International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 13 Issue: 01 | Jan 2026
p-ISSN: 2395-0072
www.irjet.net
A Ranking-Based Prediction Model for Formula 1 Race Outcomes: An Alternative to Machine Learning Approaches Aadit Tuli 1 Student, The Shri Ram School Moulsari, V-37, Moulsari Ave, DLF Phase 3,
Sector 24, Gurugram, Haryana, 122002, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Formula 1 race predictions have traditionally
most important for predicting race outcomes. Additionally, large-scale ML systems raise concerns about energy consumption, carbon emissions, data privacy, and resource usage that are increasingly important in sustainable research practices. This lack of transparency and environmental impact makes it difficult for teams, analysts, and fans to understand what truly drives racing performance while also contributing to broader sustainability challenges in computational research.
relied on complex machine learning algorithms that are difficult to interpret and understand. This study presents a novel mathematical formula-based approach for predicting race outcomes using historical performance data and team capabilities, making the prediction process completely transparent and easier for novices to perform. The objective was to develop and validate a simple yet effective mathematical model that predicts Formula 1 race outcomes while providing complete insight into how predictions are made. Historical data from multiple Formula 1 seasons were analyzed, incorporating driver rankings (DR) and team rankings (TR) normalized across regulation eras. The model was validated using three 2024 Grand Prix races: Belgian, Australian, and Italian. Driver rankings incorporated historical performance metrics weighted by tire degradation across different regulatory periods, while team rankings considered car performance, reliability, and technical specifications. The mathematical model demonstrated competitive predictive accuracy across test races, achieving 52% exact position accuracy and a 73% top-5 driver identification rate. The proposed mathematical formula offers a transparent and computationally efficient alternative to machine learning approaches for Formula 1 race prediction, providing a clear understanding of how driver skill, team performance, and regulatory factors influence race outcomes.
One significant gap in existing research is the lack of a simple, transparent formula that effectively balances driver skill, team performance, and regulatory effects. 4 Most models either focus heavily on machine learning predictions without explaining their methodology or rely on subjective assessments of driver capabilities. This research aims to answer several key questions: Can a simple mathematical formula predict Formula 1 race outcomes as accurately as complex machine learning models? What factors are most important in determining race results - driver skill, team performance, or car technology? How can we account for the major changes in Formula 1 regulations over the years when making predictions? The primary objective is to develop a transparent prediction model allowing anyone to see exactly how predictions are calculated.
2. METHODS
Key Words: Mathematics, Probability and Statistics, Formula 1, Race Prediction, Mathematical Modelling
2.1 Data Collection
1.INTRODUCTION
Data were sourced from multiple authoritative sources, including the Official Formula 1 website (formula1.com), 5 which provided current race results, driver standings, team performance data, and official race classifications for the 2024 season. The StatsF1 database (statsf1.com)6 provided comprehensive historical Formula 1 statistics, including lap times, qualifying positions, finishing positions, and mechanical reliability data across multiple seasons. Additional data were obtained from official FIA statistics and Formula 1 archives, which covered driver performances, team results, and regulatory changes. The primary data sources provided comprehensive coverage of Formula 1 statistics from 2006, encompassing multiple regulatory periods.
Predicting race outcomes in motorsports, particularly Formula 1, has been an ongoing challenge for analysts and researchers. Traditional approaches have included statistical models, Elo rating systems,1 Monte Carlo simulations, and advanced machine learning algorithms trained on historical race data. While these methods can achieve reasonable accuracy, they often function as "black boxes" - you can see the prediction, but you can't understand how or why the prediction was made. Machine learning models such as neural networks have been used to assess the impact of race conditions, driver performance, and car capabilities.3 These approaches can be powerful, but they require substantial computational resources and provide little insight into which factors are
© 2026, IRJET
|
Impact Factor value: 8.315
|
ISO 9001:2008 Certified Journal
|
Page 531