So our error estimates should have error estimates of their own? ;-) More seriously, though, both numbers are acceptable, and quite like far from the truth, but

So our error estimates should have error estimates of their own? ;-) More seriously, though, both numbers are acceptable, and quite like far from the truth, but yours gives off an aura of accuracy with little or no basis. That’s not a problem on hn, given a highly numerate audience, but in other settings it could be problematic. Reminds me of a javelin throwing contest years ago, where results were measured in feet and inches. Given that an inch is 25.4 mm, a Norwegian paper did report the results with submillimeter accuracy.