WNBA Team Performance Analytics

Team Performance Evaluation

WNBA team analytics follows similar principles to NBA analysis but accounts for differences in pace, physicality, and strategic emphasis. The Four Factors framework—shooting, turnovers, rebounding, and free throws—remains highly predictive of team success.

Python: WNBA Team Analytics

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns

# Sample WNBA team data (normally loaded from API/database)
team_stats = pd.DataFrame({
    'team': ['Las Vegas Aces', 'New York Liberty', 'Connecticut Sun',
             'Minnesota Lynx', 'Seattle Storm', 'Phoenix Mercury',
             'Dallas Wings', 'Atlanta Dream', 'Chicago Sky',
             'Indiana Fever', 'Washington Mystics', 'Los Angeles Sparks'],
    'wins': [34, 32, 27, 25, 20, 19, 18, 15, 13, 10, 9, 8],
    'losses': [6, 8, 13, 15, 20, 21, 22, 25, 27, 30, 31, 32],
    'points_for': [2840, 2795, 2650, 2588, 2520, 2490, 2455, 2380, 2350, 2280, 2260, 2200],
    'points_against': [2520, 2550, 2590, 2600, 2680, 2720, 2750, 2820, 2890, 2950, 2980, 3020],
    'fgm': [998, 985, 940, 920, 895, 880, 870, 845, 835, 810, 800, 785],
    'fga': [2180, 2160, 2140, 2120, 2100, 2090, 2080, 2070, 2060, 2050, 2040, 2030],
    'fg3m': [285, 295, 268, 255, 248, 242, 238, 230, 225, 218, 215, 208],
    'fg3a': [785, 810, 750, 720, 710, 705, 700, 695, 690, 685, 680, 675],
    'ftm': [560, 545, 505, 495, 485, 490, 480, 465, 460, 445, 450, 425],
    'fta': [720, 710, 680, 670, 665, 675, 665, 655, 650, 645, 660, 625],
    'orb': [385, 370, 395, 380, 375, 365, 360, 355, 350, 345, 340, 335],
    'drb': [1185, 1165, 1155, 1145, 1130, 1120, 1110, 1100, 1085, 1070, 1060, 1050],
    'ast': [680, 695, 645, 630, 615, 605, 595, 580, 570, 555, 550, 540],
    'tov': [495, 505, 485, 490, 510, 515, 520, 535, 545, 560, 570, 580],
    'stl': [285, 295, 280, 275, 268, 265, 260, 255, 250, 245, 240, 235],
    'blk': [165, 155, 175, 168, 160, 158, 155, 150, 145, 140, 138, 135],
    'opp_fgm': [890, 900, 915, 920, 945, 960, 970, 990, 1010, 1035, 1045, 1060],
    'opp_fga': [2050, 2060, 2070, 2080, 2090, 2100, 2110, 2120, 2130, 2140, 2150, 2160]
})

# =============================================================================
# 1. Calculate Advanced Team Metrics
# =============================================================================

def calculate_team_efficiency(team_df):
    """Calculate offensive and defensive efficiency metrics"""

    # Estimate possessions (simplified formula)
    # Poss = FGA + 0.44*FTA - ORB + TOV
    team_df['possessions'] = (
        team_df['fga'] + 0.44 * team_df['fta'] -
        team_df['orb'] + team_df['tov']
    )

    # Offensive Rating (points per 100 possessions)
    team_df['ortg'] = (team_df['points_for'] / team_df['possessions']) * 100

    # Defensive Rating (points allowed per 100 possessions)
    team_df['drtg'] = (team_df['points_against'] / team_df['possessions']) * 100

    # Net Rating
    team_df['net_rtg'] = team_df['ortg'] - team_df['drtg']

    # Pace (possessions per 40 minutes, estimated)
    games = team_df['wins'] + team_df['losses']
    team_df['pace'] = (team_df['possessions'] / games) / 40 * 40

    # Win percentage
    team_df['win_pct'] = team_df['wins'] / games

    return team_df

team_stats = calculate_team_efficiency(team_stats)

print("=== Team Efficiency Ratings ===")
print(team_stats[['team', 'wins', 'losses', 'ortg', 'drtg', 'net_rtg']]
      .sort_values('net_rtg', ascending=False))

# =============================================================================
# 2. Four Factors Analysis
# =============================================================================

def calculate_four_factors(team_df):
    """Calculate Dean Oliver's Four Factors for team success"""

