Basketball Analytics for Beginners: Beyond Points Per Game
Points per game is the most commonly cited statistic in basketball. It is also one of the most misleading. A player who scores 25 points per game on 45 field goal attempts is hurting their team. A player who scores 15 points per game on elite efficiency while playing lockdown defense might be far more valuable. Yet casual fans, media narratives, and even some front offices still treat PPG as the primary measure of a player's worth.
Basketball analytics exist to cut through these misleading surface-level numbers and reveal what actually drives winning. This guide is your introduction to advanced basketball stats — what they measure, how they are calculated, why they matter, and where to find them. By the end, you will never look at a box score the same way again.
Why Points Per Game Is Misleading
Before diving into advanced metrics, it is worth understanding exactly why the most popular stat in basketball fails as a comprehensive measure.
PPG tells you how many points a player scores on average, but it tells you nothing about:
- How many possessions it took — A player averaging 25 PPG on 30 shots per game is far less efficient than one averaging 22 PPG on 18 shots
- What kind of shots they took — Two-pointers, three-pointers, and free throws are all worth different amounts, but PPG treats them equally
- What happened on the other end — A 25 PPG scorer who gives up 28 PPG as a defender is a net negative
- Team context — A player on a bad team with a high usage rate will accumulate points that would not be available on a good team
- Pace — Teams that play faster have more possessions, meaning more raw counting stats for everyone
The fundamental problem: PPG measures volume, not value. Analytics exists to measure value.
The Essential Advanced Stats
Here are the key metrics that any serious basketball fan or analyst should understand. Each one addresses specific limitations of traditional stats.
True Shooting Percentage (TS%)
What it measures: Overall shooting efficiency, accounting for the different values of two-pointers, three-pointers, and free throws.
Why it matters: Field goal percentage (FG%) treats all made shots equally. A player who shoots 45% but takes mostly threes is actually more efficient than one who shoots 50% on all two-pointers. TS% solves this by incorporating all scoring into a single efficiency number.
The formula:
TS% = Points / (2 × (FGA + 0.44 × FTA))
The 0.44 multiplier on free throw attempts accounts for the fact that not all free throws represent full possessions (and-ones, technical free throws, three-shot fouls).
How to interpret it: The league average TS% typically hovers around 56-58%. An elite scorer will post a TS% above 60%. A TS% below 52% on significant volume indicates inefficiency.
Example: In the 2023-24 season, several high-volume scorers posted TS% numbers above 65%, showing that efficiency and volume are not mutually exclusive at the highest levels.
Player Efficiency Rating (PER)
What it measures: A per-minute rating of a player's total statistical production, adjusted for pace.
Why it matters: PER attempts to distill everything a player does into one number — scoring, rebounding, assists, steals, blocks, turnovers, and missed shots — adjusted for the pace of the team and the league.
How to interpret it: League average PER is always 15.0 by design. An MVP-caliber season typically produces a PER above 27. A PER above 20 indicates an All-Star level player. Below 10 suggests a player struggling to contribute.
Limitations: PER has significant flaws. It overvalues volume scorers, undervalues defensive specialists, does not capture spacing or gravity, and relies heavily on box score stats that miss important contributions. It was groundbreaking when John Hollinger created it but has been largely superseded by more sophisticated metrics. Still, it is useful as a quick gut-check.
Box Plus/Minus (BPM)
What it measures: A player's estimated contribution per 100 possessions above a league-average player, based on box score statistics.
Why it matters: BPM estimates how many points per 100 possessions a player contributes relative to an average player, separated into offensive (OBPM) and defensive (DBPM) components. Unlike PER, BPM accounts for position and is centered on zero rather than an arbitrary league average.
How to interpret it:
| BPM Value | Player Level |
|---|---|
| +10 or above | All-time great season (MVP) |
| +6 to +9 | All-NBA caliber |
| +3 to +5 | All-Star / high-quality starter |
| +1 to +2 | Above-average starter |
| -1 to +1 | Average player |
| -2 to -1 | Below-average player |
| Below -2 | Replacement level or worse |
Limitations: Like PER, BPM is box-score based, so it cannot capture everything a player does. Defense is particularly hard to measure from the box score alone.
Value Over Replacement Player (VORP)
What it measures: The total box score contribution of a player over what a replacement-level player would provide, over the course of a season.
Why it matters: While BPM is a rate stat (per 100 possessions), VORP accumulates over the season. This means it rewards players who contribute at a high level while also staying healthy and playing significant minutes. It answers the question: "How much did this player help their team over the course of the full season compared to a freely available replacement?"
The formula:
VORP = (BPM - (-2.0)) × (% of team possessions played) × (team games / 82)
The -2.0 represents replacement level.
