As mentioned in another thread, maybe there could be an algorithm that uses multiple data points to determine availability? So it would take into account that you haven't opened that 5,345th email from a year ago, and you're unlikely to any time soon.