Mike,

You are right, of course.  Previously I was multiplying Sharpe and
KRatio and when I changed to adding them, I didn't remove that
unnecessary code.  And I'd for some inexplicable reason changed
CompositeMetric * - 1 to Abs().

All fixed now.  Thanks for pointing it out.

Joe

--- In [email protected], "Mike" <[EMAIL PROTECTED]> wrote:
>
> Joe,
> 
> Your code snippet does not make sense to me. If both NetProfitpc and 
> RecFactor are negative, then ProfitMetric will never be positive 
> (i.e. negative plus a negative equals a greater negative), and your 
> condition will never be true.
> 
> Thus, your expression ProfitMetric = abs( ProfitMetric ); will never 
> be evaluated and would have been redundant anyway, since you would 
> only have executed it if you had found ProfitMetric to be positive in 
> the first place and abs() just renders a negative number positive 
> while having no affect on a positive number.
> 
> Am I missing something?
> 
> Mike
> 
> --- In [email protected], "Joe" <j0etr4der@> wrote:
> >
> > If you are using a Custom Metric as your Optimization target in the
> > Walk-Forward settings, you should hand-check your results. I am
> > frequently getting incorrect results, in that the highest value of 
> the
> > Custom Metric is not the one selected by W-F for a date range.
> > 
> > I documented the problem in the 5.08beta to AB support on May 13th 
> and
> > have not received a reply as yet.
> > 
> > This is the metric that is currently causing(?) the problem:
> > [code snippet]
> > ProfitMetric = ( NetProfitpc + RecFactor * 20 ) * 10 ;//make them
> > ~equivalant
> >     if( 0 > NetProfitpc  AND 0 > RecFactor AND 0 < 
> ProfitMetric )//if
> > both negative & result positive
> >             ProfitMetric = abs( ProfitMetric );
> > 
> >     bo.AddCustomMetric( "Profit", ProfitMetric );
> > [end code snippet]
> > 
> > Other of my metrics exhibit the same type of incorrect results in my
> > testing. I'm using Release 5.10.1 now, so I thought it was 
> appropriate
> > to bring this to the attention of the group.
> > 
> > Good trading,
> > 
> > Joe
> >
>


Reply via email to