Ryder Cup Singles Ordering

In the Ryder Cup golf competition, team captains must decide the ordering of their 12 players for the final day's singles matches; lineups are submitted without seeing the opponent's order. This ordering determines which players face which opponents and can significantly affect the probability of achieving the target score needed to win or retain the cup. With 12 players per side there are approximately 479 million possible orderings, and the goal is to maximize the probability of earning at least k points (where k depends on the score after two days), not merely to maximize the expected points. We model match win, loss, and tie (WLT) probabilities using strokes gained as the skill measure, estimating both a logistic and a linear specification from hole-level data. We maximize the right tail of the point distribution over all orderings using an efficient forward induction algorithm that computes exact probabilities. For the logistic model we prove optimality of strong-vs-strong and strong-vs-weak orderings in the extreme target cases; for the linear model we derive closed-form variance expressions and prove structural results for optimal orderings-matching by skill rank (strong-vs-strong) maximizes variance and is optimal when a team needs many points (e.g., coming from behind), while the reverse order (strong-vs-weak) minimizes variance and is optimal when protecting a lead. Prior work relied on heuristics and world-ranking-based skill; we show that exact optimization is now feasible for the full Ryder Cup problem and that strokes-gained-based WLT models yield clear, interpretable rules. Captains can apply these rules to choose orderings that improve the probability of winning or retaining the cup compared with ad hoc or historical practice.