|
PureBytes Links
Trading Reference Links
|
At 01:21 PM 5/27/01 EDT, you wrote: >>>> Hi Herman,
Here are some questions that are currently asked to rate a system:
<<
Thanks Jean, some of those I have in MS, what is missing is a means to compare optimized test results with statistical methods. The stats I have are just max and min values, nothing to tell me which curve is the most linear. I believe that a straight line equity is the best prove you've got a good system.
Herman.
>>>> LCD at portfolio level Recovery Time Performance (Compounded annual ROR) Winning Percentage Frequency Distribution of Daily PortfolioReturns Gain/Loss Ratio Average Gain and Average Loss Average Days in a Winning and Losing Trades Strings of Trades – Maximum ConsecutiveWinners & Lowers Edge per Trade (average win per trade * winning percentage)/(average loss per trade * losingpercentage)
cumulative contribution to total profits from the five best trades impact of the five worst trades
and of course the classicals Sharpe ratio, Sterling ratio etc.
Personally I consider that tests have to be performed over a 5 years period minimum (one economic cycle) + 5 years blind test.
Jean Jacques Chenier Alternative Asset Management, Inc. www.alterama.com
<<
|