|
|
|
|
|
|
|
downtrend, a system based on just T-bond data for the last five years might not be effective. My portfolio includes several futures, such as crude oil and the yen, which have had a pronounced downward trend over the last several years. |
|
|
|
|
|
|
|
|
Neal: Have you ever used any of the more popular software, such as TradeStation or Metastock? |
|
|
|
|
|
|
|
|
C.V.: No, I started using Quattro Pro. I felt it was better to learn the basics writing in fairly simple spreadsheet language. After a few years, it became clear that I needed software to handle a portfolio. |
|
|
|
|
|
|
|
|
Neal: What sort of correlation do you find between your historical testing and real-time trading? |
|
|
|
|
|
|
|
|
C.V.: I have found that invariably the return is lower and drawdowns higher in real time. If historical testing shows a maximum drawdown of 20%, I expect to have a higher drawdown. The rule of thumb I use is that tested drawdown represents only four to five standard deviations, and in real time you will quickly get to six standard deviations, so a 20% historical test drawdown could easily be 25 to 30% in real time. The same principle works in reverse with historically tested return. |
|
|
|
|
|
|
|
|
Neal: In 1997 you had an eight-month drawdown. How did you react to that? |
|
|
|
|
|
|
|
|
C.V.: No trader likes a prolonged drawdown, but they are an inescapable problem with long-term trend-following systems. My historical testing had shown a longer drawdown of 9 to 10 months, so I was mentally prepared for it. |
|
|
|
|
|
|
|
|
Neal: Have you ever had a market make a limit move against your position? |
|
|
|
|
|