    # Factor 1: Effective Field Goal % (eFG%)
    team_df['efg_pct'] = (team_df['fgm'] + 0.5 * team_df['fg3m']) / team_df['fga']
    team_df['opp_efg_pct'] = (team_df['opp_fgm'] + 0.5 * (team_df['points_against'] -
                               team_df['opp_fgm'] * 2 - team_df['fgm'])) / team_df['opp_fga']

    # Factor 2: Turnover Rate (TOV%)
    team_df['tov_pct'] = team_df['tov'] / team_df['possessions']

    # Factor 3: Offensive Rebounding Rate (ORB%)
    total_rebounds = team_df['orb'] + team_df['drb']
    team_df['orb_pct'] = team_df['orb'] / (team_df['orb'] + team_df['drb'])

    # Factor 4: Free Throw Rate (FT/FGA)
    team_df['ft_rate'] = team_df['ftm'] / team_df['fga']

    return team_df

team_stats = calculate_four_factors(team_stats)

print("\n=== Four Factors Rankings ===")
print("Shooting (eFG%):")
print(team_stats[['team', 'efg_pct']].sort_values('efg_pct', ascending=False).head(5))

print("\nTurnover Rate (lower is better):")
print(team_stats[['team', 'tov_pct']].sort_values('tov_pct').head(5))

print("\nOffensive Rebounding %:")
print(team_stats[['team', 'orb_pct']].sort_values('orb_pct', ascending=False).head(5))

print("\nFree Throw Rate:")
print(team_stats[['team', 'ft_rate']].sort_values('ft_rate', ascending=False).head(5))

# =============================================================================
# 3. Predict Win Percentage from Net Rating
# =============================================================================

from scipy.stats import pearsonr

correlation, p_value = pearsonr(team_stats['net_rtg'], team_stats['win_pct'])

print(f"\n=== Net Rating vs Win Percentage ===")
print(f"Correlation: {correlation:.3f}")
print(f"P-value: {p_value:.4f}")

# Linear regression for win prediction
from sklearn.linear_model import LinearRegression

X = team_stats[['net_rtg']].values
y = team_stats['win_pct'].values

model = LinearRegression()
model.fit(X, y)

team_stats['predicted_win_pct'] = model.predict(X)
team_stats['win_diff'] = team_stats['win_pct'] - team_stats['predicted_win_pct']

print("\nTeams Over/Under-performing Expectations:")
print(team_stats[['team', 'win_pct', 'predicted_win_pct', 'win_diff']]
      .sort_values('win_diff', ascending=False))

# =============================================================================
# 4. Offensive vs Defensive Efficiency Scatter
# =============================================================================

fig, ax = plt.subplots(figsize=(10, 8))

scatter = ax.scatter(team_stats['ortg'], team_stats['drtg'],
                     s=team_stats['win_pct']*1000,
                     alpha=0.6, c=team_stats['net_rtg'],
                     cmap='RdYlGn')

# Add team labels
for idx, row in team_stats.iterrows():
    ax.annotate(row['team'],
                (row['ortg'], row['drtg']),
                fontsize=8, ha='center')

# Add reference lines (league average)
ax.axhline(team_stats['drtg'].mean(), color='gray', linestyle='--', alpha=0.5)
ax.axvline(team_stats['ortg'].mean(), color='gray', linestyle='--', alpha=0.5)

ax.set_xlabel('Offensive Rating (points per 100 poss)', fontsize=12)
ax.set_ylabel('Defensive Rating (points allowed per 100 poss)', fontsize=12)
ax.set_title('WNBA Team Efficiency Map\n(Size = Win%, Color = Net Rating)', fontsize=14)
ax.invert_yaxis()  # Lower DRtg is better
plt.colorbar(scatter, label='Net Rating')
plt.tight_layout()
# plt.show()

# =============================================================================
# 5. Team Style Analysis
# =============================================================================

def categorize_team_style(row):
    """Categorize team playing style"""

    if row['pace'] > team_stats['pace'].median():
        tempo = "Fast"
    else:
        tempo = "Slow"

    if row['fg3a'] / row['fga'] > 0.35:
        shooting = "Three-Heavy"
    else:
        shooting = "Inside-Focused"

    return f"{tempo}-{shooting}"

team_stats['style'] = team_stats.apply(categorize_team_style, axis=1)

print("\n=== Team Style Distribution ===")
print(team_stats[['team', 'pace', 'style', 'net_rtg']]
      .sort_values('net_rtg', ascending=False))

