There is a popular saying- 'lies, damn lies, and statistics.' Most Risk Management philosophies that Traders base their systems on are grounded on some form of statistics be they Expectancy, RR, Win Loss Ratio, Drawdowns etc. But how realistic are these really?
So a system with an Expectancy of 1 means you will make 1 on every 1 Risked. But over how many Trades has this Expectancy been calculated? 100, 1,000, 10,000? Statistics on such types of analysis generally follow a Normal Distribution represented by the famous Bell Curve
To get a good representation of how a system will work out, you need a very large set of observations. Very few Traders capture these. Secondly, and more importantly, the figure is not absolute. It is subject to a fairly wide range of variance
So a system with a Win Loss Ratio of 60% will give you, on an average, 60 winning Trades over every 100 over a large series of 100 Trades. But that is not to say that you will not run into an observation, or a series of 100 Trades where only 10 will work out.
Black Swans don't only happen with individual Trades- they happen with Systems too. Ultimately your System may work out and perform as expected but that's of little comfort if you're bankrupt before it happens
So while it's good to have a Statistical underpinning to your Trading System, one should be prepared for bad things to happen and not have blind faith in a backtested System ultimately working out. Remember Markets can remain irrational longer than you can remain solvent
My personal belief is that given enough time and Occurances, every system, no matter how good, will suffer catastrophic failure. For most it may not happen in their lifetime. For others, it may happen quickly. Markets are always capable of producing an event which will force this
What is well known and written about is that Richard Dennis designed the Turtle Trading System and made humongous amounts of Money. Less known is that his last hedge fund had to be shut down when it suffered unacceptable losses which were well beyond projected parameters
• • •
Missing some Tweet in this thread? You can try to
force a refresh
