How a 0-100 "did they rise or fade" score gets built from box score deltas
Playoff riser stat looks at a player change in box score statistics, taking into account the overall drop in offense that teams in the playoff experience. Defense gets more physical, tighter, and home advantage increases. Because of this, playoff stats are expected to drop regardles on if a player is playing below their level. From the past ~10 seasons, TS drops on average ~3% (adj per season), PTS drop on average ~1.5 pts per player, around half a rebound and half an assist is lost due to simply the shift from regular to playoff basketball. Adjusting to this anchors a baseline that we can expect before analyzing actual player change.
90+ requires both: an absolute elite level of play and a real rise above expectation. A role player can rise just as much as a star in z-score terms, but without the raw production to back it up they cap out around the 55-65 range. Jumping from 10 points to 14 points in the playoffs is great, but cannot be compared to rising from 25 to 29 points.
For seven stats (points, assists, rebounds, steals, blocks, turnovers, true shooting %), we start from the player's regular-season per-game average and add the league-wide average playoff drop for that stat and season — scaled by how tough their specific playoff opponents were on defense, relative to league average.
Turnovers are flipped (fewer than expected is good). The league-average drop is calculated per-season from every qualifying player-season back to 2012-13, so a year with unusually stingy playoff defense doesn't unfairly punish everyone in that postseason.
The league-average drop above is a flat number for the whole postseason — but a team that runs through three elite defenses on the way to the Finals faced a much tougher gauntlet than a team that got a soft first-round matchup and bowed out. opponent_difficulty rescales the expected drop to account for exactly who a player's team actually played.
A team's playoff opponents are pulled from the actual playoff bracket results, and each opponent's defensive rating (DRTG, points allowed per 100 possessions) for that season is averaged across every series the player's team played. That average is compared against the league-average DRTG for the same season.
Facing tougher-than-average defenses (a lower average opponent DRTG) pushes opponent_difficulty above 1, which scales the league-average drop further negative. The expected stat line gets pulled down harder, so the bar for "met expectation" is lower. The same can be said for the opposite if you face outlier weaker playoff defenses we don't expect you to drop as much so the bar is slightly higher.
Every stat's over_expectation value is z-scored within its season so a point of points isn't worth the same as a point of steals — each stat is measured against how spread out that stat naturally is. The same is done with raw playoff production (not the delta) to capture absolute level. That gives two numbers per player:
Each of the seven stats is weighted before being summed into these two scores:
| Stat | Weight |
|---|---|
| Points | 0.40 |
| Assists | 0.16 |
| True Shooting % | 0.10 |
| Rebounds | 0.10 |
| Turnovers | 0.08 |
| Steals | 0.09 |
| Blocks | 0.07 |
RISE_Z and LEVEL_Z are then combined into three additive components on top of the 50-point baseline:
RISE_POINTS rewards every player for rising above expectation, capped so a role-player rise alone tops out around +22 — it can't carry someone to superstar territory by itself.
ELITE_BONUS is the multiplier that separates a star's rise from a role player's rise. Rising from already a great player to a elite player is something that is rewarded much heavier than a role player increasing their scoring 3 points in the playoffs. Looking at purely raw delta we can players who jump 10 to 15, over players who jumped 25 to 29. While going from 10-15 is rewarded in this system it is not rewarded as much as a jump from an already All-Star level player.
LEVEL_POINTS is a small cushion so an elite player who simply maintains their level (rise near 0) still lands a bit above 50 — being elite is itself worth something, even without a rise.
The raw score is bucketed into a label, with one guardrail: Legendary and Elite Run both require reaching at least the Conference Finals — a gaudy score from a first-round exit can't claim either label.
| Score | Tier (Conf Finals+) | Tier (else) |
|---|---|---|
| 90+ | Legendary | Elite Run |
| 75-89 | Elite Run | Rose |
| 65-74 | Rose | Rose |
| 55-64 | Slight Rise | Slight Rise |
| 45-54 | Maintained | Maintained |
| 40-44 | Slight Dip | Slight Dip |
| 30-39 | Faded | Faded |
| <30 | Ghost | Ghost |
Live from the same pipeline described above.
Loading...
See the full Playoff Risers leaderboard for every player-season back to 2014-15, career averages, and head-to-head comparisons.
Limitations
Minimum-sample filters (4+ playoff games, 15+ minutes, 5+ FGA, 25+ regular-season games) exist to cut out noise, but a hot or cold week in a 4-game sweep can still swing a score more than it would across an 82-game regular season.
Opponent difficulty is based on team-level defensive rating, not on the specific defender a player faced It can't credit someone for torching an elite point-of-attack defender on an otherwise average defense.
Simple box score stat analysis, doesn't capture other forms of impact that Plus Minus metrics capture.
Strengths
Anchored to an expectation that already accounts for the league-wide playoff scoring drop, not a flat "regular season vs. playoffs" comparison.
Separates "did they rise" from "are they elite" so role players and stars are graded on different curves instead of the same one.