I think it'd be interesting to take the estimate and actual numbers and figure out the actual distribution. Possibly per user. I'd like to use that myself. I be

I think it'd be interesting to take the estimate and actual numbers and figure out the actual distribution. Possibly per user. I'd like to use that myself. I bet it'd be heavily right-weighted since there's a finite amount of time to turn something in early, but an infinite amount to delay.