Ah, yeah I guess the article could be doing on a day-to-day basis instead of multiple times a day. Maybe the more misleading part of the article is that the nu

Ah, yeah I guess the article could be doing on a day-to-day basis instead of multiple times a day. Maybe the more misleading part of the article is that the number of data points is relatively constant. I would think there's some rough constant on frequency of data points (ie, 1/d), and as you get closer to a true minimum, you have much fewer data points. Having fewer data points as time goes on makes the inaccuracy more obvious, and thus this article less interesting. Def an interesting read tho.