[RT] Re: Mechanical S&P Trading Systems

[Bob Fulks <bfulks@xxxxxxxxxxxx>]

At 7:58 PM -0800 12/17/99, Gary Funck wrote:

>In reviewing this system with Dennis Holverstott, I think the 
>calculation of "pl" should be changed as shown below. I don't know 
>how well this tests out, and for now, do not have an ELA to provide.


I think it is more instructive to write it as shown below. This 
reduces to the same as in the original posted by John - just an 
algebraic manipulation to show that the system uses a modified 
"highest high" and modified "lowest low".

Both values are modified by the same (h-c)/2 term which would tend to 
be zero on up moves and tend to be equal to "Range/2' on down moves, 
effectively adding a dead band which would reduce whipsaws.

The system test out pretty well on 90 minute SPX data if 

     length = about 7.

Bob Fulks


{************************************
  Blitzkrieg System
  Trades 90 min. bars on the S&P?
  courtesy of: "Dr. John Cappello"
	       <jvc689@xxxxxxxxxxx>
  "[RT] Mechanical S&P Trading Systems"
	(Dec 17,  5:57pm)
************************************}
input:ph(h-(h-c)/2),  pl(l-(h-c)/2),
      length(12), length2(12), ToEXIT(0);

buy highest(ph, length2) + 1 point stop; 

if Toexit=1 then
    exitlong lowest(L, length2) stop;

sell lowest(pl,length) - 1 point stop;

if ToEXIT=1 then
   exitshort highest(h,length2) stop;