Best and Worst NFL Stadiums by Adjusted Field-Goal Accuracy Since 2018

Kicking environments are not created equal — or at least the kicks aren't. We measured every NFL regular-season field-goal attempt from 2018 through the completed 2025 season (8,241 kicks, league make rate 84.6%) against a leave-one-season-out expectation built only from kick distance and season, then ranked stadiums by how many makes they produced above or below that expectation.

The headline ranking

Among the 28 eligible stadiums (at least 80 attempts and 40 games), the largest positive adjusted residuals per attempt belong to Allegiant Stadium (+0.0660, 95% CI 0.0039 to 0.0861, 214 attempts), AT&T Stadium (+0.0543, 95% CI -0.0025 to 0.0819, 272 attempts), Mercedes-Benz Stadium (+0.0444, 95% CI 0.0020 to 0.0646, 301 attempts), and SoFi Stadium (+0.0419, 95% CI 0.0055 to 0.0617, 407 attempts). On the negative side, the biggest deficits are FirstEnergy Stadium (-0.0675, 95% CI -0.1002 to 0.0066, 226 attempts), Gillette Stadium (-0.0617, 95% CI -0.0885 to 0.0008, 245 attempts), and Raymond James Stadium (-0.0544, 95% CI -0.0753 to -0.0029, 255 attempts).

What the intervals actually say

Per the deterministic exclusion summary, exactly three eligible stadiums have 95% intervals excluding zero on the positive side — Allegiant, Mercedes-Benz, and SoFi — and exactly one excludes zero on the negative side: Raymond James. But none of the 28 eligible stadiums is flagged significant under the Holm correction across the full family; the smallest Holm-adjusted p-value is 0.535 (SoFi). Treat these as suggestive stadium effects, not statistically confirmed ones.

How stable is the ranking?

The ordering holds up well to most stress tests: eligibility thresholds of 60 and 120 attempts each give Spearman 1.0 versus the primary ranking with full top-5 and bottom-5 overlap, alternate distance binning gives Spearman 0.996, and turning shrinkage on or off gives 0.998. The one soft spot is the time window: restricting to 2021–2025 drops rank stability to Spearman 0.844 with only 4/5 top-5 and 3/5 bottom-5 overlap, so the 2018–2025 ordering may not reflect the most recent seasons alone.

Why it matters for fantasy

If you're weighing two kickers with similar leg strength and matchups, a road game in Las Vegas, Dallas, Atlanta, or Los Angeles has historically come with a small accuracy tailwind after distance is accounted for, while trips to Cleveland, New England, or Tampa have come with a headwind. The effect sizes are modest — roughly five to seven makes per 100 attempts at the extremes — and because no stadium survives Holm correction, this is a tiebreaker, not a reason to bench a clearly better kicker. The adjustment controls only for distance and season, so residual 'stadium' effects may partly reflect unmodeled weather or venue characteristics rather than the building itself.

Caveats worth knowing

• 17 of the 45 analyzed stadiums are ineligible for the primary ranking (fewer than 80 attempts or 40 games), including several with extreme point estimates on tiny samples — Deutsche Bank Park at -0.221 per attempt on 5 kicks, Arena Corinthians at -0.105 on 6 — whose intervals are degenerate or extremely wide and cannot support inference.
• The expectation model uses only kick-distance bins and training-season outcomes; stadium identity never enters it, and raw per-stadium make percentage is reported for transparency only, never for ranking.
• Uncertainty comes from 200 game-clustered bootstrap resamples with 95% intervals and Holm-adjusted p-values across the 28 eligible stadiums; shrinkage toward the league mean used a pre-locked rule (k = 100).

The bottom line

After adjusting for kick distance and season across the 2018–2025 regular seasons, Allegiant Stadium shows the largest positive field-goal residual among eligible stadiums and FirstEnergy Stadium the largest negative, with only Allegiant, Mercedes-Benz, and SoFi clearing zero on the upside and only Raymond James on the downside. Because no stadium survives Holm correction, use these splits as a soft tiebreaker when comparing otherwise similar kickers and matchups — not as a reason to override a clear talent or workload edge.