Finding True Optimal Risk
Any method that seeks to optimize risk without regard to drawdown is seriously flawed. Who wants to live through a dizzying drawdown just to maximize mean profits? What is needed is an approach to risk that considers the risk to equity first and profits second.
Define True Optimal Risk as the risk level that maximizes mean profits while limiting drawdown to some prescribed level. Using the data from the previous topic, we will find the risk level that limits drawdown to about 10% at the 90 percentile level. Using TradeSim to step through various risk levels, a 3% risk gives a 90 percentile drawdown of 10.9% - close enough. Here is the drawdown chart.
Although this risk level does not produce billions in profit, it does give very respectable earnings. Here is the equity chart.
This is for 50 trades starting with $6,000. Mean ending equity is $15,000 or about $9,000 in profit. Furthermore trading frequency is about 15-25 trades per day so this profit accumulates quickly. This method is very profitable without assuming too much risk.
This example is for the purpose of exploring optimum risk and for purposes of comparing with optimal-f, it is done in %Risk mode (Fixed Fractional Risk).
The method from which the original data was taken was for a fixed risk method. The data represents trading 2 S&P e-mini contracts with no increase as profits accumulate. Taking the data on this basis the 90% percentile drawdown is 8.8% and the mean ending equity is $13,700 - really just slightly less than trading 3% of equity.
The charts on this page are produced using the TradeSim software.
Copyright 2002, Larry Sanders
Last update 2002.04.12