Sports Economics & Analytics

THE birth certificate of sports economics. Introduced the Invariance Principle (predating Coase's Theorem) showing talent allocation would be identical under free agency or the reserve clause. Also articulated the Uncertainty of Outcome Hypothesis: competitive balance matters for fan interest.

Pioneered empirical estimation of player marginal revenue product (MRP). Using two-stage regression (performance→wins→revenue), demonstrated players earned just 10-20% of their MRP under the reserve clause—empirical ammunition for free agency litigation.

Synthesized 40 years of sports economics research into the definitive framework. Striking finding: revenue sharing generally has no impact on competitive balance (reinforcing Rottenberg's invariance), though salary caps do affect balance.

Demonstrated NFL draft picks are systematically overvalued. Top picks generate less surplus value than late first-rounders, teams exhibit overconfidence and the winner's curse. Implied discount rate in multi-year trades reached 136%. Influenced teams like New England to trade down systematically.

Analyzed 1,417 penalty kicks from top European leagues. Kickers and goalkeepers play exactly according to mixed-strategy Nash equilibrium: success probabilities are statistically identical across strategies, and choices show true serial independence. First field support for both implications of von Neumann's Minimax Theorem.

Confirmed equilibrium behavior in tennis serves using championship matches featuring McEnroe, Borg, Becker, and Sampras. The same statistical tests that soundly reject equilibrium play in experimental data confirm equilibrium behavior among elite professionals.

Extended minimax analysis to nearly 500,000 serves across 3,000 matches. Found higher-ranked players conform more closely to equilibrium theory, and substantial differences in degree to which men and women conform to equilibrium predictions.

Found minimax violations in baseball pitch selection and football play-calling. Pitchers throw too many fastballs; football teams pass less than they should. Suggests the strategic simplicity of penalty kicks and tennis serves may be key to equilibrium play.

Established betting markets as legitimate venues for testing rationality. Despite quick feedback that should facilitate learning, the favorite-longshot bias persists—bettors systematically overweight longshots.

Used 5.6 million horse race starts and exotic bet pricing to discriminate between competing theories. Evidence favors probability misperceptions (prospect theory) over risk-love: longshots return about -61% while favorites return approximately -5%.

Explains why bookmakers set prices rather than equilibrating supply and demand. Bookmakers are more skilled at predicting outcomes than bettors and systematically exploit bettor biases by choosing prices that deviate from market clearing.

First mathematically rigorous league model. Key finding: equalization of playing strengths is inconsistent with profit maximization under standard rules—challenged assumptions about natural competitive balance.

Provides the theoretical foundation for understanding prize structures in sports. Key predictions: effort increases with the spread between winning and losing prizes, and only the difference matters, not absolute prize sizes. Introduced competitive handicapping for efficient competition among heterogeneous abilities.

Comprehensive review covering optimal prize structures, incentive-balance tradeoffs, and revenue sharing effects. Draws together research on individualistic sports (golf, footraces) with team sports (baseball, soccer) through the lens of contest theory.

Used dynamic programming to show NFL teams punt and kick field goals far too often. Going for a touchdown from the opponent's 2-yard line instead of taking the field goal increases win probability by approximately 3%. Originally titled 'It's Fourth Down and What Does the Bellman Equation Say?'

Introduced Expected Possession Value (EPV) using player tracking data—the foundation for real-time decision valuation in basketball. Models possessions at multiple resolution levels, differentiating continuous player movements from discrete events like shots and turnovers.