> Sure, it's unintuitive that I shouldn't go all in on the smallest variance choice. Is it? You have ten estimates of some distance with similar accuracy of t

> Sure, it's unintuitive that I shouldn't go all in on the smallest variance choice. Is it? You have ten estimates of some distance with similar accuracy of the order of 10m : you take the average (and reduce the error by more than half). If you increase the precision of one measure by 1% you will disregard all the others?