Power Rankings
Each team’s projected net rating (points per 100 possessions better or worse than an average team), built from last season and every roster move, plus its odds from 10,000 simulated seasons. Once games start, every result moves it.
Players are ranked by lineup data. Every game is split into stretches between substitutions, when the same ten players are on the court. Because each player shows up in hundreds of those stretches, with different teammates and opponents, the rating can separate what he adds from what the players around him add. Last season counts too, so a few games can’t swing it. The on/off numbers in the write-ups are simpler: the team’s margin with him on the floor and with him off it. Example: if the Spurs are +17 whenever Wembanyama plays, but most of those minutes are with good starters who are also good without him, part of the +17 belongs to them.
How each team wins
Every team placed by its offense and defense. The further up and to the right, the better. Hover or tap a logo, or a team in the table, for its numbers.
Biggest offseason moves
Each player’s estimated effect on his new team’s net rating, and on the team he left.
Biggest climbs and drops
Against last season’s final ranking.
The rankings
All 30 teamsSan Antonio went 62-20, reached the Finals and brings back 94% of its minutes, the most in the league. Victor Wembanyama ranks third in the league by lineup data and second on defense; the Spurs were +17.0 per 100 possessions with him on the floor and +0.3 without him. Keldon Johnson (seven straight seasons as a regular) and Devin Vassell (six) were there before Wembanyama arrived, and newcomer Tobias Harris ranks in the top 50 (Detroit +10.0 with him on the floor, +6.3 without).
The team with the league’s top net rating and a 64-18 record returns 71% of its minutes from last season. Shai Gilgeous-Alexander ranks second in the league by lineup data, with Oklahoma City +16.3 per 100 possessions when he played and +3.7 when he sat, and Chet Holmgren ranks fifth on defense. The core goes back years, with Gilgeous-Alexander a regular for seven straight seasons, Jalen Williams four (he is back after playing 33 games) and Holmgren and Cason Wallace three.
Boston went 56-26 while Jayson Tatum played only 16 games. He came back from his torn Achilles in March and averaged 21.8 points, down from 26.8 the year before. The projection counts those games, where his lineup numbers dipped, and expects him to miss more games than usual. Derrick White ranks 11th in the league (Boston +11.3 per 100 possessions with him on the floor, +1.3 off), and Payton Pritchard has been a Boston regular for six straight seasons, White for five. Jaylen Brown averaged 28.7 points, but Boston allowed 9.6 more points per 100 with him on the floor, so losing him costs less than the scoring suggests.
Jalen Brunson and Josh Hart have been playing together for four seasons, OG Anunoby for three and Mikal Bridges and Karl-Anthony Towns for two, and that group won the title. Towns ranks 10th in the league by lineup data, and Brunson, the 2024-25 Clutch Player of the Year and last season’s Finals MVP, ranks ninth on offense. Mitchell Robinson, a top-60 defender now in Boston, is the biggest loss. The champions that slipped in recent years were the ones that got hurt, like the 2020-21 Lakers (45 games from LeBron James, 36 from Anthony Davis), so health is the question for a core that is mostly 30 or older.
Houston went 52-30 without Fred VanVleet, who missed all of last season after ranking 12th in the league by lineup data the year before. Marcus Smart, a top-25 player, arrives from the Lakers, who were +6.8 per 100 possessions with him on the floor and −2.4 without him, and Kevin Durant, at 38, still ranks 11th on offense. Amen Thompson ranks in the top 45, with Houston +6.6 per 100 when he played and +1.4 when he sat.
Detroit won 60 games and finished +6.9 per 100 possessions, but 22 of the 26 teams at +6 or better since 2016-17 were worse the next season, and Tobias Harris left (Detroit was +10.0 with him on the floor, +6.3 without him). Cade Cunningham ranks in the top 20 by lineup data, with Detroit +11.4 when he played and +3.8 when he sat, and Ausar Thompson ranks 11th on defense. The newcomers grade worse: Taurean Prince ranked near the bottom of the league in 2024-25 and played 26 games last season, and John Collins is below average.
