League context for RNP (percentile & rank)

How we compare a player’s season Running Net Points to everyone else in the same season slice on the site — same data as the leaderboard, with a minimum games rule.

Credit: nba-rnp web

What the numbers mean

Percentile (~N%) is the approximate share of qualified players whose season RNP total is strictly lower than this player’s. It is not opponent-adjusted and not a projection — only a rank within the population we have loaded for the active season.

Rank (e.g. 12 / 240) is the player’s position when qualified players are sorted by season RNP total (highest first). Ties: anyone with the same total shares the same “worse than” counts, so rank and percentile line up with strict > / < comparisons.

Who counts as “qualified”

To avoid noisy totals from tiny samples, league percentile and rank only apply to players with at least 5 games in the season (player_game_stats rows joined to that season’s games). The threshold is the MIN_GAMES_QUALIFIED_RNP constant in web/lib/rnp-league-rules.ts (keep the doc and constant in sync when changing it).

Below that cutoff, the UI shows an em dash (—) on the leaderboard and a short note on the player page instead of a percentile.

What we aggregate

We use the same query as /players: per-player sum of running_net_points over the season and count of games played. Team shown on the leaderboard is still the most recent team in that season (trades appear as one row per player).

Limitations

Single-game and short-sample RNP stay noisy (play-by-play tagging). Percentiles summarize season totals only; they do not replace reading the game log or methodology on Running Net Points.