Methodology · no black box

How a prop
gets seeded

Short version: we build our own number for a stat, compare it to the book's number, and rank the whole slate by how big that gap is. The gap is the seed. Everything below is the long version.

1–4

The chalk

Largest projection gap on the board, at a book that's usually sharp on that market. Flat bets. If you play three props a night, these are the three.

5–8

Quiet edges

Real value that nobody's posting about. Often stale because the market hasn't priced a role change yet.

9–12

Correlation plays

Thin on their own, stronger paired with a specific higher seed. We label which pairing actually works.

13–16

Live dogs

Plus-money longshots with a real path to hitting. Quarter units. Graded in their own bucket so they never inflate the headline record.

The four inputs

1. Volume, not vibes

Every projection starts with opportunity: targets, carries, minutes, batters faced, shot attempts. A player's talent doesn't score if he isn't on the field. We weight recent usage heavier than season averages, and we re-weight the second a depth chart or a rotation changes.

2. Matchup, at the right grain

Team defensive rank is close to useless for props. What matters is the specific matchup: slot corner versus slot receiver, catcher framing versus a pitcher who lives on the edge, a center's rebound rate against a team that misses a lot of long twos. We grade the matchup at the level the prop actually resolves at.

3. Game environment

Pace, total, spread, weather, and expected game script. A running back's rushing prop and his receiving prop want opposite outcomes — if a team is favored by two touchdowns, one of those numbers is wrong.

4. Market context

Where the line opened, where it is now, and which book is offering it. A number that has moved three points against the public is telling you something. A number that hasn't moved at all on a stale roster note is telling you something else.

The Madness meter

The meter is a 0–100 score built from three things:

  • Projection gap — how far our number sits from the book's, in units of that stat's normal variance. A 6-yard gap on a receiving line means more than a 6-yard gap on a rushing line.
  • Market agreement — whether the line has moved toward us since open. Moving toward us is confirmation. Moving away is a warning we take seriously.
  • Book sharpness — a discount applied when the number comes from a book that's historically tight on that market. Beating a soft book on an obscure prop is worth less than it looks.

Anything above 80 is where 1s and 2s live. Below 45 doesn't get posted at all.

What we don't do

  • We don't sell parlays as locks. We'll show you correlated legs and what the real combined price should be. What you build is yours.
  • We don't hide the losers. Every posted seed is graded automatically. Nothing gets pulled after the fact.
  • We don't quote “record.” A 62% record on −180 favorites loses money. We report units and closing line value, which are the two numbers that survive a bad week.
  • We don't chase. There's no “get it back tonight” play. If the board is thin, the board is thin.

How grading works

Props grade against final official box scores the morning after. Voided or pushed props are removed from the record rather than counted as wins. A prop you tailed is graded at the odds shown when you tailed it, not at whatever the closing number was — you get credit or blame for the price you actually got.

See it applied to tonight

Fill out your bracket.

A free 1-seed drops every morning at 9am ET. No card to look, no calls, no “DM me for the play.”