elmerdata.ai blog

My blog

Dan Quinn Has No Underwriter

Commanders coach Dan Quinn does not regret the call that cost him his quarterback. That's because nobody priced the exposure - and that would be the guardrail.

Three seconds remained in the first half in Dallas on September 20. Washington trailed 17 to 10 and had first and goal at the one-yard line. The field goal unit was on the sideline and stayed there. Dan Quinn called a touchdown play from under center, left guard Chris Paul stepped on Jayden Daniels, Daniels braced his fall with his left arm, and the elbow he had dislocated in Week 9 of last season dislocated again. Marcus Mariota finished the game. Asked afterward whether he would take the call back, Quinn said no: "We had touchdowns drawn up, so I don't regret the call."

Take him at his word, because his word is the point. Quinn is not being callous. He is answering the question his job asks. A head coach is evaluated on whether he won the next 60 minutes, and a touchdown from the one-yard line at the end of a half is a good bet by that standard. The injury itself was a freak: a teammate's cleat, an awkward fall, the wrong arm. Nothing about the play design broke Daniels' elbow. What the play did was expose the franchise's most valuable asset, a quarterback with a documented weak joint, to a goal-line collision when a kick was on offer. Quinn's evaluation system has no column for that exposure. It records outcomes, and outcomes are mostly luck. That is the structural problem, and it does not go away by asking coaches to be more careful.

Outcomes are what a coach sees, exposure is what an actuary prices

Every risk-bearing industry eventually learns that you cannot govern by outcomes, because outcomes are too noisy to teach anything. A driver who runs red lights gets home safely most nights. A bank that lends without collateral books profits for years. What an insurer prices is not the crash or the default but the exposure that precedes them, aggregated over enough cases that the luck washes out. The premium on a red-light runner goes up before he hits anyone. That is the entire technology of actuarial science: convert a pattern of exposures into a present cost, and hand that cost to whoever controls the pattern.

Football has no such mechanism at the point where exposure is actually decided. The people who decide it are coaches, whose contracts are settled on wins and whose tenure ends when the wins stop. A quarterback's elbow, or a lineman's brain, depreciates on a schedule measured in decades, and none of that depreciation appears on the ledger the coach is judged by. The same afternoon in Dallas supplied the lineman. On Washington's final drive, in a game that ended 37 to 20, right guard Sam Cosmi was blocking when his own running back, Kaytron Allen, crashed into his helmet. The replay left nothing to interpret. Cosmi dropped to his knees, stayed on the turf for several minutes, and was walked off by the medical staff, and Quinn confirmed after the game that he was in concussion protocol. Cosmi is the team's most expensive lineman, signed in 2024 to a four-year, $74 million extension with $45 million guaranteed. He tore an ACL in the January 2025 playoff game, returned in Week 7 of last season, and in December went on season-ending injured reserve with a concussion. Nine months later he has another one, sustained on a drive that could not change the result. The protocol governs when he returns. Nothing governs whether a player with that file should have been on the field for that drive at all, because that is a coaching call, and coaching calls are evaluated by the scoreboard.

The bill for that depreciation is real and it is now measurable. On August 25 The BMJ published a study led by Daniel Daneshvar of Mass General Brigham, with colleagues at Boston University and the Concussion and CTE Foundation, that counted every former NFL player who died between 2016 and 2021. There were 878. Families donated 235 brains, and 215 of those, 91%, showed chronic traumatic encephalopathy. The authors' conservative estimate assumes that not a single undonated brain had the disease, which gives a floor of 24.5% of all deaths in the period. The 24.5% figure is therefore a floor, not an estimate of the likely prevalence. "At least one in four" is the phrase the coverage settled on, and the qualifier matters, because it is the most favorable reading of the data the league could hope for. Daneshvar describes the mechanism as "the cumulative effect of hundreds, if not thousands, of hits, big and small, over years of playing." Concussion protocols catch the diagnosed events. The disease is built from the routine ones.

So the exposure is known, the long-run cost is known to be large, and the person deciding the exposure faces none of it. In any other industry that combination has a name: an unpriced externality sitting on a balance sheet that belongs to someone else. Here the someone else is the player, and after him the public health system.

Northwest Stadium in Landover, Maryland, during the Atlanta Falcons at Washington Commanders game on December 29, 2024

Northwest Stadium, Landover, Maryland, during Atlanta at Washington, December 29, 2024. Photo by Dissident93, via Wikimedia Commons. CC0 1.0.

The instruments already exist

The proposal is not to legislate courage out of football. It is to route the long-run cost back to the entity that controls the exposure, using instruments that corporate risk management already uses everywhere else, and to make that cost bind competitively rather than just financially.

Start with liability. Franchises should be required to carry third-party coverage for neurodegenerative disease in their former players, underwritten by carriers who set premiums the way carriers set every premium: from the observed behavior of the insured. Underwriters would want the data the league already collects. Designed quarterback runs per game. Snaps played by a lineman in the week after a concussion clearance. Fourth-quarter snaps by starters in games decided by three scores. A coach who runs his quarterback 12 times a game would see it in the premium the following season, the way a fleet operator sees his drivers' telematics in his. The premium becomes a continuously updated score of how much long-run health the coaching staff is spending, computed by a party with money on the line and no interest in the standings.

