Sitemap

Reading the Ruck: How Game Theory and Simulation Reveal Strategic Layers in Rugby – from Uni to Six Nations

3 min readMar 27, 2025

By Jefferies

“Rugby is chess played at sprint speed.”

That’s how I’ve come to see it. Underneath the bruising tackles and high-speed offloads lies a silent war of decision-making. But what changes when you move from a university pitch to the Stade de France? How do decisions evolve between MLR, Pro D2, and the Six Nations?

To answer that, I built a few simple simulations – each based on game-theoretic models tailored to the rugby landscape at different levels.

Level 1: University Rugby – The Predictable Player Game

At the university level, decision-making is often reactive rather than strategic. Most teams play a set pattern with little variation – perfect for modeling with a simple static game.

Model:

• 2 strategies: Run or Kick

• Players assume known tendencies (e.g., 70% run, 30% kick)

• Payoffs based on yardage gained and turnover risk

Insights:

• Teams overuse default strategies (like pick-and-go in the red zone)

• Defenders can exploit predictability with hard line speed

• The best response? Introduce low-cost variation – e.g., occasional chip kicks or pull-back passes

Level 2: MLR – The Adaptive Game

In Major League Rugby, players are fitter, and analysis has crept in – but still limited budgets and fewer elite analysts.

Model: Repeated Game with Limited Memory

• Teams adjust strategy based on last 2 – 3 phases or previous matchups

• Payoffs include fatigue, weather, and lineup rotation

Simulation Output:

• Defensive systems adapt well to patterned attack pods

• Offenses that change point of attack every 3 phases create better line breaks

• Winning teams maximize probabilistic unpredictability within their own shape

Level 3: Pro D2 / Top 14 – The Mixed-Strategy Grinder

Here, strategy deepens. Coaches optimize possession, kicking territory, and scrum penalties. It becomes a mixed-strategy game with fatigue, game clock, and psychology baked in.

Model: Dynamic Bayesian Game

• Uncertainty over opponent’s intent (attack/kick, maul/drift)

• Updates beliefs each phase based on observed choices

Findings:

• Teams with strong kicking 10s can manipulate backfield cover, opening mid-field runs

• Defensive edges shift based on early-game signals (e.g., first 5 penalties drawn)

• A Nash equilibrium is often reached: balanced kick-run ratios, rotating forward pods, targeted breakdown pressure

Level 4: Six Nations – The Stochastic Battle for Margins

At the Six Nations level, the game becomes stochastic, with micro-decisions determining macro-outcomes. One missed defensive read at phase 16 becomes a try. Welcome to the land of Quantal Response Equilibria and multi-agent simulations.

Model: Stochastic Extensive-Form Game + Agent-Based Simulation

• Each “agent” (player) makes decisions under pressure, bounded rationality, and fatigue decay

• Simulates over hundreds of match scenarios

Observations:

• Teams like Ireland use structured chaos: multiple runners at the line with subtle reads

• France’s unpredictability increases pressure on defenders – forcing them into probabilistic errors

• England’s recent inconsistency aligns with unstable equilibrium – teams unable to settle on a dominant pattern under dynamic pressure

Why This Matters: Coaching with Game Theory

Whether you coach a university side or analyze Six Nations tape, game theory offers value:

• Static games teach pattern recognition and exploitation.

• Repeated games model adaptation and memory.

• Bayesian games help understand hidden info (injuries, game plan changes).

• Stochastic games simulate full-match volatility and marginal gain theory.

Next Step: Building Your Own Model

Want to model your own team’s red-zone attack or decision trees for scrum penalties?

Start with:

1. A simple matrix: strategies vs. responses

2. Assign values: based on game footage or stat tracking

3. Simulate iterations: change strategy mixes and track outcomes

4. Add complexity: fatigue, decision errors, timing constraints

I’ll be releasing a public GitHub model soon to help coaches and analysts build this out.

Conclusion:

Rugby is far more than a contact sport – it’s a mathematical dance of decision and deception. Whether you’re coaching U20s or facing France in Paris, the same logic applies: think strategically, stay unpredictable, and play the long game.

Jefferies Jiang
Jefferies Jiang

Written by Jefferies Jiang

I make articles on AI and leadership.