Hi Carl -- The bar size is a valid optimization variable.
Just be sure you run a walk forward test to validate that the in-sample results represent recognition of a pattern in the data, rather than the system just got lucky or fit to the noise. Thanks, Howard On Tue, Sep 23, 2008 at 8:49 AM, Carl Vanhaes <[EMAIL PROTECTED]> wrote: > Have you ever noticed that some systems work better on specific time > frame whatever the optimization settings? > Using TimeframeSet I was wondering if someone has tried to use > different time frame as optimisation values, for example, optimize a > given system on 5,10,15....240 minutes using TimeFrameSet in the > formula, something like: > timeframe = Optimize( "Time_frame", 300,60,3600,60 ); > TimeFrameSet(timeframe); > > I tried this and ran an optimisation; it worked but gave totally > inaccurate results: when I tested the system on the best time frame > coming out of the optimisation it gave totally other (and of course > worst) results. > > Any advice? > Carl > > >
