I modelled how much the Azteca actually changes England–Mexico
I wanted to move beyond the general claim that the Azteca is “a difficult place to play” and estimate how much the setting should actually change the probability of England or Mexico progressing.
I fitted a simple independent Poisson model to the opening 90-minute market, then constructed a neutral-venue counterfactual by removing an assumed combined adjustment for:
- altitude and acclimatisation
- crowd pressure
- familiarity with the pitch, ball and venue
- travel and recovery
- weather
My central scenario estimates:
Neutral venue
England 68%
Mexico 32%
At the Azteca
England 58%
Mexico 42%
So the combined venue effect is worth roughly ten percentage points to Mexico.
The interesting part is that current best prices imply closer to 45.5% for Mexico, suggesting that the market may now be pricing an even larger Azteca effect than the model does.
The assumptions, sensitivity scenarios and full workbook are all public: https://substack.com/home/post/p-204909017
I’d be interested in criticism of the neutral counterfactual and the +0.35 expected-goal-difference central assumption.