|
PureBytes Links
Trading Reference Links
|
Thanks for the code!
As a general comment, if you want a (long-term) backtest that actually has a
chance at meaningful profit/loss numbers I think you need to normalize the
position size to be based on your account size, and to be independent of the
price and volatility.
i.e. back in the 80s the volatility in the SP was about 1/20 of what it is
now, so then you would have traded 20 times the number of contracts you'd be
trading today, to get (more or less) constant dollar swings in your account.
You could modify the buy/ sell statements as follows:
buy
12500/(Xaverage(TrueRange,50)*Bigpointvalue) contracts
open tomorrow + LVol *
MaxList(Highest(High,Llen)-Lowest(Close,Llen),
Highest(Close,Llen)-Lowest(Low,Llen)) stop;
In this case12500 is the dollar swing your account will experience per
50-day xatr. This will "approximate" using a constant percent of a fixed
account size for margin for each trade.
Warning: Whatever system you've been working on for all these years could
look completely different if you do this! There are many variations. You can
set up a variable account size which will include net profits, and then use
some percentage of that instead of a fixed 12500 or whatever. If your system
works well, the compounding could be sufficient to overflow your TS4 - you
get huge NEGATIVE numbers!
good luck,
Phil
http://patterntrader.com
|