# Aggregate statistics by style
style_performance = team_stats.groupby('style').agg({
    'net_rtg': 'mean',
    'win_pct': 'mean',
    'ortg': 'mean',
    'drtg': 'mean'
}).round(2)

print("\n=== Performance by Playing Style ===")
print(style_performance)

print("\n=== Team Analytics Complete ===")
print("Metrics calculated:")
print("✓ Offensive/Defensive Ratings")
print("✓ Net Rating and Pace")
print("✓ Four Factors (Shooting, TOV, REB, FT)")
print("✓ Win prediction models")
print("✓ Team style classification")

library(wehoop)
library(tidyverse)
library(scales)
library(ggrepel)

# Load WNBA team box scores
team_box <- wehoop::load_wnba_team_box(seasons = 2024)

# =============================================================================
# 1. Calculate Season-Long Team Statistics
# =============================================================================

team_season_stats <- team_box %>%
  group_by(team_id, team_display_name) %>%
  summarise(
    games = n(),
    wins = sum(team_winner, na.rm = TRUE),
    points_for = sum(team_score, na.rm = TRUE),
    points_against = sum(opponent_team_score, na.rm = TRUE),

    # Shooting
    fgm = sum(field_goals_made, na.rm = TRUE),
    fga = sum(field_goals_attempted, na.rm = TRUE),
    fg3m = sum(three_point_field_goals_made, na.rm = TRUE),
    fg3a = sum(three_point_field_goals_attempted, na.rm = TRUE),
    ftm = sum(free_throws_made, na.rm = TRUE),
    fta = sum(free_throws_attempted, na.rm = TRUE),

    # Other stats
    orb = sum(offensive_rebounds, na.rm = TRUE),
    drb = sum(defensive_rebounds, na.rm = TRUE),
    trb = sum(total_rebounds, na.rm = TRUE),
    ast = sum(assists, na.rm = TRUE),
    stl = sum(steals, na.rm = TRUE),
    blk = sum(blocks, na.rm = TRUE),
    tov = sum(turnovers, na.rm = TRUE),

    .groups = "drop"
  ) %>%
  mutate(
    # Win percentage
    win_pct = wins / games,

    # Estimate possessions
    possessions = fga + 0.44 * fta - orb + tov,

    # Efficiency ratings
    ortg = (points_for / possessions) * 100,
    drtg = (points_against / possessions) * 100,
    net_rtg = ortg - drtg,

    # Pace (possessions per game)
    pace = possessions / games,

    # Four Factors
    efg_pct = (fgm + 0.5 * fg3m) / fga,
    tov_pct = tov / possessions,
    orb_pct = orb / (orb + drb),
    ft_rate = ftm / fga,

    # Shooting percentages
    fg_pct = fgm / fga,
    three_pct = fg3m / fg3a,
    ft_pct = ftm / fta,

    # Per game averages
    ppg = points_for / games,
    rpg = trb / games,
    apg = ast / games
  )

cat("=== WNBA Team Efficiency Rankings ===\n")
print(team_season_stats %>%
  select(team_display_name, wins, games, ortg, drtg, net_rtg) %>%
  arrange(desc(net_rtg)))

# =============================================================================
# 2. Four Factors Analysis
# =============================================================================

cat("\n=== Four Factors Leaders ===\n\n")

cat("1. Shooting (eFG%):\n")
print(team_season_stats %>%
  select(team_display_name, efg_pct) %>%
  arrange(desc(efg_pct)) %>%
  head(5))

cat("\n2. Turnover Rate (lower is better):\n")
print(team_season_stats %>%
  select(team_display_name, tov_pct) %>%
  arrange(tov_pct) %>%
  head(5))

cat("\n3. Offensive Rebounding %:\n")
print(team_season_stats %>%
  select(team_display_name, orb_pct) %>%
  arrange(desc(orb_pct)) %>%
  head(5))

cat("\n4. Free Throw Rate:\n")
print(team_season_stats %>%
  select(team_display_name, ft_rate) %>%
  arrange(desc(ft_rate)) %>%
  head(5))

# =============================================================================
# 3. Net Rating vs Win Percentage Correlation
# =============================================================================

correlation <- cor(team_season_stats$net_rtg, team_season_stats$win_pct)
cat(sprintf("\n=== Net Rating vs Win%% Correlation: %.3f ===\n", correlation))

