Scorecard
How we actually did
Every ranking published here gets scored after the games: start exactly who we said, and see what share of a perfect week you'd have had. The same test is run on two lists anybody could have used instead — the betting market's, and the plain box score — so the gap between the columns is the story. Nothing is left out. A week where the model was beaten appears the same way a week where it won does.
| Week | Ours | Market | Box score | No Vegas |
|---|---|---|---|---|
| 78% | 77% | 75% | 78% | |
| 83% | — | 73% | — | |
| 78% | 71% | 73% | 78% | |
| 73% | 76% | 72% | 74% | |
| 79% | 78% | 73% | 78% | |
| 77% | 80% | 74% | 72% | |
| 81% | 77% | 77% | 76% | |
| Season | 78% | 77% · 6 wk | 74% | 76% |
Week 7
scored 2026-10-27This week we beat the market by 4 percentage points, and beat the box score by 4 percentage points. Vegas earned its place: +5.
| Position | Ours | Market | Box score | No Vegas | Got right |
|---|---|---|---|---|---|
| Quarterbacks top 12 | 81% | 67% | 75% | 80% | 7 of 12 |
| Running Backs top 24 | 90% | 86% | 83% | 79% | 18 of 24 |
| Wide Receivers top 36 | 82% | 80% | 82% | 81% | 24 of 36 |
| Tight Ends top 12 | 64% | 58% | 47% | 51% | 5 of 12 |
| Kickers top 12 | 82% | 73% | 71% | 73% | 7 of 12 |
| Defenses top 12 | 61% | — | 60% | — | 8 of 12 |
Quarterbacks
Ranked high, bustedBryce Young — ranked 9, finished 28Brock Purdy — ranked 3, finished 21Patrick Mahomes — ranked 13, finished 26
Ranked low, broke outMalik Willis — ranked 26, finished 4Geno Smith — ranked 22, finished 5Bo Nix — ranked 16, finished 3
Running Backs
Ranked high, bustedDerrick Henry — ranked 8, finished 43Kenneth Walker — ranked 4, finished 38Aaron Jones — ranked 18, finished 52
Ranked low, broke outAnthony Tyus — ranked 100, finished 32Dylan Sampson — ranked 67, finished 28Trevor Etienne — ranked 88, finished 49
Wide Receivers
Ranked high, bustedNico Collins — ranked 11, finished 77Jalen Coker — ranked 17, finished 80Ted Hurst — ranked 45, finished 90
Ranked low, broke outBarion Brown — ranked 147, finished 28Brandon Smith — ranked 153, finished 57Jacob Cowing — ranked 115, finished 38
Tight Ends
Ranked high, bustedCade Otton — ranked 15, finished 57Juwan Johnson — ranked 14, finished 50Isaiah Likely — ranked 5, finished 39
Ranked low, broke outNick Muse — ranked 96, finished 18Thomas Fidone — ranked 101, finished 32Davis Allen — ranked 55, finished 4
FLEX
Ranked high, bustedNico Collins — ranked 28, finished 169Jalen Coker — ranked 36, finished 172Isaiah Likely — ranked 42, finished 156
Ranked low, broke outDavis Allen — ranked 227, finished 34Dylan Sampson — ranked 239, finished 64Sean Tucker — ranked 213, finished 52
Kickers
Ranked high, bustedBrandon McManus — ranked 1, finished 18Eddy Pineiro — ranked 11, finished 27Ryan Fitzgerald — ranked 13, finished 28
Ranked low, broke outChase McLaughlin — ranked 19, finished 5Chad Ryland — ranked 23, finished 9Andre Szmyt — ranked 24, finished 12
Defenses
Ranked high, bustedCAR Defense — ranked 9, finished 27HOU Defense — ranked 2, finished 17SF Defense — ranked 13, finished 25
Ranked low, broke outGB Defense — ranked 25, finished 7ATL Defense — ranked 19, finished 2TB Defense — ranked 15, finished 1
Statistical detail: rank correlation
| Position | Ours | Market | Box score | No Vegas |
|---|---|---|---|---|
| Quarterbacks | 0.430 | 0.176 | 0.306 | 0.376 |
| Running Backs | 0.776 | 0.658 | 0.679 | 0.747 |
| Wide Receivers | 0.742 | 0.556 | 0.669 | 0.745 |
| Tight Ends | 0.501 | 0.347 | 0.411 | 0.456 |
| FLEX | 0.643 | — | 0.562 | — |
| Kickers | 0.510 | 0.490 | 0.240 | 0.429 |
| Defenses | 0.440 | — | 0.322 | — |
How to read it
Imagine you started exactly the players we said, no thinking. Add up their points. Now compare that to the most points anybody could have scored that week with a crystal ball. That share is the grade. An 83% means our starters scored 83% of what perfect would have. "Starters" means the top twelve quarterbacks, twenty-four backs, thirty-six receivers and twelve tight ends on each list — a full twelve-team league's worth — and points are real box-score stats in half-PPR scoring.
The two columns beside ours are the same question asked of two other lists. Box score is "what if you just started whoever has scored the most so far this season" — the list most people actually use. The market is "what if you started whoever the sportsbooks project highest" — their yards, catches and touchdown odds turned into points. Same players, same week, same maths. So the numbers read like a race, and the gap is the story: ten points clear of the box score means the model added real information over just reading last week's stats. The market only prices the top twenty-odd at a position, so where it covers fewer players than we name, both it and we are graded on that shorter list and the cell says so.
One thing to know about the scale. Even a lazy list captures about 70%, because starters all score something. So read the grade against the other two columns, not against 100. "Got right" is the number a fantasy player feels: of the twelve quarterbacks who really finished top twelve, how many were in our top twelve.
Kickers and defenses get the same treatment on their own rows, but they are not in the week grade: that model is weaker by design, the list says so, and the market does not price them.
"Without Vegas" is what the grade would have been with no betting lines in the model. The published rankings lean on them when we have them; the model-only version is kept and graded on the same games, so you can see every week whether Vegas earned its place.