If median performance (50th %) is "everyday performance", what is 95th percentile? "Worst-case" or "outlier" imply less common than reality
@dhh yeah totally. Nudging them away from bell curve, especially since all competitors use mean, is hard but important.
-
-
@dhh if you tell someone "the mean is 100ms" they believe it's bounded close to that. "median" doesn't change a thing. -
-
@pitosalas@dhh yeah, but it hides the difference between normal distribution outliers and lognormal. -
New conversation -
-
-
@wycats I would hit that hard. As long as that's the competition, your best weapon is to out-teach why that's a laughably poor metric. -
@dhh it's just a nonsense metric. Mean approximates median in a normal distribution, but we're not looking at one so it means nothing. -
@dhh rand() would not be that much worse in terms of how closely it matches the actual distribution.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.