

From the engine's power ratings, adjusted for home field and rest, then mapped through the margin distribution. Not a blend of the market — this is our own number.
We disagree with the market here. Worth knowing: across 1,139 backtested games our disagreements were right 41% of the time. We publish them regardless — and we publish how they land.
Venue and rest come from the schedule feed and feed directly into the model; surface and weather are context we show but don't yet model.
Our rating has Indianapolis Colts as the stronger side by 8 rating points, which on its own is worth about 0.3 points on the scoreboard (IND 1,459, WAS 1,451).
This is a neutral-site game, so no home-field adjustment is applied at all.
That lands on a projected margin of -0.3, and running that through the historical spread of actual results against projections gives Indianapolis Colts at 51%.
The market disagrees. The closing line makes Washington Commanders the favourite by 1.5. Our engine sees only results — it has no view on injuries, roster changes or who a team has signed. When we split from the line, history says we're wrong more often than right, and we show you that rather than dressing the disagreement up as an edge.
Other games already through the engine.
Every sport runs on the same engine, and every pick is published before the event and graded after.