Big Game AI Prediction

I used advanced AI through the Genuine Technologies chat to create an AI prediction that would be based off of both teams season, their individual players, and take into account the difficulty of their past games into account to how much those games should factor into the performance of the teams during Superbowl 2026

Here is the AI prompt used with GPT 5.2 Pro, level 10 extreme thinking, medium sized web content.

The prompt took about 20 minutes to finish, 140,613 tokens total, and $7.98 in API usage costs.

Result:

Context (what “Super Bowl 2026” means here)
The Seahawks and Patriots are scheduled to play in Super Bowl LX on Sunday, February 8, 2026 at Levi’s Stadium (Santa Clara, CA). (en.wikipedia.org)

 
1) Team win–loss research + head-to-head comparison
Regular-season records (2025 season)
Both teams finished 14–3. (espn.com)

Overall records entering the Super Bowl (regular season + playoffs)
Seahawks: 16–3 (2–0 postseason due to 1st-round bye) (espn.com)
Patriots: 17–3 (3–0 postseason; they played Wild Card weekend) (espn.com)
Points scored/allowed (regular season)
Seahawks: 483 PF / 292 PA (28.4 scored, 17.2 allowed) (espn.com)
Patriots: 490 PF / 320 PA (28.8 scored, 18.8 allowed) (espn.com)
 
2) Accounting for “difference in skill level” (schedule-adjusted strength)
Raw W–L is identical, so the key adjustment is quality of opposition. Two strong, simple proxies:

Schedule strength (SOS) and overall team rating (SRS)
Seahawks: SRS 12.85 (1st), SOS 1.61 (pro-football-reference.com)
Patriots: SRS 5.51 (6th), SOS -4.49 (pro-football-reference.com)
Interpretation (practical): Seattle’s 14–3 was achieved against a meaningfully tougher slate, and their margin-based rating (SRS) is much higher. That is the cleanest “difficulty adjustment” available without doing play-by-play opponent adjustments yourself.

 
3) “Track the players’ stats of each of those games” (a manageable, decision-relevant view)
Doing every player’s full 17-game log for both teams would be enormous. The most predictive slice for a one-game Super Bowl forecast is usually:

Each QB in team losses (what causes the floor outcome), and
Each QB/RB/WR in the postseason (closest to Super Bowl intensity).
QB performance in team losses (how the offense fails)
Patriots (Drake Maye) — 3 regular-season losses

vs Raiders (L 13–20): 287 pass yds, 1 TD, 1 INT, RTG 80.6 (statmuse.com)
vs Steelers (L 14–21): 268 pass yds, 2 TD, 1 INT, RTG 102.1 (statmuse.com)
vs Bills (L 31–35): 155 pass yds, 0 TD, 1 INT, RTG 62.8 (statmuse.com)
Seahawks (Sam Darnold) — 3 regular-season losses

vs 49ers (L 13–17): 150 pass yds, 0 TD, 0 INT, RTG 87.2 (statmuse.com)
vs Buccaneers (L 35–38): 341 pass yds, 4 TD (high ceiling even in a loss) (seahawks.com)
at Rams (L 19–21): 279 pass yds, 0 TD, 4 INT, RTG 45.5 (statmuse.com)
What this implies:

Maye’s “bad” games tend to be low-output/efficiency dips, but not typically a meltdown.
Darnold’s distribution is wider: he can put up explosive TD volume (Bucs) or have a multi-INT collapse (Rams). That volatility matters against a Patriots defense that plays well in coverage.
 
4) Use only players not injured for Super Bowl LX (final injury designations)
From the final injury report statuses going into the game:

Patriots questionable: LB Robert Spillane, OLB Harold Landry III, DT Joshua Farmer (IR/return window). QB Drake Maye is off the injury report. (patriots.com)
Seahawks questionable: FB Robbie Ouzts (only Seattle player with a game-status designation). (patriots.com)
Assumption (explicit): To follow your rule “use only players that are not injured,” I exclude the questionable players above from the matchup-impact analysis. Everyone else I reference below is treated as available.

 
5) Player/stat comparisons (healthy players only) and opponent-difficulty logic
Quarterbacks (primary driver)
Drake Maye (NE) — season + playoffs

Regular season: 4,394 pass yds, 31 TD, 8 INT, 113.5 RTG (statmuse.com)
Playoffs (3 games): 533 pass yds, 4 TD, 2 INT, plus 141 rush yds, 1 rush TD; but also 15 sacks taken. (espn.com)
Difficulty adjustment cue: ESPN notes Maye won three playoff games vs top-5 defenses (Chargers #5, Texans #1, Broncos #2). (espn.com)
Sam Darnold (SEA) — season + playoffs

Regular season: 4,048 pass yds, 25 TD, 14 INT, 99.1 RTG (statmuse.com)
Playoffs (2 games): 470 pass yds, 4 TD, 0 INT, including 346 yds, 3 TD in the NFC title game. (espn.com)
Matchup implication: Maye is the more efficient regular-season passer, but the “pressure/sack” trend is concerning against a Seattle front that creates negative plays.

