Week 3 Power 10: Tennessee Topples Penn State and the Data Gap Behind the Upset

On September 21, Penn State — ranked No. 9 in NCAA.com's Power 10 — lost 3-1 to Tennessee. The recap published by the school itself pinned the defeat on "unforced errors." That is it. No set scores, no specific error count, no attacking efficiency figures for the opponent.

Individual stat lines were provided. Gabrielle Nichols recorded 38 assists and 12 digs — her third double-double of the season. Ava Falduto led the team with 15 digs. Ryla Jones was named but given no stat line. Tennessee, the winning side, vanished entirely from the data.

Data is like dust: it only means something when you are calm enough to look through it. Here the dust is thick enough that I have to say it plainly — the problem does not yet have enough variables to solve.

What the Power 10 is, and what it is not

This needs clarifying immediately, because it is the point most readers skim past and misread. The Power 10 is a ranking curated by hand by an NCAA.com analyst, Michella Chester, and updated weekly. It does not belong to the AVCA Coaches Poll system, it is not the RPI, and it plays no part in selecting the 64-team NCAA Tournament field in December. The NCAA selection committee uses RPI plus direct evaluation; it does not use the Power 10.

In other words, Penn State leaving the Power 10 is a media event. Competitively, it costs them nothing.

Week 3 of the US women's college season falls in the non-conference window — matches outside the conference schedule. This is when rankings swing hardest, because no team has enough of a sample to prove anything yet. A good win lifts a team two spots, a bad loss drops it three, and neither says anything about November.

Based on my experience tracking matches across both volleyball and football, the defining feature of early season is the enormous gap between the noise of the rankings and the thickness of the evidence. Week 3 sits in the noisiest bracket.

Two numbers worth dissecting

Back to the match. We have two data points worth pulling apart: 12 digs from a setter, and 15 digs from the team leader.

Setter is a distribution position. In a 3-1 loss, a setter recording 12 digs and finishing second on the team in defense indicates a high frequency of balls dying in the back court. The team had to keep digging, rallies stretched, and in transition they converted poorly. That is the signature of an overloaded defensive system: good digging that produces no points off those very digs.

The phrase "unforced errors" in a university's own recap almost always means self-inflicted serving and attacking errors rather than pressure created by the opponent. But it is a symptom, not a diagnosis. With no error count, no perfect-pass rate, no attacking efficiency, we do not know which phase the errors concentrated in, nor when they landed.

I have misread a match in exactly this way. In 2026 I predicted Sanna Khanh Hoa BVN would beat Hanoi FC 2-0 in the V-League on "strong form." Hanoi won 4-1. But Khanh Hoa's xG was higher — 2.8 against 2.1. They were unlucky, not inferior. My article got the nature of the match completely wrong.

I deleted it, sat down, coded all 38 rounds of the 2026 V-League season, and learned to calculate xG shot by shot. The model was wrong; I do not blame the data, I blame myself for trusting it blindly.

That lesson applies intact to Penn State. The biggest problem here is that the dataset runs in one direction only. Tennessee's statistics, for a team described as "resume-building," are entirely absent. Without set scores, we cannot even tell whether this was a narrow 23-25 defeat or a rout. The two data points needed to validate the "unforced errors" thesis are the error count and the set scores — neither exists.

If volleyball had an equivalent of football's PPDA — passes allowed per defensive action — I would have measured it long ago. It does not. So the most honest thing I can say is that confidence in any tactical conclusion drawn from this report is low.

That night I watched Germany press and understood that a champion is only ever a variable. Germany fell in 2026 because their pressing lied, not because they lacked talent. Tennessee this week is the inverse — they won a match in which Penn State lied to itself, and that self-deception is not measured by any number.

One data point does not build a tier

What the report calls "Tennessee entering the sport's top tier" is a claim built on exactly one data point.

Remember: the Power 10 carries the personal imprint of one analyst. Structurally, its week-to-week volatility is higher than the AVCA Coaches Poll and higher than RPI, because it is designed to react fast to new results. One match can move a team in, one match can move it out.

Which means both readings are being pushed too far.

The first reading — Penn State in decline — has no basis. This was their first loss to a ranked opponent this season and their first absence from the Power 10 this year. Their sustained presence in prior weeks reflects a real quality baseline that a single September 21 defeat cannot erase. Penn State's true risk sits in their RPI resume, not in an editorial ranking.

The second reading — Tennessee belongs among the elite — is also a single data point. In the history of US college volleyball, "top tier" declarations born from one non-conference win tend to fade quickly once conference play begins and real defenses show up. TCU and Tennessee entering the top 10 in the same week suggests this is a structural rearrangement of perception, not the story of one team.

One more thing my model cannot see: US college sport runs on student emotion and recruiting pressure. A "top tier" label has real value to a 17-year-old recruit, even before it is validated on court. The recruiting market reacts faster than the RPI.

I do not bet on passion; I bet on probability verified three times. In this case, there has not been one verification.

Signals for the next round

Three things to track. The set scores from September 21, if a full box score is released. The positions of Tennessee and TCU in the Week 4 and Week 5 Power 10. And Penn State's results once conference play begins, where RPI starts accumulating for real.

What I want to keep from this piece is not who moved up or down. It is this: an editorial ranking can shift the perception of thousands of people within 48 hours, while verifiable data takes weeks to shift belief. Whoever can measure that lag reads the market before the market reads itself.