What you are doing is called equity curve trading. If you search under that title on this board, you will find a number of discussions on this topic. You can also try various definition of drawdown. Dr is generally defined as Highest(Eq) - Eq, now if make Dr = HHV(eq, lookback) - eq, you have a slightly numbers to play with. Or you simply do a Eq - Ma(eq, pds). There are many combinations. I think you should try to use optimization to determine which of the various scenarios suits you the best. Another area for consideration is what you do when you get the filter signal. You can stop buying, You can buy at reduced size or you can sell. One word of caution - you should put in one bar delay in your filter ie, you only act on your filter the next bar.
--- In [email protected], "haders2003" <[EMAIL PROTECTED]> wrote: > > Hi > > I'm interested in any comments regarding the use of % system drawdown > as a filter. By this I mean, only buy when % System drawdown is less > than, say -5%. > > I have applied this to 2 trend following systems (Linear Regression > and MA)and in both cases, although CAR was slightly reduced, most > other measures of system performance were much improved. These > included RAR, exposure, CAR/MDD and of course Max % System Drawdown. > > The way I implement this filter is to run the system as normal and > save a plot of % system drawdown, renamed to Sysdd. I then add code > to the original system so that no new positions are opened if Sysdd > is below -5%. > > Am I missing something here (it wouldn't be the first time!)? > > All comments are appreciated > > Regards > > Haders >
