Talent or scheme

You can watch this happen somewhere every October.

A coordinator has a year where the offense hums. He gets head-coaching interviews. Two years later the roster has turned over, the offense is average, and the same people say he was never that good — the players were covering for him.

Nothing got measured in between. The verdict flipped because the personnel did.

The argument nobody can win

Sunday morning. Twelve plays. The run game produced nothing.

The line coach says the backside end is unblocked by design and the back has nowhere to go. The coordinator says the play's been sound for fifteen years and the guard is getting walked into the mesh point.

Both have film. Neither can prove anything.

So it goes the way it always goes — longest tenure, most recent win, whoever the head coach came up with. That's a fine tiebreaker. It isn't evidence.

What changed is the cost of being wrong. That argument used to decide Tuesday's reps. Now it decides who gets paid, who gets replaced in December, and which coordinator gets a job.

Money buys the range, not the result

AP-ranked teams average around $35M in roster spend this year. Unranked teams average around $22M. So money matters.

Then look inside the band. Georgia was ranked second nationally at an estimated $31–34M. Florida, Nebraska, Auburn and South Carolina all spent $31M or more and finished unranked.

Across nine seasons and 131 FBS programs, the relationship between spending and production is null. Not weak. Absent.

So the gap everyone argues about is real. There's a large, unexplained space between what a roster costs and what it produces. Everybody assumes that space is coaching.

Baseball just ran the experiment

Milwaukee swept Los Angeles out of the NLCS last Thursday. The Dodgers' payroll is about $339M. Milwaukee's is about $146M. One Dodger makes $70M by himself — nearly three times Milwaukee's highest-paid player at $26M. Eighteen Dodgers make $5M or more. Seven Brewers do.

Easy read: smart small-market team beats dumb rich one. Wrong, and the way it's wrong is the point.

The Athletic polled 40 executives on the best front offices in baseball this year. Top four: the Dodgers, Rays, Brewers and Guardians.

Those are the same four teams that made the league championship series.

So the sweep wasn't the small-market team beating the rich one. It was an elite front office with $339M losing four straight to an elite front office with $146M — in a final four that the front-office rankings called before a pitch was thrown.

Four games proves nothing on its own; I'll argue in a minute that small samples are where people fool themselves. The evidence is that Milwaukee was the best team in its league for six months at 43 cents on the dollar, and that the four best-run organizations in baseball were the last four standing.

But notice what baseball can do that football can't. You can poll 40 executives on front-office quality and get an answer they'd broadly agree on, because baseball spent twenty years building the measurement underneath that judgment. Run that poll in college football and you get name recognition and last year's record.

Both instruments are half-blind

Football isn't short on data. It's short on data that answers this question.

A grade tells you how a player did the job he was handed. It can't tell you whether it was the right job. A guy executing a bad call grades badly. A guy protected by a good call grades well. The plan disappears into the player's number.

Call data has the opposite hole. You know the concept averaged four yards. You don't know whether four was good for that concept, with those five linemen, against that front. Four yards with a true freshman at left tackle and four yards with a returning all-conference tackle are not the same play.

Grades hide the call. Call data hides the player. That's why two coaches watch the same twelve snaps, both holding real evidence, and walk out with opposite answers.

The sample doesn't help either. Baseball settles arguments over 162 games. A college season is twelve. Across 4,967 close games, the spread of records looks about like what you'd get flipping coins. Coaches get fired over that tail.

It was never a coaching question

Line the two sources up snap by snap and the question changes. Not who was wrong. How well did the plan and the players fit each other.

That used to be the end of it. Fit was something you coached, because you had the roster you had.

Not anymore. Watch the best version of this now.

A staff is built around an identity — whatever the coordinator has run for twenty years. The roster doesn't fit it. The prized recruit at the trigger isn't producing, but there's a back who is. Mid-season, they abandon the identity. The offense becomes something that coordinator has never been known for, and it works. Then in the offseason they go to the portal and buy the quarterback who fits the offense they became.

That's not scheme. That's not talent. That's managing fit across a season and an offseason with money as the instrument — and it's probably the most valuable thing a staff can do right now.

It's also completely unpriced. Nobody has a number for how fast a staff caught the mismatch, what it cost while it lasted, or whether the money spent closing it was spent well. We grade the players. We chart the calls. The thing in between, where the decision actually lives, has no instrument at all.

So the problem with "talent or scheme" isn't that it's a false binary. It's that it was never a coaching question. It's a roster construction question wearing a coaching costume. Every answer implies a purchase — a player signed, a contract renewed, a coordinator kept or fired — and those purchases get made on an argument settled by tenure.

Which cuts both ways, and that's why it's hard to adopt. Sometimes the answer is that a unit executed fine inside calls that didn't work, and the problem is upstream of the players. Grades point outward at people who don't get a vote. This points in both directions, and one of them is the guy deciding whether to use it.

Here's the narrow version I'd defend anywhere: the argument every staff in America is having this morning can't be settled with what those staffs keep. Both sides are honest and both sides are half-blind. The side that wins has more tenure in the room — and the bill now shows up in dollars.

Next in this series: the constrained-program case. Everything above costs money to act on. Most programs don't have it.  All research →