One in four is a floor, not a finding: CTE research finally gets a denominator

The quarter is not an estimate. Of the 878 former NFL players who died between 2016 and 2021, the investigators had pathology on 235, about 27 percent of the cohort ^[5][1]. Vox's newsletter account puts the in-window examined count at 235; the full-period donation total reported elsewhere is 338, at a 93 percent positivity rate [c6b]^[6]. Apply that rate to the in-window subset and you get roughly 219 confirmed cases, or 24.9 percent of everyone who died in those six years ^[3]. The headline figure is confirmed disease divided by all decedents, which is why it behaves like a floor rather than a central estimate.

Everything contested sits in the 643 brains nobody examined ^[2]. The paper's own bound runs to about 97 percent if every non-donor except the 20 definitively cleared brains had the disease ^[7]; the authors allow that the real rate may be near 90 percent ^[19]; the 2019 Binney and Bachynski modelling exercise put it closer to 10 percent ^[9]. Donation is symptom-driven and non-random ^[10], so the ceiling is as much an artifact of who volunteers as the floor is of who does not.

For anyone pricing physical risk, the floor is the usable number precisely because selection bias cannot inflate a count. It sits over a near-complete roster of every man to play a game since 1949 and a national vital-statistics denominator ^[3], and it is conservative in a specific way: another 215 non-donors had a neurodegenerative disease listed on the death certificate, on a document type that undercaptures those diagnoses ^[11].

The defence material is in the same paper. Dementia showed up in more than 90 percent of stage IV cases, but also in roughly 40 percent of donors whose brains were clear of CTE, with weak correlation at lower stages ^[8]. Uzma Samadani, a neurosurgeon who consulted for the NFL from 2015 to 2019, warns against repeating the amyloid hypothesis and treating mislocalized tau as proven cause of symptoms ^[13]. A defendant can concede a 25 percent pathology rate and still contest causation in any individual case.

Then there is vintage. Mean age at death among donors was about 66, so most careers ended before 2005, ahead of helmet grading, kickoff redesign, targeting enforcement and contact-practice limits ^[12]. Vox's newsletter reaches for black lung as the analogy ^[15], but black lung came with an exposure standard; the closest regulatory move here runs the other way, OSHA having proposed excluding hazards inherent and inseparable from the core nature of a professional activity, a proposal not yet finalised ^[14]. Meanwhile the labour supply thins while demand does not: tackle participation among 13-to-17-year-olds fell 9 percent from 2019 to 2024 and now trails tennis ^[16], while NFL games were 83 of last year's 100 top-rated television programmes ^[18].

Both the league and the plaintiffs' bar can quote this study accurately and land in opposite places, because a 72-point interval is wide enough to hold both readings ^[5].