rtucker 318 posts msg #33636  Ignore rtucker 
10/23/2004 5:50:21 AM
need help with trying to find flags. this filter only sees 3 to 7 days consolidation. if i try to add more or's it dosent work. am i doing something wrong or am i reaching a SF limit. Oh.. thank you Cegis / others for instructing as well as posting great filters.
set{v,count(3 day slope of close is between .02 and .04,1)}
set{a,count(5 day slope of close 3 days ago > .08,1)}
set{and1,v * a }
set{t,count(4 day slope of close is between .02 and .04,1)}
set{r,count(5 day slope of close 4 days ago > .08,1)}
set{and4,t * r }
set{b,count(5 day slope of close is between .02 and .06,1)}
set{c,count(5 day slope of close 5 days ago > .08,1)}
set{and2,b * c }
set{d,count(6 day slope of close is between .02 and .06,1)}
set{e,count(5 day slope of close 6 days ago > .08,1)}
set{and3,d * e}
set{h,count(7 day slope of close is between .02 and .06,1)}
set{i,count(5 day slope of close 7 days ago > .08,1)}
set{and5,h * i }
set{or1,and1 + and2}
set{or2,or1 + and3}
set{or3,or2 + and4}
set{or4,or3 + and5}
or4 > 0

SurfnDestiny 78 posts msg #33645  Ignore SurfnDestiny 
10/23/2004 6:45:01 PM
Hey, thanks for showing me the answer to my back day slope question in your example.

cegis 235 posts msg #33661  Ignore cegis 
10/24/2004 12:19:08 PM
rtucker,
After the great compliment, how could I refuse an answer?! <grin>
SF has a limit as to the "computational expense" of a filter, which it attempts to calculate. One thing that is "expensive" as far as SF is concerned is "nesting" set{}s. That is, one set{} based on another set{}, based on another, and so on, such as in your calculation of or4 (in particular).
Sometimes, but not always, you can get further in your calculations by removing a level of nesting, if possible. In your case, you don't HAVE to use set{}s for v, a, t, r, b, c, d, e, h or i. Simply substitute the count()s for these set{}s in the calculation of and1, and2, and3, and4, and and5  thus removing the extra level of nesting.
Also, it may help (although I have not proven it to myself yet) if you calculate the or's as such:
set{or1,and1 + and2}
set{or2,and3 + and4}
set{or3,or1 + or2}
set{or4,or3 + and5}
You may notice that the number of "levels down" the set{}s nest in my implementation is 3, where in yours it's 4.
HTH,
C
