Skip to content

Essential storage, first-party measurement, cookieless Google Tag Manager. Sentry, visit log off unless you Accept all.

Settings
$1 first week

Edge Lab · choose a source-backed question

Run a saved question, or type free text naming home dogs, division games, primetime, wind, rest-day edge, week, public tickets, line move, spread, or total. If the graded history can't score what you asked, you get "Couldn't match this question", never a different question's number.

Every answer is a count of what happened against the closing line, not a priced bet. A ?query= link reopens the same saved answer and is marked stale once it is more than a day old. Selecting or running a question does not award XP.

Edge · Lab · Bench

Calibration

See whether each model's stated chances line up with what actually happened over the last 30 days of graded pregame calls.

Brier and log loss both reward accurate chances and punish confident misses. Lower is better. The 50/50 check asks whether a model scored better or worse than calling every game even.

Calibration scores are unavailable right now

The graded history could not be read. No replacement scores are shown.

Quests · claims and evaluations

No data

No claim quest progress yet

Sign in to track claim/evaluation quests. Missing progress is never converted to XP or a fabricated count.

  • NO RUN No run has produced a measurement yet, so no guardrail can speak.

Coach

Ask a question → the Lab counts how it did over every graded game in the window. Read the caption → games counted, seasons covered, and the as-of date sit under every number. Compare on the Bench → calibration and leaderboard count their own games, so their numbers are not this one.