> ## Content Index
> Fetch the complete content index at: https://www.statsdigest.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Minnesota at Tampa Bay, Week 3: What the Numbers Actually Say
- URL: https://www.statsdigest.com/min-at-tb-week-3-2026-quantitative-preview/
- Published: 2026-09-25T16:08:40.000Z
- Updated: 2026-09-25T16:08:40.000Z
- Description: Tampa Bay's offense is the one number that doesn't match its 2025 baseline — the rest is noise.
- Author: Bill Seifert
- Tags: Vikings, Buccaneers, NFL Week 3, EPA, betting odds, NFL analytics

The market has Minnesota as a 1.5-point road favorite in Tampa on Sunday, with a total of 42.5\. That is a close-to-pick line, which is the correct place to start: two games in, there is not much these teams have done that is trustworthy on its own. This piece is a breakdown of how the game is likely to be decided — which units matter, which matchups are measurable, and where the early-season numbers should be discounted.

## Table of Contents

1. The market and a model number
2. Two-game EPA, with play counts
3. Opponent adjustment — the whole story of week 2
4. Where each offense throws, and what that implies
5. Situational and run splits

---

## 1\. The market and a model number

Kickoff is week 3 in Tampa. The closing-line data has TB at +1.5 with a 42.5 total (home perspective; MIN favored by 1.5).

Our team-spread model projects TB +2.44 against that -1.5 line, a gap of about 3.9 points toward Tampa versus the market. Two structural notes before anyone reads too much into that:

A descriptive spread built from 2025 net-EPA ratings using the `spread_projection` recipe lands at TB -1.8 in this game — essentially the market number. In other words, the disagreement between the model and the line is coming from what has happened in the last two weeks, not from either team's 2025 body of work.

## 2\. Two-game EPA, with play counts

Through two games, cumulative EPA/play:

| Team | Side    | EPA/play | Plays | Pass EPA/play                 | Rush EPA/play      |
| ---- | ------- | -------- | ----- | ----------------------------- | ------------------ |
| MIN  | offense | \-0.047  | 110   | +0.020 (weighted from weekly) | \-0.108 (weighted) |
| MIN  | defense | \-0.150  | 136   | —                             | —                  |
| TB   | offense | \-0.164  | 113   | \-0.304 (weighted)            | +0.048 (weighted)  |
| TB   | defense | +0.003   | 115   | —                             | —                  |

Week-by-week the picture is choppy — as it should be at this sample:

- **MIN week 1** (39-22 win over GB): offense -0.020 on 59 plays, defense -0.175 on 67.
- **MIN week 2** (9-3 win over CHI): offense -0.078 on 51 plays, defense -0.125 on 69.
- **TB week 1** (33-27 loss at CIN): offense -0.185 on 51 plays, defense -0.002 on 62.
- **TB week 2** (23-19 win over CLE): offense -0.146 on 62 plays, defense +0.009 on 53, with a pass-defense EPA/play of +0.324.

Denominators of 51-69 plays are small enough that a single explosive drive can swing the number a tenth. Read these as game-level facts, not season-level rates.

For comparison to a real sample — full 2025 seasons:

- **MIN 2025**: offense -0.115 EPA/play (943 plays), defense -0.101 (1016).
- **TB 2025**: offense -0.009 (1064), defense +0.013 (990).

Minnesota's defense number so far (-0.150) is directionally consistent with last year (-0.101), maybe a touch better. The offense (-0.047) is *better* than 2025's -0.115\. Tampa's defense (+0.003) is right on top of 2025's +0.013\. The one apparent departure from baseline is Tampa's offense: -0.164 through two games versus -0.009 across all of 2025\. That is the finding worth building the game around, so long as we handle the schedule honestly.

## 3\. Opponent adjustment — the whole story of week 2

An unadjusted two-game comparison is often just a schedule comparison. Look at the four opponents through week 2:

| Opponent (of…) | Off EPA/play (2026) | Plays | Off rank | Def EPA/play (2026) | Plays | Def rank |
| -------------- | ------------------- | ----- | -------- | ------------------- | ----- | -------- |
| GB (MIN wk1)   | \-0.195             | 117   | 29       | \-0.067             | 130   | 11       |
| CHI (MIN wk2)  | +0.144              | 139   | 7        | +0.064              | 113   | 21       |
| CIN (TB wk1)   | \-0.045             | 120   | 19       | \-0.222             | 127   | 2        |
| CLE (TB wk2)   | \-0.084             | 102   | 23       | +0.077              | 115   | 24       |

The distortion is real and it cuts both directions.

