Our calls
Our calls: what we said, and what happened
Updated weekly · scored Nov 12, 2026, 3:03 AM PT · every call is recorded when the card is published and scored only on games or days after it
Usage-shift calls (did the share move the way we said?)
Right 10, wrong 3, no clear move 0 (about 8 in 10 right when there was a move). 8 still open, 0 not scored.
| Player | Call | Result |
|---|---|---|
| Saquon Barkley PHI | up | hit (+35.6 pts) |
| Will Shipley PHI | up | miss (-8.9 pts) |
| Elijah Mitchell PHI | up | hit (+20.6 pts) |
| Tyler Allgeier ARI | down | miss (+18.2 pts) |
| Bam Knight ARI | down | hit (-30.6 pts) |
| Braelon Allen NYJ | down | hit (-20.8 pts) |
| Christian McCaffrey SF | down | hit (-4.6 pts) |
| Jacob Cowing SF | up | hit (+5.2 pts) |
| Jordan Watkins SF | up | hit (+13.4 pts) |
| Brandin Cooks SF | up | hit (+3.6 pts) |
| Tez Johnson TB | up | hit (+8 pts) |
| Ted Hurst III TB | up | hit (+4.8 pts) |
Risers and fallers (did Our Value move the way we said?)
Right 2, wrong 0, no clear move 0 (about 10 in 10 right when there was a move). 3 still open, 0 not scored.
| Player | Call | Result |
|---|---|---|
| Erick All Jr. CIN | down | hit (a bit down) |
| Christian McCaffrey SF | down | hit (slightly down) |
How the estimates work, and how good they were before we started
When a fantasy-relevant player is ruled out (or returns), we estimate which teammates pick up (or give back) his share of the team's targets and carries. The first idea we tried, spreading his share in proportion to what teammates already had, was no better than a coin flip when we tested it on 2024–25 games, so we dropped it. What we use now is a set of averages by teammate type from 2022–25 games. Tested on past seasons, it pointed the right way for about 6 of 10 teammates overall, but only for the biggest expected changes was it dependable: we now name a teammate only when the expected change is at least 12 percentage points of the team's targets or carries, where past calls were right roughly 7 in 10. That is a small share of teammates, so we publish a card only when we can name at least one teammate, and hold the rest. Cards published before this change (labelled in the table by their date) named every teammate.
Usage is measured from weekly player stats (targets and carries), and the injury status comes from the league's public player feed. Dynasty value pressure is a rule of thumb (a big usage move, or an injured-reserve player), and we score it against Our Value 14 days later. A call that did not move by at least 1 point of share (or a small move in value) counts as "no clear move", never as right.
Every call here is also in the public prediction log, with the rookie projections: How accurate are we? →