Hmmm. But the open price the next day will not be the same as the close price. So, we loose the overnight value change....
--- In [email protected], "Mike" <[EMAIL PROTECTED]> wrote: > > I believe that the advised approach is to alter your arrays to be > state arrays rather than signal arrays. > > e.g. Buy = Flip(Buy, Sell); > > Your position will still be closed in the previous OOS, but it will > immediately be reopened in the current OOS if the condition still > applies. > > Mike > > --- In [email protected], "dloyer123" <dloyer123@> wrote: > > > > Is there a way to keep trades open at the end of the out of sample > > walkforward step, rather than "mark to market" at the end of the > step? > > > > The problem is that closing trades at the end of each step does not > > model trades that last longer than the walkforward step size. > > > > I reduced the walkforward step size to 1 day, but this has the > implied > > effect of closing each trade at the end of the day in the out of > sample > > data. > > > > What I really want is to confine trade entries to the step window, > but > > let the trade end natrually in computing the out of sample > performance, > > as if I updated the entry parameters each step. > > >