**MIN's defense (-0.150 EPA/play allowed) faced two below-average offenses.** Green Bay ranks 29th on offense through two games (-0.195 EPA/play, 117 plays), and Chicago, despite ranking 7th on offense overall, was held to their week-2 line largely because Minnesota was on the other sideline in the game the aggregate contains. Take the MIN defense number at face value and you are crediting the schedule too.

**TB's offense (-0.164) faced two above-average defenses**, one of them a lot better than average. Cincinnati is second in the league in defensive EPA/play allowed through two games (-0.222 on 127 plays). Cleveland is 24th but held Tampa to +0.009 EPA/play on 62 offensive plays. Every offense that has stepped on the field against CIN this year has looked worse for it — TB's -0.185 EPA/play in that game is heavily colored by opponent.

The two-game picture the model has to work with is: Minnesota's defensive number is inflated by opponent quality, and Tampa's offensive number is depressed by it. Both regress toward what 2025 said about these teams (defense: MIN better; offense: TB a hair better than MIN). That is why the descriptive rating-season projection is close to the line and the recent-form model pulls further toward Tampa: recent form is largely a schedule artifact.

## 4\. Where each offense throws, and what that implies

Air-yards splits from the play-by-play, pass plays only (sacks excluded), 2026 through week 2:

**Minnesota (44 charted attempts):**

- Behind LOS: 10 attempts, avg air yards -3.9, -0.07 EPA/play
- Short (0–9): 18 attempts, +0.19 EPA/play
- Intermediate (10–19): 12 attempts, +0.41 EPA/play
- Deep (20+): 4 attempts, +0.26 EPA/play

**Tampa Bay (62 charted attempts):**

- Behind LOS: 18 attempts, -0.26 EPA/play
- Short (0–9): 19 attempts, +0.29 EPA/play
- Intermediate (10–19): 14 attempts, -0.07 EPA/play
- Deep (20+): 11 attempts (avg air 33.9), -0.26 EPA/play

Two things stand out and both should be flagged as small-sample.

First, Tampa has tried a *lot* of deep shots — 11 of 62 charted attempts are 20+ air yards, at an average depth of nearly 34 yards. They have not connected on them profitably (-0.26 EPA/play), but the volume itself is a stylistic tell: the offense is stretching the field. If MIN's coverage is strong on the top and soft underneath, that gets uncomfortable. Minnesota's own defensive-depth profile from the larger 2025 sample would be the natural check here — I wasn't able to pull the bucketed 2025 splits cleanly this session, so I'll flag the specific matchup as "worth watching, unresolved from data" rather than manufacture the number.

Second, the interior of the field has been the productive zone for Minnesota's offense (+0.41 EPA on 12 intermediate attempts) and the *unproductive* zone for Tampa's (-0.07 on 14). Twelve and 14 attempts respectively is a shaky base for anything definitive, so I'd read this as consistent with the aggregate — Minnesota's passing game has been the more efficient of the two through two weeks — rather than as its own standalone matchup edge.

## 5\. Situational and run splits

Two-week rushing EPA:

- MIN rush offense: cumulative -0.108 EPA/play (weekly weighted). Neither game showed a real run game — -0.054 vs GB, -0.160 vs CHI.
- TB rush offense: cumulative +0.048 EPA/play. Uneven, but the week-2 mark against CLE was +0.180 — one of the few clean things Tampa's offense did against a defense that was giving up nothing through the air (their pass EPA that game: -0.366).
- MIN rush defense: -0.429 vs GB, -0.160 vs CHI. Both very good.
- TB rush defense: +0.097 vs CIN, -0.470 vs CLE.

The Tampa rush-defense line is the loudest small-sample number in the whole preview: -0.470 EPA/play against CLE on the defensive side is not a repeatable rate; it's a game where Cleveland could not run. Take it with the appropriate skepticism.

The most stable read across both teams: Minnesota's defense has stopped the run at a real level in both games, and Tampa's offense has needed the ground game to survive against tough pass defenses. If MIN takes that away — and their two games say they can — Tampa's path narrows to the same deep passing game that hasn't converted yet.

## Where this leaves things

If you had to write the game in one paragraph:

> Minnesota's defense number is real but flattered by two below-average offenses; Tampa's offense number is bad but faced two above-average defenses including the league's stingiest through two weeks. Regress both toward the 2025 baselines and the teams look closer to even, which is roughly where the market has priced it. What breaks the game either way is likely one of: Tampa's deep passing volume finally landing (11 attempts of 20+ air yards without a payoff so far, -0.26 EPA/play), Minnesota's interior passing game continuing at its current +0.41 EPA/play clip against a Tampa defense that has been soft in the middle of the field, or the run-game asymmetry where MIN has stopped the run in both games and TB has needed it to score.