Fourth & Value · Case study · Week 2, 2026

Did the market overreact to the Eagles–Titans injury news?

Philadelphia at Tennessee is a useful first test of the new injury screen: the total fell three points, while our injury layer saw a much smaller change.

Pregame snapshot. This is an analysis of the information and prices available before kickoff. It is not a claim that the model has already proved the bet.

42.5captured opening total

39.5current consensus total

−3.0market move

−0.6model injury impact

The injury report

The report included Philadelphia linebacker Jonathan Greenard as Out, with Andrew Mukuba listed Questionable. Tennessee had Cedric Gray, Cordale Flott and James Williams listed Questionable. Jalen Carter was limited in practice without an official game-status downgrade.

Those are real pieces of information, but most of the listed players are defensive players. The first version of our totals injury layer gives modest weight to defensive injuries because we have not yet calibrated how individual defensive absences move a team total.

What the models said

Before the injury adjustment, the totals model projected 44.2 points. Applying the current availability inputs moved that to 43.6, a reduction of about 0.6 points.

The market moved from 42.5 to 39.5. That is a three-point move, roughly five times the model’s estimated injury effect. Our screen therefore labels the game a possible downward overreaction.

The books are not identical at the new number. LowVig had the best listed Over 39.5 price at −107. Fanatics had the best listed Under 39.5 price at −105, while Bovada was the only book showing the better Under number of 40.0 at −110. The opinion and the price are separate decisions.

Is the market wrong?

We do not know yet. The market may be incorporating information the public report does not capture: a late change in quarterback expectation, a defensive matchup concern, weather, or sharper private injury information. The screen is useful because it tells us exactly what would have to be true for the move to be justified: the missing information would need to be worth roughly another 2.4 points beyond the report-based estimate.

That is a hypothesis, not a bet recommendation. The current position weights are deliberately provisional. A defensive injury should not be allowed to create an apparent edge until we have enough historical announcements and line snapshots to estimate its effect by position, player quality and replacement depth.

What we will grade afterward

Did the final total stay near 39.5, move back toward 42.5, or move farther down?
Did the actual game total land closer to 43.6 or 39.5?
Was the report accurate about who played and how much?
Did the same pattern repeat across quarterback, skill-player, offensive-line and defensive injuries?

The useful edge is the disagreement. The market moved three points. Our first injury model moved six tenths. That gap is worth tracking, but it becomes a strategy only after the injury coefficients survive a walk-forward backtest.

See the totals board · Open the injury reaction screen · Back to the blog