Shannon Entropy by Stephen Massel
Here, we define Shannon entropy and show how you can apply it to your trading results
and derive a measure of the uncertainty or predictability of your method.
There are many different trading systems and strategies that can be used to trade
the markets, and these systems (including discretionary trading) all generate trade
records that can be analyzed, whether generated from backtested results or live
trading. There are also many techniques that can be used to analyze these trades (for
more on this, see my March 2011 Stocks & Commodities article on mathematical
expectation, What Can You Expect, Mathematically?).
Of course, you can review the profit & loss of these trade records as well as many other
metrics, but is there a formal way to quantitatively measure the structure (nonrandomness)
of the data and compare it to the purely random case in order to provide some insight into
the predictability of the method? Welcome to the world of Shannon entropy!