How we predict padel
Our padel predictions work the same way as our table-tennis ones — an Elo rating, calibrated against real results. Here's the padel-specific part.
We rate pairs, not players
Because pro padel is doubles, we treat each pairing as the entity that earns a rating. When Galán/Chingotto beat another pair, their pair rating rises and the opponent's falls — by more if it was an upset, less if it was expected.
Elite padel pairs tend to stay together for a season, so pair ratings settle into something meaningful surprisingly fast.
From rating to probability
The gap between two pairs' ratings becomes a win probability, temperature-calibrated so that a stated 70% really happens about 70% of the time. No market odds feed in — it's pure model output.
New and honest about it
Padel is our newest sport, so the model is still building depth from the available history. Early predictions are less settled than our established leagues — and the track-record page is upfront about that. Accuracy sharpens as more matches accumulate.