Under the hood
The Seer is a statistical model — no pundits, no hype, no star-name bias. Here's exactly what goes into every call, in plain English.
Elo is the same system used to rank chess players. Each national team has a single number — its strength. Beat a stronger team and you gain points; lose to a weaker one and you drop. Ratings are seeded from the current World Football Elo values and then updated by replaying real results.
For a fixture, the Seer compares the two ratings. A bigger gap means a more lopsided match. That difference is converted — via the Elo expected-score curve — into win, draw and loss probabilities that add up to 100%. A small home advantage is added to the nominal home side (it only really matters for the host nations).
Goals in football follow a Poisson distribution. The Seer turns each team's strength into an expected number of goals, then finds the single most likely exact scoreline — that's the "Seer predicts 2–0" you see on a card.
After full time, the Seer's pick is checked against the actual result — right alongside the crowd's vote and your own. Nothing is hidden: the accuracy you see on the scoreboard is the model keeping score on itself.
To keep it honest: the model is built purely on results, ratings, recent form, head-to-head history and venue. It does not read formations, squad ages, coaching changes or injury news — those aren't reliably available as data, and we won't pretend the maths accounts for them. That's also exactly why the Seer can be wrong — and why beating it feels good.