Two teams out of each league go straight to Lane. Everyone else is chasing two at-large spots.
Chute plays the rest of the season out, from here to the finish at Lane, and then does it again. Thousands of times. Every league meet, every selection decision, every state race. The numbers on the board are simply how often each thing happened.
Line every runner up at their season best and you learn who is fastest on paper. You do not learn how likely anything is — and paper does not run in November.
Nobody races the same twice. Some days the legs are there, some days they are not, and the model treats that as the main event rather than an afterthought.
Two things about those days matter more than people expect. The first is that teammates tend to have them together — the same bus, the same week of training, the same course that suits them or does not. The second is the shape of the days themselves.
One perfect afternoon in September tells you less than a season of racing does. Where a runner has more than one result, Chute uses the shape of their season rather than the single best line of it.
This is the part we would want to see if we were you.
We rebuilt the 2024 and 2025 seasons exactly as they stood in mid-September — 96,265 race results, nothing after that date, no hindsight — then ran the season forward and compared it to what actually happened at Lane in November. Both genders, four points in each season, 197 team-seasons scored.
The bottom pair is the one that matters. It is easy to sort teams; it is hard to be right about how sure you are. Those two say that when this model claims near-certainty it is nearly always correct, and when it writes a team off it is usually correct too.
Picking the winner is a different and much harder question, and we will not dress it up: replaying those four championship races from mid-September, the model's favourite went on to win two of them. Eight weeks out, the eventual champion is genuinely uncertain — and a model that told you otherwise would be lying to you.
Where it is still loose is the middle, this early. Teams we put between 70% and 90% in mid-September made it about half the time — honest odds there are wider than they look, and they tighten every week as results come in. We know that because we measured it, which is the whole point. A model nobody has scored is just an opinion with decimals.
One caveat we would rather state than bury. Those checks were run on 6A, which is where the history is. Every classification runs on the same engine and the same data, and each one's qualifying rules — how many teams a league sends, how many at-large places exist, and for 3A and 2A/1A girls the fact that only four runners score — are taken straight from OSAA. But 2026 is the first year of four-runner scoring, and last season the small-school girls raced as one combined division, so there is no like-for-like history to score those two boards against yet. Treat 6A's numbers as evidence about the method, not as a measured accuracy for every board.
Drop five seconds off a runner and the board moves. The trick is making sure it moved because of the five seconds and not because the dice rolled differently — so every comparison is run against identical races. What you see is the improvement, with the luck held still.
Courses. A flat park and a hilly farm are treated alike, and that is the biggest gap. We have measured it and we are working on it.
The future. Nobody gets injured, nobody has a breakthrough, and the top seven in September are the top seven in November. Real seasons are less tidy, so the further out you are, the more the true spread exceeds what any model shows. Read early-season numbers as a shape, not a promise — they sharpen as the season fills in.
The committee. At-large selection here is a sensible stand-in for a room full of people who also weigh league strength and head-to-head.
None of that is hidden in the small print, because the honest version is more useful than the confident one.
Take time off one or more of your seven and see what it does to your league and state odds. Both worlds are run against the same simulated races, so the difference is the improvement, not luck.