500, I remember coming up with an "average nonzero cell" estimate and then multiplying it by my estimate of how many of these there were. All done in under 5 se

500, I remember coming up with an "average nonzero cell" estimate and then multiplying it by my estimate of how many of these there were. All done in under 5 seconds, a side effect of which is that I don't really remember what process the first estimate came from.