Re: Buying without selling

["Gary Fritz" <fritz@xxxxxxxx>]

> The two systems were designed to compliment each other and when
> they were opposing each other in the market there was no way of
> telling which would win - if there was, you'd only go with that
> system. However, both have stops and in testing it was found that
> the system which was "wrong" got stopped out while the system that
> was "right" carried on to make a profit. This proved more
> profitable than having the systems stop each other out of the
> market. 

The "having the systems stop each other out of the market" part might 
have been the problem.  You might have implemented that wrongly.

Let's say sys1 goes long, then sys2 goes short.  Now the market goes 
up, so sys2 was "wrong."  In your model sys2 gets stopped out, and 
sys1 stays long and goes on for a profit.

You can exactly model this with a single-system approach, using the 
kind of code I posted yesterday:

sys1 goes long:      add 1 to your position (now long)
sys2 goes short:     subtract 1 from your position (now flat)
sys2 stops out:      add 1 to your position (now long)
sys1 takes profit:   subtract 1 (now flat)

So you are long from the time sys1 goes long until sys2 goes short, 
and from the time sys2 stops out until sys1 exits.  The profits on 
this model will be identical to the profits from your 2-system model.

Gary