All projects
Research · Causal Inference
Does the Undercut Pay Off?
Causal analysis of Formula 1 pit-stop strategy — does undercutting an opponent actually improve race outcome?
Problem
In Formula 1, teams frequently "undercut" opponents by pitting earlier — but whether this strategy causally improves race position, rather than merely correlating with it, is a harder question.
Approach
Combines lap-level race data with a causal inference design to separate the effect of the undercut from confounding factors like car pace, tyre degradation, and track position, alongside statistical and machine learning analysis of the outcome.
Pipeline
- 01Research QuestionDoes the undercut work?
- 02DataLap-level race data
- 03Causal DesignConfounder-aware setup
- 04Statistical AnalysisEffect estimation
- 05Machine LearningSupporting models
- 06ResultsEvidence-based conclusion
Technologies
- Python
- Causal Inference
- Bayesian Analysis
- Machine Learning
- SQL
Notes
- Full research project — from question framing to causal design to results — documented on the Research page.