Toronto went 46-36 and adds Kawhi Leonard, who ranks fourth in the league by lineup data; the Clippers were +7.5 per 100 possessions with him on the floor and −5.8 without him. He is 35 and played 65 games last season. Losing Brandon Ingram barely hurts, since Toronto was 5 points per 100 better with Ingram off the floor.
Portland went 42-40 and adds two guards coming off injuries: Damian Lillard, 36, who missed all of last season after ranking ninth on offense by lineup data in 2024-25, and Ja Morant, who ranked in the top 25 in 2024-25 but played 20 games for Memphis last season. Jrue Holiday is the steadiest piece, a top-30 player with Portland +5.5 per 100 possessions when he played and −4.0 when he sat.
Orlando brings back 90% of the team that went 45-37, and its core has been together longer than all but two teams’: Wendell Carter Jr. has been a regular for six straight seasons, Franz Wagner and Jalen Suggs five and Paolo Banchero four. Wagner ranks sixth in the league by lineup data but played only 34 games, and Desmond Bane, a top-40 player, had Orlando +2.1 per 100 possessions on the floor and −2.5 off. Banchero averaged 22.2 points but ranks below average, and Orlando was 2 points per 100 better with him on the bench.
Nikola Jokić is still the best player in the league by lineup data, and Denver was +10.8 per 100 possessions with him on the floor and −2.9 without him. But only 61% of last season’s minutes return, and DeMar DeRozan, 37 and projected for 29 minutes a night, ranks near the bottom on defense; Sacramento allowed about 10 more points per 100 with him on the floor.
Philadelphia added Jaylen Brown and LeBron James, who turns 42 in December, but Brown’s defense is the catch: Boston allowed 9.6 more points per 100 possessions with him on the floor. Tyrese Maxey (28.3 points a game) ranks in the top 50 by lineup data, and Joel Embiid ranks in the top 20 but played only 38 games. Only 56% of last season’s minutes return, among the lowest in the league.
Cleveland won 52 games and two playoff series, and its core goes back years: Jarrett Allen has been a regular for six straight seasons, Evan Mobley for five and Donovan Mitchell for four. Mitchell (27.9 points a game) ranks in the top 15 by lineup data, with Cleveland +7.1 per 100 possessions when he played and −0.7 when he sat. James Harden, 37, ranks 14th on offense but near the bottom on defense.
Devin Booker ranks in the top 20 by lineup data, and Phoenix was +3.8 per 100 possessions with him on the floor and −1.4 without him. New forward Miles Bridges ranks about average, though Charlotte was 3 points per 100 better without him, and Jalen Green, back after a 32-game season, ranks near the bottom on defense. Oso Ighodaro is the quiet plus, a top-40 defender with Phoenix +5.0 when he played and −1.5 when he sat.
Atlanta went 46-36, and Dyson Daniels is its best returning player by lineup data: the Hawks were +5.7 per 100 possessions with him on the floor and −3.3 without him. Jalen Johnson averaged 22.5 points, 10.3 rebounds and 7.9 assists, yet his lineup numbers are only around average. Luguentz Dort, arriving from Oklahoma City, ranks below average, and the Thunder were 5 points per 100 better with him on the bench.
Minnesota went 49-33 but lost Julius Randle and Naz Reid, and added LaMelo Ball, a top-25 player by lineup data. Rudy Gobert ranks first in the league on defense, with Minnesota allowing 6 fewer points per 100 possessions when he played, while Anthony Edwards averaged 28.8 points but the team was only 0.6 points better with him on the floor. Ayo Dosunmu, projected for 15 minutes a night, ranks in the league’s bottom 25.
Charlotte went 44-38 with the league’s eighth-best net rating, then lost LaMelo Ball, a top-25 player by lineup data; Charlotte was +9.1 per 100 possessions with him on the floor and +0.4 without him. Moussa Diabaté is the best of what remains, with Charlotte +10.4 when he played and −0.1 when he sat. The arrivals grade worse: Royce O’Neale ranks well below average and Dennis Schröder below average.
