2026 season · Version 1
Every game on NC State's schedule, predicted before the season started and frozen exactly as it was written.
This is the model's first attempt at a whole season. All twelve games, predicted before a single snap was played, written down on July 30 2026 and never recalculated since. When it turns out to be wrong — and it will, repeatedly — the record of what it said beforehand is still sitting here unchanged.
What it does reasonably well is margin: who wins, and roughly by how much. That comes from every play of every FBS team last season, adjusted for the quality of who it happened against, then regressed toward the average to account for a year of roster turnover. It is a real method, not a hunch.
Known limitation — Version 1
It cannot predict total points, so it doesn't pretend to. Add up any score below and you get 52.9 every time. That isn't a bug or a coincidence — it's the 2025 league average, used as a stand-in because our own attempts at predicting totals scored worse than simply guessing that average. So the model supplies the margin, the league average supplies the total, and the two scores are worked out from those.
It means the margins are worth reading and the individual scores are worth much less. A proper total needs a separate scoring-rate model — drives per game, points per drive — which is a different build rather than a rearrangement of these numbers. It's first on the list for Version 2.
The point of this page isn't to be right. It's to be a fixed, checkable record of what a homemade model said in July, so that in December you can see precisely where it went wrong and what changed in response. This is a model being built in public, flaws included — the flaws are arguably the more interesting half.
Loading the frozen forecast…
Still to come on this page
