Part II: Player Analytics

"Stats are accumulated by players. Wins are accumulated by teams." — Bill Parcells

Overview

Part II applies the foundations from Part I to the core challenge of football analytics: measuring individual player value. In a sport where 11 players must coordinate on every snap, isolating individual contribution is both essential and difficult.

The Challenge of Player Evaluation

Football presents unique challenges for player evaluation:

  • Interdependence: A quarterback's success depends on the offensive line, receivers, and play-calling
  • Sample Size: Even the busiest players have limited data compared to baseball
  • Context: Game situations, opponents, and schemes vary constantly
  • Hidden Information: Much of what determines success isn't captured in box scores

What You'll Learn

This section covers evaluation frameworks for each major position:

Chapter 6: Quarterback Evaluation The most scrutinized position in sports. We'll explore EPA-based metrics, CPOE, air yards, and the challenges of separating QB skill from supporting cast.

Chapter 7: Rushing Analytics Why traditional rushing stats mislead, how to contextualize running back value, and the economics of the position.

Chapter 8: Receiving Analytics Target quality, separation, and yards after catch. Measuring what receivers contribute beyond volume stats.

Chapter 9: Offensive Line Analytics The hardest position group to evaluate. Pass protection metrics, run blocking grades, and individual attribution.

Chapter 10: Defensive Player Evaluation Pressures, coverage grades, and tackling efficiency. Why defensive stats require careful interpretation.

Chapter 11: Special Teams Analytics Kicking, punting, and return game analysis. Often overlooked but analytically rich.

Common Themes

Several themes recur across position evaluations:

1. Efficiency vs. Volume

High-volume players accumulate impressive counting stats but may be less efficient than lower-volume alternatives. We consistently ask: "Is this player good, or just heavily used?"

2. Context Adjustment

Raw performance must be adjusted for: - Opponent strength - Game situation (score, time) - Supporting cast quality - Scheme effects

3. Stability and Sample Size

Some metrics stabilize quickly (completion percentage) while others require years of data (rushing success rate). We'll discuss reliability timelines for each metric.

4. The Replacement Level Concept

Player value should be measured against what a freely available replacement would provide. This concept, borrowed from baseball's WAR, helps contextualize production.

Key Metrics Introduced

Metric Position Meaning
EPA/Play All Expected points contribution per play
CPOE QB Completion % over expected
ADOT QB/WR Average depth of target
YAC WR/TE Yards after catch
RYOE RB Rush yards over expected
Pressure Rate OL/DL % of dropbacks with pressure
Coverage Grade DB Subjective/objective coverage rating

Tools and Data

The analyses in Part II use:

  • nfl_data_py: Play-by-play and player statistics
  • nflfastR models: EPA and WP calculations
  • Next Gen Stats: When available for tracking data
  • PFF grades: For subjective evaluations (with caveats)

How to Approach This Section

For each position chapter:

  1. Understand the metrics before critiquing them
  2. Build the analysis with provided code
  3. Interpret with caution — context matters
  4. Compare to alternatives — what else could we measure?

Player evaluation is never "solved." The best analysts continuously refine their approaches as new data and methods become available.

Let's begin with the position that generates the most discussion and the most data: quarterback.

Chapters in This Part