How strong was he while he played?
A 0–100 score compared only with players from the same NBA season.
Menu
Last updated Oct 6, 2026
How the rankings work
How hussEyquation ranks NBA players
hussEyquation estimates how much a player helped, then turns that estimate into season, career, and team rankings. Here is the method in plain English.
76 NBA seasons450 minutes before an official rank0–100 season scale2026–27 rules locked in advance
The short version
What each ranking answers
How much did he contribute all year?
Impact plus total workload. Playing more matters here.
How high—and how often—did he rank?
A peak-weighted blend of the best three, five, and ten season finishes.
How much did he contribute over his full career?
The sum of every regular season's fixed-baseline value contribution.
How strong was the whole roster?
Every player's contribution, weighted by how much he played for that team.
01 · Player seasons
How Impact Rate is built
Impact Rate is about level of play, not total games played.
- 1
What the player did
The model uses scoring and shooting, passing, rebounds, steals, blocks, turnovers, fouls, and minutes. It accounts for position, role, and how fast the player's team played.
- 2
How small samples are handled
A hot stretch can make a short-season player look better than he really is. Below 450 minutes, the estimate is pulled toward a more ordinary level and the player does not receive an official rank.
- 3
How on-court results are used
The model looks at how the team performed during the player's lineups and tries to separate his effect from his teammates. Several seasons are used to make that check steadier. When only one season is available, the model uses 75% of that adjustment. With at least two seasons, it uses the full adjustment, subject to the same limit for every player.
Maximum effect: the on-court check can move the estimate by no more than 1.25 impact points in either direction.
- 4
How shorter nightly roles affect the score
Players at 30 or more minutes per game keep full credit. Below 30, the part of the score above 25 is reduced gradually. The largest reduction is 20% at 12 minutes per game or less. The rankings also include a switch that removes this adjustment.
Example: at the maximum reduction, a score of 85 becomes 73 because only the 60 points above the 25-point floor are reduced by 20%.
- 5
How players are compared within a season
The final result is placed on a 0–100 scale. A score of 100 means the best result in that season—not automatically the greatest season ever.
02 · Full-season contribution
How Season Value is different
Impact Rate measures how effective a player was while he was on the court. Season Value adds how much he played over the full regular season.
Season Value starts with the player's impact above a reference line set near the lower end of established rotation players. It then counts how many possessions he played over the full regular season.
That means an excellent player in limited minutes can lead in Impact Rate but finish much lower in Season Value. A durable star can rank higher in Season Value because he sustained strong play over a much larger workload.
Example: at the same impact level, 2,000 possessions create twice as much Season Value as 1,000 possessions.
No double penalty.
The minutes-per-game adjustment used by Impact Rate is not applied again. Season Value already accounts for playing time through the player's actual workload.
03 · All-time careers
How the career ranking works
The career list rewards peak, prime, and staying power, with its peak emphasis stated explicitly.
- 1
Rank every season
Each qualified season receives a league finish: first, second, third, and so on.
- 2
Use three views of the career
The model averages the player's best up to three qualified finishes, best up to five, and best up to ten. The three-season average receives 50% of the score, the five-season average 30%, and the ten-season average 20%.
- 3
How shorter qualified careers stay eligible
At least one season must reach the 450-minute floor. After that, missing slots in the Best 3, Best 5, and Best 10 views receive a replacement finish one place below the median qualified-player field size across that player's qualified seasons. Short careers stay eligible, but a season that was never played is no longer treated like a top-25 season.
- 4
Understand the effective weights
Because the windows overlap, the best three seasons appear in all three views and supply 74% of the score. Seasons four and five supply 16%. Seasons six through ten supply 10%. There is no additional bonus for No. 1 finishes.
Version: rank-first-v2-1-2026-10-06. This release corrects the treatment of incomplete careers without changing any observed season rank or rescaling one era against another.
Lower is better on the career list.
The score is an average league finish, so a 2.8 is stronger than a 6.4. If two players tie, the model first favors more No. 1 seasons, then larger gaps over the league, more top-five seasons, and finally more total season value.
Download the pre-registered full-field audit · Download the named sanity panel
04 · NBA history
How older players are handled
The farther back the season, the less the NBA recorded.
The archive begins in 1950–51. Complete modern box-score inputs begin in 1973–74. For earlier seasons, hussEyquation estimates missing pieces from the statistics and team information that still exist.
Defense is especially difficult. For seasons without modern lineup data, the model adds a small, limited Historical Defense Estimate based on the defensive information available for the player and his team. Every season that uses an estimate is labeled.
These are informed estimates, not recovered facts. Older players therefore receive wider possible rank ranges. Most importantly, every player is judged against the league he actually played in.
Maximum effect: the Historical Defense Estimate can move a season by no more than one internal impact point in either direction.
05 · Teams
How the team rankings work
Add the roster together, then compare it with what happened on the court.
The upcoming-season projection starts with the current roster. Each returning player brings his completed full Impact estimate from last season, rookies receive a clearly labeled draft estimate, and the full roster is assigned exactly 240 healthy-roster minutes per game. The model then estimates how many games each player is likely to deliver from his recent availability, role and age. The Box Forecast helps estimate roles but does not determine the primary team-strength score.
Most of the power rating comes from projected player strength, roles and expected games. Last season's point differential supplies a smaller reality check. When a returning roster looked strong on paper but the team performed poorly, half of that unexplained gap carries forward in proportion to how much of the same rotation returns.
The rank bands are calibrated against past preseason forecasts and final point-differential strength. The main range is the tightest span containing 68% of 25,000 calibrated outcomes. A broader 80% range remains available in each team's details. Confidence is higher when a team returns most of its rotation and lower when availability, roster turnover, rookies, or returning players create more uncertainty. On the untouched 2021–26 test, the main range covered 65.0% of teams and had a median width of 12 ranking places.
Each player's season impact is weighted by how many possessions he played for that team. If a player was traded, his minutes are split between the teams that actually received them.
Roster Impact shows the average quality of the roster's minutes. Total Roster Value shows the full amount delivered over the schedule.
The result is compared afterward with the standings and each team's point differential. Wins are not added to the formula just to make the ranking match the standings.
What the ranking does not claim
Important limits
Playoffs, championships, awards, and reputation do not add points.
Shot difficulty, defensive attention, off-ball movement, and every tactical detail are not directly recorded.
The model cannot recreate possessions that were never tracked.
Players who share most of their minutes can still be difficult to separate.
The preseason team projection estimates next season. Player availability, trades, rookies, and role changes can move it quickly.
The same rules apply to every player, even when one result feels surprising.
How changes are handled
The rules lock on opening night
The preseason development window remains open through opening night. Beginning with the first 2026–27 regular-season game, new statistics will update the rankings, but the formulas will stay fixed through the season. Any emergency correction will receive a new version, a dated changelog entry, and a preserved prior edition. After the season, proposed improvements must again be judged across the full league.