How to interpret it: A VORP of 0.0 means the player performed at replacement level. A VORP above 3.0 represents a strong starter's season. Above 5.0 is All-Star caliber. The highest single-season VORPs in history exceed 10.0, achieved by all-time greats in their peak years.
Win Shares (WS)
What it measures: An estimate of the number of wins a player contributes to their team through their offense and defense.
Why it matters: Win shares translate individual performance directly into team wins, making it intuitive to understand. If a team wins 50 games, the win shares of all its players should sum to approximately 50.
Components: Win shares split into Offensive Win Shares (OWS) and Defensive Win Shares (DWS). The offensive component is generally considered more reliable than the defensive component.
How to interpret it: A typical starter accumulates 4-6 win shares per season. An All-Star caliber player earns 8-10. The greatest individual seasons in NBA history have produced 20+ win shares.
Net Rating
What it measures: The point differential per 100 possessions when a player (or lineup) is on the court.
Why it matters: Net rating is conceptually simple — how much does the team outscore opponents per 100 possessions with this player on the floor? It combines offensive rating (points scored per 100 possessions) and defensive rating (points allowed per 100 possessions).
How to interpret it: A positive net rating means the team outscores opponents when the player is on the court. Elite players post net ratings of +8 to +15. The difference between a player's on-court net rating and the team's net rating when they sit (on/off differential) is particularly revealing.
Limitations: Net rating is heavily influenced by teammates and opponents. A player who always shares the court with other stars will have an inflated net rating. Lineup data with small sample sizes can be very noisy.
Usage Rate (USG%)
What it measures: The percentage of team possessions a player "uses" (by shooting, getting fouled, or turning the ball over) while on the court.
The formula:
USG% = 100 × ((FGA + 0.44 × FTA + TOV) × (Team Minutes / 5)) / (Minutes × (Team FGA + 0.44 × Team FTA + Team TOV))
Why it matters: Usage rate provides critical context for all other stats. A player averaging 20 PPG on 30% usage is doing something very different from one averaging 20 PPG on 22% usage. Generally, efficiency decreases as usage increases, so maintaining high efficiency at high usage is one of the hallmarks of elite scorers.
How to interpret it: League average usage is 20% (since five players share possessions). Usage above 28% is extremely high. The highest usage rates in NBA history approach 40%.
Traditional vs. Advanced Stats: A Side-by-Side Comparison
To see why advanced basketball stats tell a different story, consider two hypothetical players with similar traditional stat lines.
Hypothetical Player Comparison
| Statistic | Player A | Player B |
|---|---|---|
| Points per game | 24.5 | 21.0 |
| FG% | .440 | .475 |
| 3P% | .340 | .395 |
| FGA per game | 22.0 | 15.5 |
| FTA per game | 5.5 | 7.0 |
| FT% | .810 | .870 |
| Rebounds per game | 5.0 | 6.5 |
| Assists per game | 4.0 | 5.5 |
| Turnovers per game | 3.5 | 2.0 |
| True Shooting % | .535 | .627 |
| Usage Rate | 31.0% | 24.5% |
| PER | 18.5 | 22.0 |
| BPM | +1.5 | +5.5 |
| VORP | 2.0 | 4.5 |
| Win Shares | 5.0 | 9.5 |
A casual fan looking at traditional stats sees Player A as the better player — more points, clearly the "go-to guy." But the advanced stats reveal that Player B is dramatically more valuable: far more efficient, fewer turnovers relative to creation, better rebounding and passing, and nearly double the win contribution. Player A is a volume scorer whose high usage actually costs his team possessions better used by teammates.
This is exactly the kind of insight that basketball analytics provides.
Pace-Adjusted Stats: Why Context Matters
One of the most important concepts in basketball analytics is pace adjustment. Teams play at wildly different speeds — some push the ball and play 100+ possessions per game, others slow it down to 90 or fewer.
Raw counting stats (points, rebounds, assists) are inflated by pace. A player on a fast-paced team will accumulate more raw stats simply because there are more possessions. Per-100-possession stats normalize for this, allowing fair comparison between players on fast and slow teams.
Example: A center averaging 10 rebounds per game on a slow-paced team might be rebounding at a rate of 15.0 per 100 possessions. A center averaging 12 rebounds on a fast-paced team might actually be rebounding at only 13.5 per 100 possessions. The first player is the better rebounder, despite the lower raw number.
Always check whether a stat is per-game (raw) or per-possession (pace-adjusted) before drawing conclusions.
Moreyball and the Three-Point Revolution
No discussion of basketball analytics is complete without addressing how analytics transformed the way basketball is played, not just measured.
Daryl Morey, as general manager of the Houston Rockets starting in 2007, pioneered the application of rigorous analytics to basketball strategy. The core insight was deceptively simple: the expected value of different shots is not what traditional basketball wisdom assumed.