# Predict win percentage from net rating
win_model <- lm(win_pct ~ net_rtg, data = team_season_stats)

team_season_stats <- team_season_stats %>%
  mutate(
    predicted_win_pct = predict(win_model, newdata = .),
    win_diff = win_pct - predicted_win_pct
  )

cat("\nTeams Over/Under-performing Net Rating:\n")
print(team_season_stats %>%
  select(team_display_name, win_pct, predicted_win_pct, win_diff) %>%
  arrange(desc(win_diff)))

# =============================================================================
# 4. Visualization: Offensive vs Defensive Efficiency
# =============================================================================

ggplot(team_season_stats,
       aes(x = ortg, y = drtg, size = win_pct, color = net_rtg)) +
  geom_point(alpha = 0.7) +
  geom_text_repel(aes(label = team_display_name), size = 3) +
  geom_hline(yintercept = mean(team_season_stats$drtg),
             linetype = "dashed", alpha = 0.5) +
  geom_vline(xintercept = mean(team_season_stats$ortg),
             linetype = "dashed", alpha = 0.5) +
  scale_color_gradient2(low = "red", mid = "yellow", high = "darkgreen",
                        midpoint = 0, name = "Net Rating") +
  scale_size_continuous(name = "Win %", labels = percent_format()) +
  scale_y_reverse() +  # Lower DRtg is better
  labs(
    title = "WNBA Team Efficiency Map 2024",
    subtitle = "Offensive vs Defensive Rating (size = Win%)",
    x = "Offensive Rating (pts per 100 poss)",
    y = "Defensive Rating (pts allowed per 100 poss)"
  ) +
  theme_minimal()

# =============================================================================
# 5. Team Style Classification
# =============================================================================

team_season_stats <- team_season_stats %>%
  mutate(
    tempo = if_else(pace > median(pace), "Fast", "Slow"),
    shooting_style = if_else(fg3a / fga > 0.35, "Three-Heavy", "Inside"),
    style = paste0(tempo, "-", shooting_style)
  )

cat("\n=== Team Playing Styles ===\n")
print(team_season_stats %>%
  select(team_display_name, pace, style, net_rtg, win_pct) %>%
  arrange(desc(net_rtg)))

# Style performance comparison
style_summary <- team_season_stats %>%
  group_by(style) %>%
  summarise(
    teams = n(),
    avg_net_rtg = mean(net_rtg),
    avg_win_pct = mean(win_pct),
    avg_ortg = mean(ortg),
    avg_drtg = mean(drtg),
    .groups = "drop"
  )

cat("\n=== Performance by Style ===\n")
print(style_summary)

# =============================================================================
# 6. Four Factors Correlation with Winning
# =============================================================================

four_factors_impact <- team_season_stats %>%
  summarise(
    efg_cor = cor(efg_pct, win_pct),
    tov_cor = cor(-tov_pct, win_pct),  # Negative because lower is better
    orb_cor = cor(orb_pct, win_pct),
    ft_cor = cor(ft_rate, win_pct)
  )

cat("\n=== Four Factors Correlation with Winning ===\n")
cat(sprintf("Shooting (eFG%%): %.3f\n", four_factors_impact$efg_cor))
cat(sprintf("Turnovers (low TOV%%): %.3f\n", four_factors_impact$tov_cor))
cat(sprintf("Rebounding (ORB%%): %.3f\n", four_factors_impact$orb_cor))
cat(sprintf("Free Throws (FT Rate): %.3f\n", four_factors_impact$ft_cor))

# =============================================================================
# 7. League-Wide Trends
# =============================================================================

cat("\n=== WNBA League Averages ===\n")
league_averages <- team_season_stats %>%
  summarise(
    avg_pace = mean(pace),
    avg_ortg = mean(ortg),
    avg_ppg = mean(ppg),
    avg_3pa_rate = mean(fg3a / fga),
    avg_efg = mean(efg_pct),
    .groups = "drop"
  )
print(league_averages)

cat("\n=== Team Analytics Complete ===\n")
cat("✓ Efficiency ratings (ORtg, DRtg, Net Rating)\n")
cat("✓ Four Factors analysis\n")
cat("✓ Win prediction models\n")
cat("✓ Playing style classification\n")
cat("✓ League trend analysis\n")

Four Factors Hierarchy

Research shows that in basketball, the Four Factors have different predictive weights for team success. Shooting efficiency (eFG%) is typically the most important factor, followed by turnovers, rebounding, and free throw rate. This hierarchy generally holds true in the WNBA as well.