The why behind the numbers.
Touchdown Theory is a research tool for the NFL. It projects games and players and shows how every number is built, from the team matchup and pace to player usage and the environment the game is played in. Research, not picks.
Research tools
How all 32 teams rate on offense and defense, per play.
Compare any two teams, offense against defense.
Every venue and the conditions that shape a game there.
How every number is built, and why that is the product.
Plain answers to how the research works.
The league in 2025
Net rating is a team's per-play quality on both sides of the ball, measured across the season. Positive is above league average.
| # | Team | Net rating |
|---|---|---|
| 1 | LALos Angeles Rams | +0.129 |
| 2 | SEASeattle Seahawks | +0.123 |
| 3 | NENew England Patriots | +0.090 |
| 4 | BUFBuffalo Bills | +0.080 |
| 5 | HOUHouston Texans | +0.073 |
| # | Team | Net rating |
|---|---|---|
| 28 | MIAMiami Dolphins | -0.065 |
| 29 | CINCincinnati Bengals | -0.067 |
| 30 | LVLas Vegas Raiders | -0.109 |
| 31 | TENTennessee Titans | -0.122 |
| 32 | NYJNew York Jets | -0.156 |
This week's slate
OffseasonEvery game on the slate with a full projection: each team's points, the total, and win probability, each one traceable to the matchup, the pace, and the environment behind it.
An illustration of the layout. These are not projections, and no live numbers appear until the season begins.
Player props
OffseasonAnytime touchdown probability for every skill player on the slate, expressed as a plain percentage with the fair odds alongside it.
| Team | Player | Anytime TD | Fair odds |
|---|---|---|---|
| KC | Lead running back | 62% | -163 |
| CIN | Top wide receiver | 48% | +108 |
| PHI | Goal-line back | 41% | +144 |
| DET | Second receiver | 35% | +186 |
| BAL | Tight end | 28% | +257 |
| SEA | Third receiver | 22% | +355 |
An illustration of the layout. These are not projections, and no live numbers appear until the season begins.
Fair value vs market
OffseasonOur probability beside what the sportsbooks are offering, so you can see exactly where the two disagree. Probability first, directly comparable to prediction markets.
| Team | Market | Our probability | Fair odds | Book price |
|---|---|---|---|---|
| KC | Team total over | 58% | -138 | -128 |
| PHI | Win probability | 63% | -170 | -155 |
| SF | Game total under | 52% | -108 | +104 |
| BAL | Anytime touchdown | 41% | +144 | +155 |
| ARI | Anytime touchdown | 34% | +194 | +210 |
An illustration of the layout. These are not projections, and no live numbers appear until the season begins.
Accuracy
OffseasonOnce games are played, this is where the calibration goes: how closely the stated probabilities matched real outcomes, sliceable by season, venue, and team. Where the model missed is published, not buried.
Where the games are played
The environment is a measured part of every projection, not a footnote. These are the conditions the model reads.
Venues with a fixed or retractable roof, where weather can be removed from the game entirely.
Grass degrades through a wet game in a way artificial turf does not.
Estadio Banorte, where thinner air lets the ball carry farther.