The Math Behind Moreyball
- A two-point shot at league-average accuracy (approximately 50% from mid-range): 0.50 × 2 = 1.00 points per attempt
- A three-point shot at league-average accuracy (approximately 36%): 0.36 × 3 = 1.08 points per attempt
- A shot at the rim at typical accuracy (approximately 64%): 0.64 × 2 = 1.28 points per attempt
- A free throw (approximately 77% accuracy): 0.77 × 1 = 0.77 points per attempt, but getting to the line often comes with and-one potential and fouls on the defense
The conclusion: mid-range two-point jumpers are the least efficient shot in basketball. Teams should maximize three-point attempts and shots at the rim while minimizing mid-range jumpers.
The League-Wide Impact
The NBA has undergone a dramatic transformation. In the 2001-02 season, teams averaged about 14.7 three-point attempts per game. By the 2023-24 season, that number exceeded 35 per game. The mid-range jumper has not disappeared, but it is increasingly reserved for elite mid-range shooters and late-clock situations.
This shift has not just changed shot selection — it has changed player evaluation, roster construction, spacing philosophy, and defensive strategy. Players who cannot shoot threes have seen their market value decline. Stretch bigs who can space the floor have become premium assets. The entire geometry of the game has been redrawn by analytics.
Where to Find Advanced Stats
One of the best things about basketball analytics is the accessibility of data. Here are the essential resources for anyone wanting to explore advanced basketball stats.
Free Resources
- Basketball Reference (basketball-reference.com): The gold standard for historical and current stats. Includes TS%, PER, BPM, VORP, Win Shares, and extensive lineup data. The "Advanced" tab on any player's page is a treasure trove.
- NBA.com/stats: Official NBA statistics, including tracking data (speed, distance, touches), shot dashboard, hustle stats, and advanced metrics. The best source for on/off court data.
- Cleaning the Glass (cleaningtheglass.com): Founded by Ben Falk (former NBA analytics staffer). Free tier offers excellent analysis; paid tier provides the most thoughtfully curated advanced stats available, with garbage time and opponent-strength adjustments.
Paid and Premium Resources
- Second Spectrum / NBA tracking data: The league's official optical tracking system. Not publicly available, but some metrics derived from it appear on NBA.com.
- Synergy Sports: Play-type data (pick and roll, isolation, post-up, spot-up, etc.) showing efficiency by play type. Some access available through NBA.com.
- PBP Stats (pbpstats.com): Free play-by-play derived analytics with unique metrics not found elsewhere.
Data for Your Own Analysis
- NBA API (unofficial Python wrappers like
nba_api): Pull raw data for custom analysis - Basketball Reference's export features: Download stat tables as CSV files
- Kaggle: Various NBA datasets available for machine learning projects
Getting Started With Your Own Analysis
If you want to go beyond reading stats and start doing your own basketball analysis, here is a practical roadmap:
-
Start by watching games with advanced stats in mind. Notice shot selection, spacing, defensive rotations, and pace. Ask yourself: is this team creating good shots or bad shots?
-
Pick a question to investigate. For example: "Is Player X really better than Player Y?" Pull their advanced stats and compare across multiple metrics, not just one.
-
Learn to use Basketball Reference effectively. Explore the play index, game finder, and lineup tools. These alone can power years of analysis.
-
Learn basic data analysis. Python with pandas, or even just Excel, will let you explore NBA data more deeply than any website alone.
-
Read the foundational works. Dean Oliver's Basketball on Paper, the Thinking Basketball YouTube channel, and Seth Partnow's The Midrange Theory are excellent starting points.
-
Be humble about what stats can and cannot tell you. Every metric has limitations. The best analysts combine statistical evidence with film study and contextual understanding.
What Analytics Cannot Capture (Yet)
It is important to acknowledge the limits of even the best basketball analytics:
- Leadership and chemistry — Stats cannot measure who makes their teammates play harder or smarter
- Defensive positioning and communication — Much of elite defense is about being in the right place, which tracking data captures imperfectly
- Gravity and attention — Stephen Curry warps defenses even without touching the ball; this creates value for teammates that is difficult to attribute
- Clutch performance — Small sample sizes make it nearly impossible to statistically identify "clutch" players, though the narrative power is real
- Development trajectory — Analytics describe what has happened, not what will happen. Projecting improvement is still partly art
The best basketball minds combine analytics with scouting, film study, and human judgment. The goal of basketball analytics is not to replace watching the game — it is to watch the game with better tools for understanding what you are seeing.
For a comprehensive deep dive into how NBA teams build analytical frameworks, evaluate talent, and make strategic decisions using data, explore our full treatment in Professional Basketball Analytics.