Running games (healthy backs)
Seahawks RB Kenneth Walker III

Regular season: 1,027 rush yds (4.6 YPC), 5 rush TD; plus receiving usage (statmuse.com)
Playoffs: 178 rush yds (4.7 YPC), 4 rush TD (espn.com)
Patriots RB TreVeyon Henderson

Regular season: 911 rush yds (5.1 YPC), 9 rush TD (statmuse.com)
Playoffs: 24 carries, 57 yds (2.4 YPC), 0 TD (espn.com)
Patriots RB Rhamondre Stevenson

Regular season: 603 rush yds, 7 rush TD (14 games) (statmuse.com)
Difficulty adjustment cue: Henderson’s efficiency has dropped sharply in the postseason—exactly when he faced playoff-caliber fronts—suggesting Seattle’s top-tier run defense is a particularly bad stylistic matchup.

Primary pass-catchers (healthy)
Seahawks WR Jaxon Smith-Njigba

Regular season: 119 rec, 1,793 yds, 10 TD (NFL leader in yards) (statmuse.com)
Playoffs: 13 rec, 172 yds, 2 TD (including 10–153–1 vs LAR) (espn.com)
Patriots WR Stefon Diggs

Regular season: 1,013 rec yds (statmuse.com)
Playoffs: 11 rec, 73 yds, 1 TD (espn.com)
Patriots TE Hunter Henry

Regular season: 60 rec, 768 yds, 7 TD (statmuse.com)
Seahawks supporting (healthy)

WR Cooper Kupp: 47 rec, 593 yds, 2 TD (statmuse.com)
TE A.J. Barner: 52 rec, 519 yds, 6 TD (seahawks.com)
Matchup implication: Seattle has the single biggest “win-your-route-anywhere” receiving piece (JSN) plus enough complementary options to punish double teams. New England’s pass defense is strong, but Seattle’s ability to generate explosive plays is more concentrated in one player.

 
6) Trenches + defense (often the Super Bowl swing factor)
Pressure and protection indicators
Seattle defensive sacks: 47; New England defensive sacks: 35 (statmuse.com)
Patriots sacks allowed (offense): 48; Seahawks sacks allowed: 27 (statmuse.com)
Turnover context
Turnover differential: Patriots +3, Seahawks -3 (espn.com)
Matchup implication: Seattle looks more likely to create sacks/negative plays, while New England looks more likely to win turnover margin. Those factors often oppose each other; I treat sacks/pressure as slightly more predictive here because Maye has taken heavy sack volume in the playoffs. (espn.com)

 
7) Score modeling (strength-adjusted and injury-filtered)
Baseline points projection (offense PPG vs opponent defensive PPG allowed)
Using regular-season scoring rates:

Seahawks expected points ≈ average(SEA 28.4, NE allowed 18.8) = 23.6
Patriots expected points ≈ average(NE 28.8, SEA allowed 17.2) = 23.0 (statmuse.com)
Adjustments (to reflect your “difficulty accounted for” requirement)
Schedule-quality adjustment: Seattle’s SRS/SOS advantage (12.85 / 1.61 vs 5.51 / -4.49) implies SEA’s “true” baseline is stronger than raw W–L suggests, so I nudge Seattle up and New England down slightly. (pro-football-reference.com)
Pressure vs protection: SEA’s pass rush vs NE’s sack rate trends lower NE scoring efficiency (more stalled drives, longer down-and-distance). (statmuse.com)
Turnover offset: NE’s positive turnover differential and Maye’s low INT count partially offsets the above. (espn.com)
Net result: a low-to-mid 40s total with Seattle winning by about one score, which is consistent with the market spread/total (SEA -4.5, O/U 45.5) as a reasonableness check. (foxsports.com)

 
Prediction (winner + final score)
Seattle Seahawks 24, New England Patriots 20.

Why Seattle, in one sentence
Seattle’s schedule-adjusted team strength (SRS/SOS) plus a major trenches edge (47 sacks vs NE’s 48 sacks allowed) is the most reliable path to a narrow win in a matchup where both defenses are top-tier. (pro-football-reference.com)

If you would like, I can also produce individual stat-line projections (passing yards/TD/INT, rushing attempts, top receiver yards) for the specific healthy skill players used above, using the same opponent-strength adjustments.