I ran it through my data from the last ~9 months & showed a small loss around 2%, & actually it improved pushing out to 30% - would take a fair amount of live trialling to establish some workable parameters I would think.

__________________
"Now let me get this straight - Whatever I do don't bet this horse?"