Floor Finder ?

What this finds

Sportsbooks post a player's receptions line from his role — a TE2 gets a TE2 line. When a player's actual usage has moved (more snaps, running a route on nearly every pass play, drawing targets) but the line is still over 0.5 receptions near even money, the price is usually wrong. This board finds those spots.

The evidence

Across every closing 0.5-receptions line captured 2023–2026, flat-betting the over won in every usage bucket except one:

The 1.5 line shows no edge at all. Likely cause: books load "zero catches" probability for a player who might barely play, but if he doesn't play the bet is void, so the bettor never eats that outcome.

How a play qualifies

• Last game: at least 8 routes and a route on at least 50% of the team's pass plays (or 2+ targets when PFF route data hasn't landed).
• Same team as last game, game not started, not Out/Doubtful. Questionable shows a badge. With 4+ targets last game and a 0.5 line, any injury listing disqualifies — that bucket loses.
• A 0.5 line exists at a book and the fair probability beats the book's implied probability by at least 5.0 points.

Fair probability

Expected receptions = expected routes × targets per route × catch rate. Routes blend the last game, L3 and L5 and shrink toward the season average; targets-per-route and catch rate shrink toward the position average so one big game moves the estimate without defining it. A Poisson gives P(1+ catch), then a calibration map fitted on ten seasons of player-games corrects the raw number (which runs hot). Fair probabilities only display for a market once the Tuesday backtest shows the calibrated model beats the no-vig closing line.

Pinnacle floor

Pinnacle often posts a higher rung (1.5). From its price we back out what Pinnacle implies for 1+ catch. "Sharp disagrees" means Pinnacle's implied floor is below the soft book's price — a caution, never a veto.

Watch list

Players whose usage qualifies, whose role is small (most recent captured receptions line ≤ 1.5, or no line captured this season and ≤ 2 catches a game), and who have no 0.5 line in our captures yet. "Line" is Hard Rock's main line and over / under price for the player's previous game — captures run on the props and WR models' schedules, so "none captured" means we haven't seen one, not that Hard Rock never posted one. "Flags at" is the worst price that would still clear the edge threshold — check your book.

Record

Every flag is recorded the first time it clears the threshold (book, price, probability) and graded after the game. CLV = closing implied probability at the same book minus implied at flag time; positive means the market moved toward the play. Units are flat 1u. Voids (did not play) count for nothing.

Limits

Few books post 0.5 as a main line, so most weeks show a handful of plays. Alternate ladders arrive about 90 minutes before kickoff. PFF route data lands Tuesday; until then the fallback rule uses targets. This is a scanner for one structural leak, not a projection model.