The Lakers went 53-29 but lost LeBron James and Marcus Smart, and only 42% of last season’s minutes return, the lowest in the league. Luka Dončić (33.5 points a game) ranks eighth by lineup data, with the Lakers +3.8 per 100 possessions when he played and −1.2 when he sat. Newcomers Quentin Grimes, projected for about 28 minutes a night, and Collin Sexton both rank well below average.
Golden State went 37-45 and carries one of the league’s two biggest age adjustments, with Stephen Curry at 38, Jimmy Butler at 37 and Draymond Green at 36. Curry and Butler still rank in the top 20 by lineup data when they play (Golden State was +7.4 per 100 possessions with Butler on the floor and −3.9 without him), but they played 43 and 38 games. Brandin Podziemski, a top-30 defender, played all 82.
Giannis Antetokounmpo arrives from Milwaukee as a top-5 player by lineup data (the Bucks were +4.3 per 100 possessions with him on the floor and −9.9 without him), but he played 36 games, so the projection counts him for about half a season. Tyler Herro, Jaime Jaquez Jr. and Kel’el Ware went the other way, and newcomers Klay Thompson, 36, and Bobby Portis rank below average. Bam Adebayo ranks seventh overall and fourth on defense.
Indiana fell to 19-63 without Tyrese Haliburton, who missed the season after ranking fifth on offense by lineup data in 2024-25. Pascal Siakam (24 points a game) ranks in the top 25, but Jarace Walker, Andrew Nembhard and Kelly Oubre Jr. all rank in the league’s bottom 60, and Indiana was 6.4 points per 100 possessions worse with Walker on the floor.
New Orleans brings back 85% of a 26-56 team. Herb Jones is its best player by lineup data and ranks sixth on defense, and Zion Williamson ranks 15th on offense, with New Orleans 4.7 points per 100 possessions better when he played. Jeremiah Fears and Derik Queen both rank well below average, and New Orleans was about 5 points per 100 worse with each of them on the floor.
Milwaukee went 32-50 and lost Giannis Antetokounmpo; the Bucks were +4.3 per 100 possessions with him on the floor and −9.9 without him, a 14-point swing. Of the players arriving from Miami, Jaime Jaquez Jr. and Kel’el Ware rank below average by lineup data and Tyler Herro about average, and Kyle Kuzma ranks near the bottom of the league.
Memphis went 25-57, and Jaren Jackson Jr., a top-15 defender, went to Utah late in the season; Memphis was 3 points per 100 possessions better with him on the floor. Jerami Grant, arriving from Portland, ranks in the league’s bottom 40. Zach Edey is back after playing 11 games, in which Memphis was +18.3 per 100 with him on the floor, a small sample.
Dallas went 26-56, and Kyrie Irving, who missed last season, ranked fifth in the league by lineup data in 2024-25. Cooper Flagg averaged 21 points as a teenager, but Dallas was about even with him on or off the floor. Naji Marshall and Max Christie, projected for 28 and 26 minutes a night, rank below average, and Anthony Davis is now in Washington.
Utah won 22 games. Lauri Markkanen ranks 14th in the league by lineup data, with Utah 10 points per 100 possessions better when he played, and Jaren Jackson Jr., who arrived from Memphis late last season, is a top-15 defender. Keyonte George and Isaiah Collier rank in the league’s bottom 15, mostly because of defense: Utah allowed 8 more points per 100 with George on the floor.
Brooklyn won 20 games. Day’Ron Sharpe and Michael Porter Jr. rate as its best players by lineup data, with Brooklyn 5.9 points per 100 possessions better when Porter played, and Julius Randle (21.1 points a game) arrives from Minnesota as an above-average player. Noah Clowney and Drake Powell rank well below average, while losing Nic Claxton, who ranked in the bottom 15, is a quiet plus.
Chicago went 31-51 and brings back only 55% of its minutes. Norman Powell (21.7 points a game) arrives from Miami and ranks above average, but Nic Claxton, from Brooklyn, ranks in the league’s bottom 15. Patrick Williams ranks fourth from the bottom, and Chicago was 9 points per 100 possessions worse when he played.
