Sector-relative scoring

grading like against like · How the app turns all of this into a grade

Some numbers only make sense against a company’s peers. 30% revenue growth is mediocre for software but elite for industrials; a 40× EV/EBITDA is normal for chips but absurd for a telecom. The app scores those metrics against each name’s own sector median — with a guard so a shrinking sector can’t mint “winners” just by grading on a soft curve.

z = (value − sector median) ÷ sector MAD
score = 50 + z × 20

sector median

— the middle value of that metric across the name’s own sector

MAD

— median absolute deviation — a spread measure one outlier can’t distort

z

— how many spreads above/below the sector middle the name sits