Week 5 was the first board. We rebuilt it from the model’s records after the games, before this series existed, so read it as a sample of the format. From Week 6 on, the snapshot goes up before kickoff and does not change. It shows what the data looked like, not what to play, and the Week 5 review grades it.
Most DFS content gives you a list of names and a confidence level. This page shows the signals the model read before the games, including where they disagreed, so you can see how the board was built and judge it yourself. After the games we publish the grade, hits and misses both.
How the board was built
Each player’s row draws on three inputs:
- The in-house model’s projection, built from salary, matchup, recent form, and the week’s market lines.
- Public projections, which are free and publicly available. “Public avg” is the average of whichever ones existed for that player. This week it covers fewer public sources than later weeks will.
- Outside opinion: 9 articles and podcasts logged this week, recorded as bullish or bearish on a player, never as a point number.
All three feed one agreement-weighted score. The board below is top 5 per position by agreement score, plus the next 20 overall. That score decides which names appear and says nothing about who to play: a high score means the inputs agree, not that the player will score.
How to read the columns
- Model and Public avg are projected DraftKings points. A large gap means the model sees something the public numbers don’t, or the reverse.
- Range shows how wide a player’s outcomes look next to others at the same position. Tight is more predictable, Wide is more boom-or-bust, and Mid sits between them.
- Notes list the injury tag at snapshot time and whether the sources split on the player.
The board, by position
Each table is sorted by salary, highest first. Split marks players where the inputs disagreed sharply. For everyone else the model, the public projections, and outside opinion roughly agreed. Agreement does not make a player safe. It means fewer inputs are arguing.
Jump to: Quarterbacks · Running backs · Wide receivers · Tight ends · Defenses · Where the sources split
Quarterbacks
5 on the board. Widest gap between the model and the public average: Jacoby Brissett (model 16.3, public 22.4). Sources split on Drake Maye, Jayden Daniels.
| Player | Team | Salary | Model | Public avg | Range | Notes |
|---|---|---|---|---|---|---|
| Joe Burrow | CIN | $6,900 | 21.1 | 20.5 | Mid | None |
| Jared Goff | DET | $6,600 | 22.1 | 23.7 | Wide | None |
| Drake Maye | NE | $6,300 | 19.4 | 20.4 | Tight | Split |
| Jayden Daniels | WAS | $6,000 | 16.4 | 19.2 | Tight | Split |
| Jacoby Brissett | ARI | $5,500 | 16.3 | 22.4 | Mid | None |
Running backs
13 on the board. Widest gap between the model and the public average: Braelon Allen (model 4.7, public 13.8). Sources split on Jonathan Taylor, Chase Brown.
| Player | Team | Salary | Model | Public avg | Range | Notes |
|---|---|---|---|---|---|---|
| Jahmyr Gibbs | DET | $9,800 | 29.4 | 24.2 | Tight | None |
| Jonathan Taylor | IND | $7,800 | 15.9 | 17.8 | Tight | Split |
| Ashton Jeanty | LV | $7,200 | 14.6 | 14.7 | Mid | Questionable |
| Chase Brown | CIN | $6,900 | 20.8 | 17.6 | Tight | Split |
| Jaylen Warren | PIT | $6,700 | 15.2 | 16.9 | Tight | None |
| Aaron Jones Sr. | MIN | $6,500 | 12.0 | 15.5 | Tight | None |
| D’Andre Swift | CHI | $6,200 | 18.0 | 14.5 | Tight | None |
| Cam Skattebo | NYG | $6,000 | 12.4 | 13.6 | Mid | Questionable |
| Quinshon Judkins | CLE | $5,700 | 10.8 | 14.1 | Tight | None |
| Rhamondre Stevenson | NE | $5,500 | 14.9 | 13.1 | Mid | Questionable |
| Braelon Allen | NYJ | $5,400 | 4.7 | 13.8 | Mid | None |
| Emanuel Wilson | SEA | $5,300 | 5.7 | 13.5 | Tight | None |
| RJ Harvey | DEN | $5,200 | 14.2 | 12.5 | Mid | None |
Wide receivers
12 on the board. Widest gap between the model and the public average: Matthew Golden (model 7.7, public 15.3). Sources split on Chris Olave, Nico Collins.
