Sunday Top 50 — Fifty Overs for Week 3, Ranked by Confidence (First Run)

First run of a new card. Fifty over-side player props for the Sunday slate, one list, ranked by confidence, each with its why. It sits beside the main Sunday card rather than replacing it: eleven of these legs are already published there at a captured price, and the other thirty-nine are the best of the 101 players the overnight panel read and left at WATCH — now ranked instead of shelved.

Read this before the table

Prices. Our data platform went dark at 11:05 AM ET (the page rendered, the board never hydrated; re-tested 11:18), so the first version of this card, at 11:16, carried prices on only the eleven legs the main Sunday card had already published. The platform came back at 11:20 and every leg below was priced from the live board at 11:22–11:24 AM ET, before any kickoff. Where the board's line had moved off our number, the card follows the market and shows the original as "was" — the record grades the number first published. No price on this card was invented, and none was added after a game started.

The record. This card is graded on its own page and is not folded into the site record — eleven of its legs would double-count against the main Sunday card. If the format earns its keep, that's a decision for a later week, on evidence.

Ranks 41–50 have no ten-game median in the current role — new team, new starter, rookie, or a one-game sample. They are usage-and-matchup reads only, and they are rated below the panelled forty on purpose.

What moved this morning

Harvey is "unlikely to play" (reported this morning), which turns Dobbins' 57% share into near-workhorse usage at plus money — he moves into the top five. Jennings is OUT (personal): his line collapsing 19.5→9.5 was the market pricing the absence, not an edge, and he is off the board entirely. Collins is OUT (may miss two). Olave, McMillan and Burrow are Questionable; Olave is rated accordingly and Burrow's leg is a VOID if he sits.

The fifty

Median = ten-game median from the overnight panels (2026 games weighted heaviest); gap = median minus line. Opp = opponent's rank on the exact stat, 1 = toughest, so a high number is the soft side. Windows = hit rate of the over in the last ten, last five, 2026 and 2025. Confidence is confidence in the read, capped at 90, never a win probability.

Where the list is exposed

Three games carry six legs each. SEA@ARI (Wilson, Kupp, McBride, Brissett, Smith-Njigba, Harrison, plus Price and Love), MIA@SF (Achane, McCaffrey, Kittle, Dulcich, Samuel, Douglas) and JAC@DEN (Dobbins, Washington, Lawrence, Meyers, Sutton, Nix). Our own record work says same-game legs don't hit less often — but they make a list far swingier than its leg count suggests. Played flat, those three games decide the day.

The Atlanta legs argue with each other. London at a 91-yard median and Pitts at 47.5 both need passing volume that Cooper Rush's own 183.5-yard line says isn't coming. They are kept, demoted, because the target shares are real — but you cannot sensibly hold London and Pitts and Bijan. Pick a side of that script.

Rank-direction conflicts, named: Pierce (opp 1st), Hollins (3rd), Kittle and Samuel (5th), Wan'Dale Robinson (6th), Olave and Juwan Johnson and Price and Love (7th), Chase Brown, Brissett, Dulcich and Douglas (8th). Every one of those is a good player into a top-eight unit on that exact stat. They are on the list because the usage is real; they are where they are because the matchup isn't.

The argument against the whole card

Fifty legs is a volume play, and a volume play across fourteen games is fifty chances to be wrong at once. The honest version of this experiment is the one above: every price captured live and stamped, every moved line shown with its original, and no entry into the record until the format has shown it deserves one. If the first Sunday goes badly, that is information, and it will be printed here on Monday with the same font size as a good one.