Note that optimal branch factor would probably depend on the expected negative rate. If 98% of your samples are coming back negative, it may be more optimal to

Note that optimal branch factor would probably depend on the expected negative rate. If 98% of your samples are coming back negative, it may be more optimal to divide by into 4 or 8 groups instead of 2 groups.