One Game Told Us More Than the Whole Summer Did.

Recruiting rankings and returning starters explain about 14% of how a college season turns out. Add a single opening result and it climbs to 21%. We expected the opposite, and we were wrong in an interesting way.

The Sports Page · The Professor · what a small sample is worth

The question: after one weekend, what do we actually know?

The answer: a surprising amount. A club that wins its opener finishes about 1.5 wins better than one that loses it — and that is after accounting for talent and returning production.

Here is the surprising part. We came to this expecting to write that one game tells you almost nothing. That is usually the lesson. This time the data said the opposite, and the reason is not flattering to August.

What you get from it: a better sense of when a small sample is worth listening to.

The naive number, and why it is wrong

Start where most people start. Across 1,297 team-seasons since 2015, clubs that won their opener finished at .583. Clubs that lost finished at .429. That is a gap of 1.8 wins over twelve games.

But that number is inflated, and obviously so. Good clubs win openers because they are good. The teams that won week one had a mean recruiting-talent rating of 606; the losers, 519. Some of that 1.8-win gap is just talent showing up twice.

So we took it out. Fit final winning percentage to the two things anyone could have known in August — the talent a club recruited and the share of last year’s production it returned — and then ask whether the opening result adds anything on top.

It adds almost all of it

The week-one result survives at 1.5 wins. Of the naive 1.8, only about 0.3 was something we already knew in August. The rest is new.

And it is not an artefact of one group. Split the clubs into quarters by recruiting talent and the gap appears in all four — 1.5 wins at the bottom, 1.4, 1.6, and 1.9 at the top. Like meeting like, the opener still separates them.

Why one game moves so much

Here is the part worth carrying away, and it is not really about football.

Talent and returning production together explain 13.6% of final winning. That is a real relationship and a weak one; roughly six-sevenths of a season is coaching, health, schedule and luck. Adding the opening result takes the figure to 20.6%.

That is a 52% increase in what can be explained, bought with a single game. Not because one game is a mountain of evidence. Because the pile it was added to was so small.

This is how new information always works. Its value depends on what you already had. A single measurement lands hard on an empty table and barely registers on a full one, and the same result can be revelatory in September and trivial in November.

And it is wrong about a quarter of the time

Now the other side, because a number this useful invites overreach.

29% of clubs that won their opener finished under .500. 11% of clubs that lost it finished 8–4 or better. Put those together and 294 of 1,297 seasons — about 23% — had the opening weekend pointing the wrong way.

So the honest statement has two halves, and most coverage this week will carry only one. The opener is genuinely informative. It is also going to mislead you about roughly one club in four, and it will not tell you which one.

What to take home: ask what you knew before

When a new piece of evidence lands, do not ask how big it is. Ask how much you already had.

The identical result can be decisive or trivial depending only on that. A first data point in a new area is worth more than a hundredth in a well-studied one. A first impression of a stranger carries more than a fifty-first observation of a colleague — not because first impressions are accurate, but because there is nothing else on the table yet.

That is also the warning inside it. Evidence that arrives when you know least gets the most weight, which is precisely when you are least equipped to judge it. September is when a single game moves your view furthest, and September is when your view is worst.