Skip to main content
MissedTakes
Menu

2026 NFL Season Beta

We Keep Score on the People Keeping Score.

Society is inundated with pundits, making daily predictions and sharing opinions on everything from sports to politics. MissedTakes, a beta-phase app from The GoodApps Company, is testing a grading model for understanding pundit takes and grading their accuracy over time. Our 2026 NFL beta ingests pundit opinions from around the internet in various multimedia formats, parses them for structured takes, and reconciles them against draft, team, player, season, and award actuals.

The leaderboard →NFL predictions →How we score it →

Takes logged
27,407

across 30 kinds of prediction

Graded so far
10,011

checked against real results

Pundits tracked
760

across 45 outlets

Average accuracy
60.0

out of 100, all graded takes

Visualizing NFL Pundit Accuracy

Why so many zeros? →
This is a visualization of every graded pundit call we’ve analyzed. You can see the middle band is very thin: predictions tend to be simply right or wrong, so the challenge for our data is to see if we can assign range data that shows how right and wrong they were, along with the risk they took for disagreeing with consensus and formulating takes long before having data that better-guarantees accuracy.

Takes in each band

020406080
0 — dead wrongscore out of 100100 — nailed it
5,172 scored 80 or better3,459 scored under 40

The State of the 2025 Season

Ingesting pundit takes is only half the battle. Here’s the current state of our reconciliation of pundit takes against actual season performances. The gap closes during a few major season milestones, including:
  • The NFL draft reconciles mock draft takes
  • The preseason reconciles roster spots, trades, free agent signings, etc.
  • The season reconciles win/loss totals, along with fantasy football performances
  • The playoffs reconcile final power rankings and season awards (like MVP, Rookie of the Year, Comeback Player of the Year, Offensive and Defensive Player of the Year, All-Pro, etc.)
  • The post-season reconciles coaching changes and further free agency

10,011 / 27,407

takes graded

17,396 still pending a result

Top-Performing Pundits by MissedTakes Score

Full leaderboard →
Simple accuracy scoring is one thing, but the MissedTakes score improves upon it by measuring a take against the accuracy of a naive guess, so the net result is a score that shows whether a pundit’s take outperforms random noise.

Confidence so far: provisional— nobody has enough articles behind them for a settled rating yet.

Our Current Model Tracks 34 Kinds of Prediction

How each is graded →
The NFL is our proof of concept for this kind of modeling. Pundit calls are published constantly and we parse for unambiguous takes that can be reconciled against definitive results.

How It Works

  1. We Log Pundit Takes

    Whenever a pundit predicts something we can later reconcile against (draft order, fantasy scoring, win totals, etc.), we record it and provide users with a link to where it was said. Receipts are for winners.

  2. The Season Accumulates Data

    Players, teams, and coaches deliver actuals over the course of the season through wins, losses, fantasy points, and more. We pull the results and reconcile them against pundit takes.

  3. We Deliver Modeled Scores

    When graded takes come in, pundit scores adjust: overall, per-player, per-team, per take type, and delivering an over-time and aggregate accuracy and MissedTakes score.

No, We’re Not Out Here With Pitchforks.

Predicting football is genuinely hard. Everything from injuries and weather to a bad bounce on a third down makes it a nearly-impossible sport to predict. Pundits are reasoning about countless variables with incomplete information, limited time, and in public forums. Being wrong is in the job description and we don’t judge it as a character flaw.

What we do is show the math, link every source, and say plainly when we think someone excels above the rest. Who really has a unique eye for talent? A quality take on an underdog? And when we don’t have enough data, we’re transparent about not having enough predictions to draw a legitimate conclusion about accuracy.

What To Expect From the Beta

In order to go big, you have to start small. We’re taking our tech and testing it on the 2026 NFL season. What we learn from this beta will enable us to expand, not just to other sports and leagues, but to news, politics, and beyond.