The Clippers went 42-40 after trading James Harden and Ivica Zubac during the season, and this summer Kawhi Leonard (a top-5 player by lineup data) went to Toronto and John Collins to Detroit, so only 49% of last season’s minutes return. Brandon Ingram (21.5 points a game) arrives from Toronto ranking below average, and Toronto was 5 points per 100 possessions better with him off the floor. Brook Lopez, 38, is still in the projected rotation.
Washington went 17-65 with the league’s worst net rating. Anthony Davis, traded from Dallas in February, has yet to play for Washington; he played 20 games last season and ranked in the top 30 by lineup data in 2024-25. Deandre Ayton comes from the Lakers ranking in the bottom 50. Bub Carrington ranks last in the league, and Washington was 9 points per 100 possessions worse when he played.
Sacramento went 22-60 with Domantas Sabonis playing 19 games; in 2024-25 he averaged 19.1 points and 13.9 rebounds and ranked in the top 20 on defense by lineup data. Zach LaVine and Malik Monk both rank in the bottom 60 on defense, and Maxime Raynaud, projected for 24 minutes, ranks in the bottom 40 overall.
Move: change from last season’s final ranking. Proj. record: the average of 10,000 simulated seasons, with the middle 80% of win totals under it. Playoffs counts teams that get in through the play-in; Finals is the chance to reach the NBA Finals, Title the chance to win it. Off and Def: projected league rank, with points per 100 better or worse than average. The number beside a player is his estimated effect on his team’s net rating. Back: returning from an injury that cost him most of last season. Player ranks in the write-ups come from lineup data and compare the roughly 300 players who played 2,000 or more possessions last season.
How these rankings work
Each team gets an offensive and a defensive rating in points per 100 possessions, adjusted for the opponent and for home court, from NBA.com box scores. Net rating combines the two.
Before the season, each team starts from last season’s rating (playoff games count like regular-season games), pulled toward average, since strong and weak seasons tend to drift back toward the middle. Then every roster change is counted, player by player:
Player values: lineup-based ratings from play-by-play. Every stretch of a game with the same ten players on the floor (about 40,000 a season) is solved together, so each player’s offense and defense is separated from his teammates’ and opponents’. Box-score production (NBA’s PIE) is added on top.
Minutes: each team’s 240 minutes a night go to its best available players first, by player value, each up to his usual minutes, counting the games they tend to miss. A player who missed most of last season injured counts from the season before (marked Back), aged for the time he missed, and is expected to play fewer games than before: since 2017, players coming back from a season-long injury have played about 63% of games, not the 80% their earlier seasons suggested. A player in his first season back from a torn Achilles also gets fewer minutes a night and less offense until he has played again: nine players coming back from one since 2017 played 14% fewer minutes a night than before (healthy players their age played 4% more), and their lineup value fell about 0.6 points per 100 possessions more than those players’. This season that means Damian Lillard and Tyrese Haliburton.
Chemistry: teams whose core has stayed together tend to beat the rest of the projection a little, so it counts how many straight seasons each rotation player has been a regular on his team (up to five), weighted by his minutes. The defending champion also gets one extra win when players from its title team are back. That number is a judgment call: fitted on the last nine champions it came out to about three wins for this year's Knicks.
Age: players tend to improve until about 26 and decline after, faster as they get older. The curve is learned from more than 3,000 season-to-season changes in players’ value.
Once games start, every result moves the rating, recent games count more, and the preseason estimate fades out.
The odds come from 10,000 simulations of the season. Each one draws every team’s true strength from its rating and how sure the model is about it, plays the rest of an 82-game NBA schedule with home court, then the play-in and best-of-seven playoffs. Games already played count as they finished.
How accurate is it? It was tested on every season since 2017-18, using only what was known at the time. Before the season, half of its projections landed within 6 wins of a team’s final record. Game by game, it picked the winner 65.5% of the time. And its percentages hold up: teams it gives a 75% chance win about 75% of the time.
Rosters are NBA.com’s current 2026-27 rosters.