| Player | Team | Salary | Model | Public avg | Range | Notes |
|---|---|---|---|---|---|---|
| Jaxon Smith-Njigba | SEA | $8,800 | 25.2 | 22.0 | Tight | None |
| Amon-Ra St. Brown | DET | $7,900 | 22.3 | 20.9 | Tight | None |
| Chris Olave | NO | $7,700 | 24.3 | 17.4 | Mid | Split |
| Nico Collins | HOU | $7,500 | 21.2 | 17.1 | Tight | Split |
| Justin Jefferson | MIN | $7,300 | 9.3 | 10.1 | Mid | Questionable |
| Garrett Wilson | NYJ | $6,600 | 15.4 | 17.0 | Tight | None |
| Mike Evans | SF | $6,500 | 9.8 | 14.1 | Mid | Questionable |
| Malik Nabers | NYG | $6,300 | 15.6 | 15.7 | Mid | Questionable |
| Michael Wilson | ARI | $6,200 | 17.4 | 16.9 | Tight | None |
| Matthew Golden | GB | $5,700 | 7.7 | 15.3 | Tight | None |
| Carnell Tate | TEN | $5,300 | 11.2 | 14.3 | Mid | Questionable |
| Luther Burden III | CHI | $4,900 | 11.8 | 14.7 | Tight | None |
Tight ends
10 on the board. Widest gap between the model and the public average: Isaiah Likely (model 5.4, public 12.1). Sources split on Colston Loveland, Juwan Johnson, Harold Fannin Jr.
| Player | Team | Salary | Model | Public avg | Range | Notes |
|---|---|---|---|---|---|---|
| Brock Bowers | LV | $7,000 | 17.9 | 15.2 | Tight | None |
| Trey McBride | ARI | $6,700 | 20.6 | 17.6 | Mid | None |
| George Kittle | SF | $5,200 | 16.8 | 12.7 | Tight | None |
| Sam LaPorta | DET | $4,900 | 13.1 | 13.7 | Tight | None |
| Colston Loveland | CHI | $4,600 | 10.2 | 10.4 | Mid | Split |
| Juwan Johnson | NO | $4,500 | 12.3 | 10.3 | Mid | Split |
| Harold Fannin Jr. | CLE | $4,400 | 13.0 | 10.5 | Mid | Split |
| Isaiah Likely | NYG | $4,300 | 5.4 | 12.1 | Mid | Questionable |
| Mike Gesicki | CIN | $4,100 | 8.2 | 10.5 | Mid | None |
| Hunter Henry | NE | $3,700 | 10.6 | 9.1 | Mid | None |
Defenses
5 on the board. Widest gap between the model and the public average: HOU DST (model 6.4, public 9.2). The sources agreed on everyone here.
| Player | Team | Salary | Model | Public avg | Range | Notes |
|---|---|---|---|---|---|---|
| CIN DST | CIN | $3,500 | 9.9 | 8.4 | Mid | None |
| HOU DST | HOU | $3,400 | 6.4 | 9.2 | Wide | None |
| DEN DST | DEN | $3,200 | 8.1 | 8.1 | Mid | None |
| CLE DST | CLE | $2,800 | 7.3 | 7.3 | Tight | None |
| NYJ DST | NYJ | $2,700 | 9.8 | 7.4 | Wide | None |
Where the sources split
9 of the 45 players had at least one input that disagreed sharply with the rest. These are the open questions on the board, whatever their raw score says. When inputs disagree, the projection carries more uncertainty than its number suggests. The disagreement can sit in one source or in outside opinion, so the averaged projections above can look close even when the inputs underneath were far apart.
- Drake Maye (QB, NE)
- Jayden Daniels (QB, WAS)
- Jonathan Taylor (RB, IND)
- Chase Brown (RB, CIN)
- Chris Olave (WR, NO)
- Nico Collins (WR, HOU)
- Colston Loveland (TE, CHI)
- Juwan Johnson (TE, NO)
- Harold Fannin Jr. (TE, CLE)
What to take from the snapshot
- Where a player’s model number sits far from the public average, the model is making a specific claim. The review tests those claims.
- A split means the projection is less settled. It says nothing about whether to play or fade the player.
- Salary tells you what a player costs, not how he will score.
Come back after the games for the grade: every projection on this page against the real result. The fuller discussion happens in the community.
FAQ
What is this page?
A timestamped record of the signals the model read before Week 5 kicked off: its own projection, the average of public projections, and where the inputs disagreed. It lists no recommended plays.
What does “Split” mean?
At least one input disagreed sharply with the others, either in how it ranked the player or in how many points it projected. A split means the projection is less settled. It says nothing about whether to play or fade the player.
How is the board chosen?
It is top 5 per position by agreement score, plus the next 20 overall. The agreement score picks and orders the board but is not shown, and a high score does not mean the player will score.
Does this change after the games?
Not after publication. Week 5’s board was rebuilt after the games, so it is a sample of the format. From Week 6 on, the snapshot goes up before kickoff and the review grades what was shown.