Our NFL Model's Real Record: Every Prediction, Graded, Published

Here's a number you'll almost never see a prediction service publish: 52.8%.

That's our machine-learning model's lifetime accuracy against the spread — 821 correct, 733 incorrect, 43 pushes across more than 1,500 graded predictions, every one of them stored in our database and displayed on the site, hits and misses alike. In 2025 it went 148–134–3 (52.5%).

If you're used to services claiming 65% winners, that number probably looks underwhelming. That's exactly why we're leading with it.

What 52.8% actually means

The point spread exists to split every game into a 50/50 proposition. It's a crowd-sourced forecast, sharpened by enormous amounts of information, and decades of research show closing spreads are among the most accurate public predictions of game outcomes in any sport. Against a benchmark engineered to be a coin flip, sustained accuracy above 50% is the entire signal — and models that hold even the mid-50s against closing spreads over large samples are genuinely rare.

So 52.8% lifetime is a real, measurable edge in forecasting skill — a slim one, held over 1,500+ predictions against one of the best-calibrated forecasting benchmarks in existence. Anyone quoting you a sustained 60%+ against the spread is describing either a short hot streak or a fiction.

Validation vs. live: the gap nobody talks about

When we build a model, we train it on two decades of games and test it on seasons it has never seen. Our production model scored around 58% on that held-out history. Live, it runs 52.8%. That gap deserves an honest explanation, because every quantitative prediction service has one and almost none disclose it:

  • Hold-out tests are still friendlier than the future. A model is selected because it topped the validation set, so its validation score carries an optimism premium — the winner's curse of model search.
  • The benchmark moves. Historical spreads are static targets; the live market adapts season over season, absorbing the public statistics models feed on.
  • Live seasons are small samples. A model with true skill in the low-to-mid 50s swings multiple points season to season on ~280 predictions.

We report both numbers because the honest claim isn't "our model hits 58%" — it's "our model tested at 58% on unseen history and has delivered 52.8% in live production." The second sentence is the one that describes what you'll actually watch happen each week.

The week-by-week truth: variance is violent

The 2025 season, week by week, is the best variance education we can offer. The same model, the same process, every week:

  • Week 5: 13–1. The kind of week that gets screenshotted and shared.
  • Weeks 2–3: 21–10 combined. A month in, the model looked unstoppable.
  • Weeks 14–18: 32–45. A five-week 41.6% slump — from the same model that went 13–1 in October.
  • Final: 148–134–3.

Any single week of a forecaster's record is marketing. A five-week window is still mostly noise. The season-long number is where signal begins, and the multi-season number is the only one that matters. This is also why we publish the weekly breakdown on the site — the ugly stretches are the proof the good ones aren't curated.

Where the model is genuinely strong: picking winners

Against the spread is the hardest forecasting task in sports, because the line removes most of what a model knows. Straight-up — just predicting the winner — our model has run at 63% (85–50) since we began grading that objective mid-2025, including a 12–4 Week 15 and perfect playoff weeks late in the run. If what you want is a disciplined, stat-driven read on who wins each week's slate, that's where the model's skill is most visible.

(We also tested an over/under objective. After months of feature engineering, totals graded out as near-perfectly calibrated — the hardest target we've pointed the machine at. The totals projections we publish carry our most conservative confidence tiers because of that research, and the full story is its own article.)

How the predictions are made

No gut, no lean, no Tuesday hunch:

  1. Every completed game since 2002 is decomposed into 360 computed trend features — rolling efficiency stats, situational splits, market-derived signals.
  2. A gradient-boosted model (with a logistic-regression cousin as a sanity check) is trained on two decades of history and validated on held-out seasons it never saw.
  3. Each week, the locked model re-trains its weights on the most recent window and scores the upcoming slate. Every prediction gets a confidence rating — 3-star predictions are the top quartile by model confidence.
  4. Results are graded automatically against closing lines and published, hit or miss.

The model has no favorite teams, no narratives, and no memory of last week's embarrassment. That last property, it turns out, is the hardest one for humans to replicate.

Judge for yourself

The complete graded record — by season, by week, by confidence tier — is on the SpreadTrends Predictions page, updated every week of the season. We'd rather earn trust with a real 52.8% than rent it with a fake 65%.

Record as of the end of the 2025 season: 821–733–43 against the spread lifetime; 148–134–3 in 2025; 85–50 straight-up since SU grading began (Week 11, 2025). Pushes excluded from percentages.