Form summarises recent results — WWDLW strips on every stats site. Raw form is a weak predictor: five results carry enormous noise, opposition quality varies wildly, and one deflected goal flips a letter. Underlying rates (shots, xG, field position) predict the next match better than the letters do.
Form earns a place when it is weighted (recent matches more), adjusted for who was played, and cross-checked against the underlying numbers. Our model applies time-decay for exactly this reason — a result from last week says more than one from October, but neither overrides a season of rates.
Team A: LLWLL against the league's top five. Team B: WWWWW against the bottom five. The letters say B is flying; opposition-adjusted rates often say A is the stronger side at level terms — and the market price usually sits somewhere between the two stories.