I’d really like to see a better explanation of that, or some evidence. It’s possible to order things from “best” to “worst”, with the bottom one percent being the best, but there’s really no way to interpret “below the 85th percentile” as anything other than “among the bottom 85%” without some special, personal, opposites-day definition of “percentile”.

To show that at least some at the NIH have a perfectly reasonable understanding of percentiles, here’s a chart of boy’s blood pressure by height percentile: http://www.nhlbi.nih.gov/guidelines/hypertension/child_tbl.p...

The NIH’s percentile scoring system for grant applications certainly knows the proper meaning of “percentile”: http://www.med.nyu.edu/spa/education/PaylinesDescription.htm...

My guess is that the commenter was mistaken.