type, per trader, etc. So, they will have layers of risk controls that the independent trader is unlikely to meet.

The fact that many new traders do not understand that is why they overrate their own abilities. Yet, it is for this same reason that independent traders under-estimate the types of returns that are possible.

In fact, independent traders need to seek out and capitalize on these asymmetric inequalities to obtain superior returns.

The fact that a hedge fund may not be able to concentrate risk into a few names is the very reason that independent trader should be willing to consider concentrating risk. The fact that a hedge fund cannot take big fluctuations is why an independent trader should be willing to take bigger swings.

All traders want to limit drawdowns. One way to do that is bet less when you start losing. However, that technique can prolong drawdowns. Personally, I would rather take bigger drawdowns if it means for briefer drawdowns and if I have high confidence that I will eventually make money.

But that’s the problem: risk measures do not tell us what we really want to know which is whether or not we will make money trading some method! They can only tell us that if the future is similar to the past that we may see a similar performance.

And that gets back to my point that risk measures do have value but only when compared like-for-like. If I build 3 trading strategies, I can evaluate the risk measures to get an idea for how they will perform in the future under the assumption of that future conditions are similar to past conditions.

However, I believe it would be wrong to take those measures and compare them to something like a buy-and-hold in MSFT because it is far more likely, if one holds long enough, that one will eventually make money buying-and-holding MSFT. It is a strategy that has been working, sometimes better or sometimes worse, ever since they IPOed in the late 80’s!

About the Author

The author is passionate about markets. He has developed top ranked futures strategies. His core focus is (1) applying machine learning and developing systematic strategies, and (2) solving the toughest problems of discretionary trading by applying quantitative tools, machine learning, and performance discipline. You can contact the author at curtis@beyondbacktesting.com.