Contribute
Hometown Line is a small project with a few specific gaps. These are the ones where help would actually change something.
Local knowledge of a state
The roster is hand-curated and it shows. Illinois shipped without New Trier or either Glenbrook until somebody who knows the state said so. If you know your state's programmes, that is the single most valuable thing you can give this.
School colours
Most programmes have colours read from a source. The rest carry a plausible colour from a palette, which is honest but is not theirs. Corrections are easy to apply and immediately visible.
Coaches and athletic directors
The sites that hold every score in the country did not get them from an association — coaches typed them in. That is the whole model, and it means the data was never behind privileged access. If you coach or run athletics at one of these schools, the Team Desk lets you manage your own schedule with a staff key: add missing games, fix kickoffs and venues, post finals. No key yet? Ask for one below; it takes about a day.
Open the Team DeskOr send your season in one formAt a game
A result is shown here only where its source is named, and the stands are one of them, beside the state associations the site may read, a credited publication, and coaches. Put in what the scoreboard says and it shows for everybody once other people at the ground agree, credited to the stands with a count of people.
Report a scoreA result you saw
Results come from the state associations and leagues that let the site read them, from Cal-Hi Sports, from people at the ground, and from coaches — each credited by name. A game with none of these behind it is not shown as a result. Writing down a game you watched afterwards is lawful where a crawler is not, so there is a form for it; it asks for a link, because a claim nobody can check is worth less than none.
Send in a resultCalibration
The generated fixtures are shaped to look like real prep football — margins, home advantage, common final margins — and still miss on shutouts (0.12 against a measured 0.19). Every remaining lever trades that against how spread out the margins are, which sets the predicted margin. If you can model that better, say so.
How to reach the project
The repository is private for now, so there is no link to give you. Everything above can be sent as a message instead — a list of corrections is as useful as a pull request and takes you less time.
Get in touch