Is it possible to do a system test that uses forward looking criteria, to look for commonalities in days with larger gains and losses? Something like

long entry:

c<ref(c,1)*1.05

(and c<> various moving averages etc)

Tried it but got goose eggs..
thanks
:)