If those figures have a large amount of data to back them up. For me, I would simply bet them at a percentage of bank non reducing. Probably 1% of bank would be a realsistic figure as you would expect a 20+ run of outs at some stage, and it is always possible to hit another long run after that, so its better to be a bit consercative and keep your bank in a healthy position even when the tough times come (which they will at some stage no matter how good the selection method).