The first objection an actuary would raise is attribution. A disease diagnosed at 70 in a man who played for four franchises and a university cannot be charged to the coach who ran him on a Sunday in 2026, and no underwriter would try. But apportioning a cumulative injury across employers is not a new problem, and it is already being solved in this exact setting. When Wayne Gandy, a 15-year veteran, filed a cumulative trauma claim in California, it named the Rams, the Steelers, the Saints and the Falcons, and the proceeding's job was to decide what share each owed. Workers' compensation does that arithmetic by exposure: seasons, snaps, position. A liability carrier would do the same, charging each franchise for the exposure it added and adjusting the rate for how the coaching staff spends it. The premium never claims to know which hit caused which disease. It only has to know that exposure is measurable, that the team controls its share, and that the share should cost more when it is spent carelessly.

A premium alone does not change behavior in a league where owners are billionaires and the money is not the constraint. The constraint is the salary cap, so the risk-adjusted portion of the premium has to count against it. A coaching staff that spends its players' futures aggressively would find itself with less room to sign the next one. That converts a health cost the coach never sees into a roster cost the general manager cannot ignore, which is the only currency the building actually trades in.

The third instrument is already standard practice. Teams insure guaranteed money on large contracts. Joel Corry, a former agent who writes on NFL finance for CBS Sports, reported in 2024 that all but one of the 13 quarterbacks earning $45 million or more a year had insured portions of their deals, that the policies pay when a player is "unable to perform the services required by his NFL player contract" through injury, and that a payout comes back to the team as a salary cap credit the following year. Read that last clause again. The one place where insurance already touches the cap, it rewards the team when the insured player breaks. Underwriters can and should write redlines into the same policies: a player returned to the field before a documented clearance, or played through an injury the team's own medical staff flagged, voids the coverage on his contract. If a coach plays an uncleared player and the player is lost for the year, the owner pays the guarantee out of his own pocket rather than the insurer's. Cosmi's $45 million is exactly the kind of number an underwriter would want conditions on, and a lineman nine months removed from a season-ending concussion is exactly the file that would carry them. Nothing about that requires league approval. It requires underwriters to act like underwriters.

None of it requires deciding in advance which plays are too dangerous, or a committee to second-guess Dan Quinn at the one-yard line. The aggressive call stays available. It simply arrives with its true cost attached, estimated by people who estimate such costs for a living, and charged in a unit the coach's employer must spend carefully.

The obstacle is the settlement everyone already agreed to

None of this exists because the current arrangement was built to prevent exactly this kind of accountability. Players' on-field injuries run through workers' compensation, which state law generally makes the exclusive remedy for a workplace injury and which the collective bargaining agreement folds into the players' benefits. The bargain is that the player gets a guaranteed claim without proving fault and the employer gets immunity from suits that would ask whether a specific decision was negligent. The league's 2013 concussion settlement extended the same logic to the disease itself, paying out on a schedule rather than litigating what any team knew or decided. Every part of the structure prices outcomes after the fact and none of it prices exposure before.

Reopening that bargain means reopening the CBA, and everyone at the table has a reason to object. Owners would resist any instrument that makes a coaching decision costly to the franchise. The union has historically fought for compensation and protocols rather than guardrails on coaches, partly because its members want to play and partly because it does not employ the coaches. Insurers would want the data before they would want the risk. None of that makes the design wrong. It makes it a governance change rather than a policy memo, which is what any real fix to an incentive problem turns out to be.

There is a hypothesis circulating this month that the brutal opening fortnight of the season reflects a ramp-up deficit: teams protect their starters from preseason contact, then send them into game speed cold. Whether or not it holds, it shows why exposure has to be priced from outside the building. A coach who rests his starters in August lowers one exposure and raises another, with no instrument for weighing the two, and an underwriter with 32 teams' worth of data would know within a few seasons which choice actually costs more.

Quinn will keep his job, or lose it, on Washington's record. That is the only ledger he has ever been shown. The question this season keeps asking, in Dallas and everywhere else, is why the people who can see the other ledger are not allowed to send anyone the bill.

Further Reading

AI Assistance Statement ▾

A near daily publishing pace is possible because AI tools do a substantial share of the work between the idea and the published text. Preparation of this entry included assistance from Anthropic's Claude and from OpenAI's ChatGPT (GPT-5 series reasoning models). I use them to research a topic and gather primary sources, to organize ideas and propose structure, to draft and revise prose, to check factual claims against the cited sources before publication, and to score drafts against a set of house style rules. Longer pieces are often developed across several sessions. A written handover carries the argument, sources, and open questions from one session to the next, and the same tools help prepare those handovers. The tools also help identify candidate images and confirm that selected images appear to be released for reuse, for example through public domain or Creative Commons licensing.

A fuller explanation of the editorial process appears in How I Use AI to Write This Blog. This process is also an AI experiment in its own right. There is a live argument about what AI-assisted writing does to originality, and whether the result is thought or slop; a noteworthy example is the August 2026 dispute over a Wall Street Journal op-ed that its author acknowledged drafting with AI, and the Journal's subsequent defense of the practice (WSJ is behind a paywall, but for a public summary: click here, and here). I would rather run the experiment openly than pretend it is not happening. This blog is one sustained attempt to find out whether a person with an argument, working with these tools every day, produces writing that is still recognizably that person's, and I disclose the method so readers can judge the result.

The judgment is mine. I choose the topic, decide the argument, supply the personal and professional experience the pieces draw on, read and edit every draft, verify the sources and image licensing, and take full responsibility for the final published content. Where a post contains my own recollections, the AI did not invent them.

Statement revised September 2026.

#Governance #Observations