On the issue of holding precision constant over time--a slight modification...

[Subbarao Kambhampati <rao@asu.edu>]

So while Jadiel is right that in general precision can't be held constant w.r.t time, there are two important exceptions--namely precision at 1 and precision at 0.

If you only give the relevant answers and nothing but the relevant answers, then your precision remains at 1 at successive periods. Similarly, if you only give irrelevant answers and nothing but irrelevant answers, your precision remains at 0 at successive periods.

Rao