Rugby Analytics with R: Complete Guide to Performance Analytics in Rugby Union and League | R bloggers

Rugby is a sport characterized by clashes, structure and constant tactical adaptation. Unlike many other invasion sports, rugby alternates between highly structured moments – scrums, lineouts, restarts – and extended passages of chaotic open play. Each phase generates rich performance data: tackles, rucks, carries, kicks, meters gained, penalties awarded, turnovers and spatial changes in territory. Despite this wealth, rugby analytics has historically lagged behind other sports, especially when it comes to open, reproducible analytical workflows.

This gap offers clear opportunities. R provides a complete environment for rugby performance analysis: data collection, cleaning, modeling, visualization and automated reporting. For analysts, sports scientists and coaches, R enables evidence-based decision making that goes far beyond traditional metrics and subjective video assessment.

Why rugby analytics requires a different analytical mindset

Rugby is not a possession sport in the same way as basketball, nor is it a continuous flow game like football. Possession can be short or long, territory is often more important than time on the ball, and a single penalty can turn the match momentum. Analytics must therefore respect the unique structure of rugby.

Simple totals – tackles, carries, meters – are insufficient on their own. Analysts must take into account the state of the game, the position on the field, the quality of the opposition and the role of the player. R makes it possible to integrate this context systematically and consistently across matches and seasons.

Data collection in rugby: scraping, APIs and internal feeds

Public rugby data is fragmented and inconsistent. Analysts often combine multiple sources to build a useful data set. R is particularly well suited to this challenge because it supports web scraping, API usage, and database integration within a single workflow.

# Core libraries for rugby data acquisition
library(tidyverse)
library(rvest)
library(httr)
library(jsonlite)

# Example: pulling match data from an API
response <- GET("https://api.example.com/rugby/match/9876")
raw_json <- content(response, "text")
match_data <- fromJSON(raw_json)

Web scraping is often necessary when APIs are unavailable. This requires careful handling of HTML structure, rate limits, and data validation to ensure accuracy and reproducibility.

# Scraping a match statistics table
page <- read_html("https://example-rugby-site.com/match/9876")

team_stats <- page %>%
  html_node("table.match-stats") %>%
  html_table()

team_stats

Data cleaning and validation: a crucial but underestimated step

Rugby datasets are rarely ready for analysis. Player substitutions, injury substitutions, and inconsistencies in data entry lead to errors that can skew results if left unchecked.

# Standardizing and validating team statistics
team_stats_clean <- team_stats %>%
  janitor::clean_names() %>%
  mutate(across(where(is.character), str_trim)) %>%
  mutate(
    possession = as.numeric(possession),
    territory = as.numeric(territory)
  )

# Basic validation check
stopifnot(all(team_stats_clean$possession 

Validation logic must be embedded directly into the pipeline. This ensures that each new match is processed consistently, reducing human error and analyst workload.

Transforming events into rugby-specific units of analysis

Raw events are just the starting point. Meaningful rugby analysis requires transforming events into units such as phases, possessions, sets and passages of play.

# Creating phase identifiers from ruck events
events <- events %>%
  arrange(match_id, event_time) %>%
  mutate(
    phase_id = cumsum(event_type == "ruck")
  )

# Summarising phase-level performance
phase_summary <- events %>%
  group_by(match_id, team, phase_id) %>%
  summarise(
    duration = max(event_time) - min(event_time),
    carries = sum(event_type == "carry"),
    meters = sum(meters_gained, na.rm = TRUE),
    turnovers = sum(event_type == "turnover"),
    .groups = "drop"
  )

These structures allow analysts to study momentum, ruck efficiency and attacking intent in a way that aligns with how coaches understand the game.

Advanced player performance analysis with R

Player evaluation in rugby must be contextual and role specific. Front row players, halves and outside defenders contribute in fundamentally different ways.

# Player-level performance profile
player_profile <- events %>%
  group_by(player_id, player_name, position) %>%
  summarise(
    minutes_played = max(event_time) / 60,
    tackles = sum(event_type == "tackle"),
    missed_tackles = sum(event_type == "missed_tackle"),
    carries = sum(event_type == "carry"),
    meters = sum(meters_gained, na.rm = TRUE),
    offloads = sum(event_type == "offload"),
    penalties_conceded = sum(event_type == "penalty_conceded"),
    .groups = "drop"
  ) %>%
  mutate(
    tackles_per_min = tackles / minutes_played,
    meters_per_carry = meters / carries
  )

Rate-based metrics show impact more effectively than totals, especially when comparing starters to starters or evaluating performance in different match contexts.

Defensive systems analysis: beyond individual tackles

Effective defense is systemic. Missed tackles are often the result of spacing errors, fatigue or poor decision-making rather than individual incompetence.

# Defensive performance by field channel
defense_analysis <- events %>%
  filter(event_type %in% c("tackle", "missed_tackle")) %>%
  group_by(team, field_channel) %>%
  summarise(
    tackles = sum(event_type == "tackle"),
    misses = sum(event_type == "missed_tackle"),
    success_rate = tackles / (tackles + misses),
    .groups = "drop"
  )

Defensive analytics should reveal structural weaknesses and workload imbalances, not just the number of individual errors.

Territory, trapping strategy and spatial dominance

Territory remains a key determinant of success in rugby. Teams that consistently win the territorial battle reduce defensive workload and increase scoring opportunities.

# Kicking distance and efficiency
kicks <- events %>%
  filter(event_type == "kick") %>%
  mutate(kick_distance = end_x - start_x)

kicking_summary <- kicks %>%
  group_by(team, kick_type) %>%
  summarise(
    avg_distance = mean(kick_distance, na.rm = TRUE),
    kicks = n(),
    .groups = "drop"
  )

Spatial analysis allows analysts to quantify whether a team’s kicking strategy fits the stated game model and the constraints of the environment.

Probability and decision modeling in rugby

Win probability models convert complex match states into intuitive probabilities. In rugby, these models must take into account scoring, time, territory, possession and discipline risks.

# Building a basic win probability model
wp_data <- matches %>%
  mutate(
    score_diff = team_score - opponent_score,
    time_remaining = 80 - minute
  )

wp_model <- glm(
  win ~ score_diff + time_remaining + territory,
  data = wp_data,
  family = binomial()
)

summary(wp_model)

Even simple models provide immediate value by framing tactical decisions – such as kicking for touch versus taking the points – in probabilistic terms.

Automated reporting and reproducible workflows

The final step in rugby analysis is communication. With R, analysts can automate reporting, ensuring consistency and freeing up time to generate deeper insights.

# Creating a clean match summary table
summary_table <- team_stats_clean %>%
  select(team, possession, territory, tackles, line_breaks, penalties_conceded)

knitr::kable(summary_table)

Automated reports ensure that analysis becomes part of the weekly rhythm and not as an optional extra.

The strategic opportunity in rugby analysis with R

There is a clear and growing interest in rugby analytics, but very little comprehensive R-focused content. Analysts, sports scientists and coaches are actively looking for practical guidance.

A dedicated, end-to-end approach – encompassing data acquisition, performance metrics, modeling and reporting – fills a real gap and establishes authority in a niche with minimal competition.