The Basketball Database
  • Home
  • Players
  • Teams
  • Seasons
  • Support
Seasons
2023-242022-232021-222020-212019-202018-192017-182016-172015-162014-152013-142012-132011-122010-112009-102008-092007-082006-072005-062004-052003-042002-032001-022000-011999-001998-991997-981996-971995-961994-951993-941992-931991-921990-911989-901988-891987-881986-871985-861984-851983-841982-831981-821980-811979-801978-791977-781976-771975-761974-751973-741972-731971-721970-711969-701968-691967-681966-671965-661964-651963-641962-631961-621960-611959-601958-591957-581956-571955-561954-551953-541952-531951-52

The Basketball Database is copyright 2022. The website's name, design, and some of the methods of statistical analysis used in the development of the site are the intellectual property of its creator.

The Greek God of Stats logos are the work of Lindsey McCown. They may not be copied, used, or reproduced without the expressed written permission of the artist.

Team logos appear here courtesy of SportsLogos.Net and Sports Encyclopedia. The images are the property of the NBA and respective member organizations.

Player Impact Plus-Minus (PIPM) is a model created by Jacob Goldstein. A writeup on the model can be found at Nylon Calculus . The coefficients and methodology laid out in this article are what I have used in producing the PIPM values presented here.

Box Plus-Minus (BPM) is the second iteration of a model developed by Daniel Myers. The most recent explanation of his methodology is posted at Basketball-Reference, though it should be noted that one must use data elements from Basketball-Reference in order to return the values reported by Myers. The values presented on this site typically employ data from https://www.nba.com/stats/, and thus vary from values presented by Basketball-Reference.

Regularized Adjusted Plus-Minus (RAPM) was originally developed by Jermias Engelmann, and both he and Steve Illardi have been heavily involved in the further development of variant forms of RAPM. There are a number of explanations of the necessary data elements and statistical methods used in RAPM which are readily available on blogs, as well as useful tutorials on Github and various websites. The values presented on this site are "vanilla" RAPM - the result of a pure ridge regression without any priors to tell the regression what the results "should" look like.

Luck-Adjusted ratings were initially developed by Jacob Goldstein with an explanation published on Nylon Calculus. A number of analysts have used this methodology in the development of models, including Ryan Davis' Luck-Adjusted RAPM.

Win Shares are an expression of Dean Oliver's ORating and DRating presented originally in his book Basketball On Paper. Win Shares were developed by the team at Basketball-Reference as a means of converting Oliver's per-100 ratings into credit for wins, similarly to Bill James' concept of "Win Shares" from the baseball analytics community. The methodology used to calculate Win Shares is described in detail at Basketball-Reference.

For further reading, visit the Greek God of Stats blog by clicking on the icon below, or read The Basektball Bible , a comprehensive work of statistical analysis and data visualization covering the early seasons of the player-tracking data era.