Product

Methodology

Every number on FiveVsFive is either an official counting stat or a relative rank inside a defined pool. We do not publish scores as probabilities, and we do not invent missing data.

This page is the contract with the reader. If a block cannot be calculated, it is hidden. Zeros are not placeholders. Rankings that need ten games do not appear at game three with a fake 50.

What this site is for

Searchers do not need another points-per-game table. They need to know how many minutes a player is actually getting, whether that role is locked, and what happens to the rotation when someone is out. Those questions drive minutes leaders, Role Stability lists, injury reports, and every team hub.

Role Stability

Role Stability answers: how predictable is this player’s playing time? We take the last 15 games with minutes greater than zero (injury DNPs out), compute the coefficient of variation of minutes, and scale it so 100 is a locked role and 0 is volatile.

  • 80–100 — Locked
  • 60–79 — Steady
  • 40–59 — Variable
  • 0–39 — Volatile

Minimum sample: eight games. Below that the score is not published. The player card also shows the recent high and low so the number is not a black box.

Composite

Composite is a 0–100 blend of percentiles inside a position group (guards, wings, bigs) — not across the whole league, and not five traditional positions.

  • Scoring 25% — points per 36
  • Efficiency 20% — true shooting
  • Volume 20% — usage and minutes
  • Playmaking / defense 20% — available AST%, STL%, BLK%, DREB%
  • Stability 15% — Role Stability

Under 10 games the score is marked provisional and stays off leaderboards. If any component is missing, the composite is omitted entirely.

Minutes Impact

When teammate B is out, how do player A’s minutes change? We compare games both played versus games A played and B was OUT. A pair needs at least four games without B. That table is the backbone of “who benefits if X is out” on the injury-status page.

Opportunity Score

Game day only, and only for AVAILABLE or PROBABLE players. It is expected possession volume from projected minutes, matchup pace, and usage — not a points forecast and not a pick. In the offseason and on non-game days the block is hidden.

Season counting stats

NBA season payloads often store totals. When season minutes exceed 48, we divide counting stats by games played so a leaderboard is not a pile of raw totals. Rates (FG%, TS%, USG%) stay rates. Splits and on/off blocks render only when the sample exists.

What we will not do

  • Score or winner predictions
  • Betting odds, implied totals, or “lock of the night” language
  • News recaps or injury-reason speculation
  • Fake zeros when a feed has not arrived

Sources are listed on data sources. Errors go through corrections.