Why the NBA’s Middle Stat Line Depends on Who You Count

There is no universal set of median NBA player statistics. A credible result must identify the season, separate the regular season from the playoffs, define which players are eligible, and specify whether the figures are season totals, per game, per 36 minutes, or per possession. Without those decisions and a complete player-level dataset, an exact median stat line is not defensible.

The short answer: there is no universal median NBA stat line

The median changes with the population being measured. Counting every player who appeared in a game produces a different benchmark from counting only players who met a games or minutes threshold.

NBA’s Season Leaders page, when set to the 2025–26 playoffs, reports 155 player rows and offers controls for season, season type and statistical mode. A leaders view ordered by a category such as points cannot by itself establish a league-wide median, which requires the complete set of eligible observations (NBA Season Leaders).

The NBA Players Traditional view reports 230 rows under its displayed 2025–26 playoff settings. That count should not be combined or reconciled with the leaders-page count without confirming that the pages have identical filters, scope and treatment of player records (NBA Players Traditional statistics).

Neither page supports treating those playoff figures as regular-season norms. Nor is it valid to estimate a median from the scoring leaders or the first portion of a longer table.

Median versus mean in NBA statistics

The median is the middle value after observations have been arranged from lowest to highest. If the dataset contains an odd number of observations, the median is the single middle value. With an even number, the usual convention is to average the two central values.

The arithmetic mean is calculated by adding every observation and dividing the total by the number of observations. It is more sensitive than the median to unusually high or low values, as the NIST explanation of measures of location describes.

Consider an ordered list of player scoring rates. The rate in the middle is the median, while every rate—including the highest and lowest—contributes to the mean. The median can therefore be more resistant to a small number of extreme observations.

Neither measure automatically represents a typical rotation player. If every player who appeared is included, the calculation describes that all-player population. If eligibility requires a minimum number of games or minutes, it describes the qualified group instead.

The choices that change the median

Two analyses can both be accurate yet produce different median NBA player statistics because they define the question differently.

Methodological choice	What it measures	Main limitation
Every player who appeared	The middle among all participants	One-game and low-minute players can influence the result
Games-qualified players	The middle among players meeting a games cutoff	The analyst-selected cutoff excludes some players
Minutes-qualified players	The middle among players with an established workload	It describes a rotation group, not everyone who appeared
Per game	Production in an average player appearance	Playing time and role strongly affect the result
Season totals	Accumulated production across the season	Availability and games played have substantial influence
Per 36 minutes	Production normalized to a common minutes scale	Small samples can become misleading when scaled upward
Per possession	Production adjusted for pace and opportunity	The result depends on the chosen possession mode

None of these options is universally correct. Per-game medians describe ordinary box-score production by appearance, while totals emphasize cumulative contribution.

Shooting percentages require another explicit policy. A player with no attempts has no defined percentage, while a player with very few attempts may have an extreme but unstable result. An analysis should state how it handles zero-attempt records and whether field-goal, three-point or free-throw percentages require minimum attempts.

Regular-season and playoff medians should also be reported separately.

A reproducible calculation procedure

A defensible calculation should follow a documented workflow:

1. Select a completed season. Do not present an unfinished season as a final benchmark.
2. Choose the season type. Calculate regular-season and playoff results separately.
3. Select the statistical mode. Use totals, per game, per 36 minutes or a clearly defined per-possession mode.
4. Retrieve the complete player table. Do not calculate from leaders, search snippets or a single displayed page.
5. Define eligibility. State whether the population includes every player or requires minimum games, minutes or attempts.
6. Consolidate traded-player records. Use an official combined season record when available. Otherwise, aggregate the underlying counting totals first and recompute rates from their denominators. Do not average team-level per-game rates or shooting percentages.
7. Handle missing and undefined percentages. Establish rules for zero attempts, missing values and attempt qualifications.
8. Sort each category independently. Order all eligible observations from lowest to highest.
9. Take the middle value. For an even sample, apply the declared convention to the two central observations.

When team-level records must be combined, counting statistics should be summed across the player’s records. Per-game figures should then be calculated from the combined totals and games played. Shooting percentages should be recomputed from total makes and attempts rather than averaging the percentages shown for individual teams. Games should also be deduplicated if the source structure creates overlapping records.

The core output should cover GP, MIN, PTS, REB, AST, STL, BLK, TOV, FG%, 3P% and FT%. Each category needs its own median because the ordering of players changes from one statistic to another.

NBA player-statistics pages provide controls for season, season type, per-mode, position, starter status, experience, team, opponent, date range and other dimensions. Those controls support useful segmentation, but an unnoticed filter can change the eligible population. Every active setting should therefore be recorded before the data are exported or copied.

The published results should identify the filters, retrieval date, final sample size, eligibility threshold, trade-handling method and missing-value rules. Reporting the arithmetic mean beside each median is also useful because a substantial difference between them can show where extreme observations are influencing the mean.

A median stat line is not necessarily a real player

A table of category medians normally produces a synthetic stat line. The player in the middle for points may not be the player in the middle for rebounds, assists, minutes, turnovers or shooting percentage. Combining those separate values does not show that one player recorded the complete line.

Identifying the player closest to the overall median is a different analysis. A common procedure is to:

1. Construct a vector containing the median for each selected metric.
2. Place differently scaled statistics on comparable scales using a disclosed standardization method.
3. Measure each eligible player’s distance from that median vector.
4. Rank the players from the smallest distance to the largest.

The outcome depends on the design. Points and minutes may exert more influence under one method, while equal weighting of steals, blocks and percentages may lead to another result. The included metrics, eligible population, standardization method, category weights and distance formula must be disclosed before naming a closest player.

A community-created “most average player” analysis demonstrates this model-building approach by choosing an eligibility threshold, combining traded-player records and applying standardized distance. It is exploratory fan analysis, not an official NBA award or an objective designation, and its unverified player figures are not a substitute for a calculation from complete official data.

How to read a published median table

Before trusting a table of median NBA player statistics, check for:

• A named season
• Regular season or playoffs
• A complete dataset
• A definition of who counts as an eligible player
• The final sample size
• The statistical mode: totals, per game, per minute or per possession
• Any games, minutes or attempt cutoffs
• Treatment of traded-player records
• Rules for zero and low shooting-attempt totals
• The data-retrieval date

Do not treat a playoff table as a regular-season benchmark, and do not derive a league median from category leaders or a truncated first page. If the goal is to understand both the league-wide middle and a typical established role player, look for separate all-player and rotation-qualified results.

This standard is consistent with Free Throw Stats’ stated approach of explaining basketball statistics with official league data and relevant context. That editorial approach does not prove any particular median; the result must still come from a complete dataset and a transparent calculation.

The median NBA player is a defined analytical result, not a permanent stat line or official player designation. Without a complete official dataset and a disclosed method, an exact median should not be presented.