We run several well-known football models and show them all, so you can see where they agree. None is a crystal ball; this is analysis for interest, not betting advice.
Poisson - counts how often each team scores to estimate likely goals, like weighted dice.
Dixon-Coles - Poisson, tuned so low-scoring results (0-0, 1-1) and recent form fit better.
xG Simulation - uses chance quality (expected goals) and replays the match thousands of times.
ELO - a chess-style strength rating, updated after every match.
Neural Net - learns patterns from past matches to spot what tends to win.
Bradley-Terry - ranks teams by head-to-head strength and turns that into a result.
Market-Implied - reads the bookmakers' odds and converts them back into a likely score.
Bayesian - starts with a prior belief and updates it as new results arrive.
Form-Weighted - leans on recent games more than older ones (hot and cold streaks).
H2H History - looks at how these two teams have done against each other before.
Ensemble Avg - the straight average of all the models above.
Meta-Learner - learns which models to trust most and blends them into one call.