Behind the Scenes: How the 2026-27 NBA Preseason Rankings Were Built
Giannis Antetokounmpo plays in Miami now. LeBron James plays in Philadelphia, and Jaylen Brown and Paul George changed cities in the same trade. So our opening NBA board for 2026-27 ended up moving 25 of the 30 teams, and every one of those moves came out of a trade, a signing or an injury. Below is the whole board, what we did to each team, and then further down, how we work these ratings out in the first place.
These are preseason rankings in the sense that they come before the regular season starts. Nothing that happens in an exhibition game goes into them.
What We Did
We start with where all 30 teams finished last season, playoffs included. But we don't carry that forward whole. Our own data says only about two thirds of a season survives into the next one, so every team gets pulled roughly 70 percent of the way back toward the league average. Then we go team by team through the summer and through February's trade deadline and move the ones that changed. Every adjustment is tied to a specific trade, signing or injury, and the whole set has to add to zero, since thirty teams in a closed league obviously can't all get better at once.
So fourteen teams went up, eleven came down, and five stayed where the math left them. Indiana moved the most at five and a half points. The Clippers moved the furthest the other way at four.
The 2026-27 NBA Week 0 Rankings
This is the whole board going into opening night on October 20. Each team shows its league rank on offense and on defense, its projected wins, and in orange, how far we moved its point spread. This list is frozen at today's date and won't be updated, so the live NBA rankings will start moving away from it as soon as games are played.
A couple of things to know before you read it. Projected wins cover the 80 games currently on the schedule rather than 82, because the NBA Cup knockout games haven't been given dates yet. And the projections don't track the ranking exactly, since the rank comes from a team's overall rating while the projection comes from its offensive and defensive ratings run against its real opponents. Detroit ranks fifth and projects more wins than Philadelphia at fourth.
Oklahoma City Thunder
Northwest Division · 75-21 last season
San Antonio Spurs
Southwest Division · 75-31 last season
New York Knicks
Atlantic Division · 70-32 last season
Philadelphia 76ers
Atlantic Division · 50-44 last season
Detroit Pistons
Central Division · 67-29 last season
Denver Nuggets
Northwest Division · 56-32 last season
LA Lakers Lakers
Pacific Division · 56-35 last season
Minnesota Timberwolves
Northwest Division · 55-39 last season
Boston Celtics
Atlantic Division · 59-30 last season
Houston Rockets
Southwest Division · 54-34 last season
Miami Heat
Southeast Division · 43-40 last season
Cleveland Cavaliers
Central Division · 60-40 last season
Toronto Raptors
Atlantic Division · 49-40 last season
Atlanta Hawks
Southeast Division · 48-40 last season
Orlando Magic
Southeast Division · 49-42 last season
Portland Trailblazers
Northwest Division · 43-44 last season
Phoenix Suns
Pacific Division · 46-42 last season
Indiana Pacers
Central Division · 19-63 last season
Golden State Warriors
Pacific Division · 38-46 last season
Charlotte Hornets
Southeast Division · 45-39 last season
Chicago Bulls
Central Division · 31-51 last season
Washington Wizards
Southeast Division · 17-65 last season
Dallas Mavericks
Southwest Division · 26-55 last season
LA Clippers Clippers
Pacific Division · 42-41 last season
New Orleans Pelicans
Southwest Division · 26-56 last season
Utah Jazz
Northwest Division · 22-59 last season
Milwaukee Bucks
Central Division · 31-50 last season
Brooklyn Nets
Atlantic Division · 20-62 last season
Sacramento Kings
Pacific Division · 22-60 last season
Memphis Grizzlies
Southwest Division · 25-57 last season
The Biggest Movers
Philadelphia, up ten places to 4th. Jaylen Brown and LeBron James in, Paul George and Quentin Grimes out. That's a large net gain in talent even after we discounted James heavily, which we did, because he turns 42 in December and his minutes are going to be managed. So we moved them up three points.
Indiana, up eleven places to 18th, and that five and a half points is the largest single adjustment on the whole board. I'll get to why in the next section. But they're also the team I would argue about first, because an Achilles at that position is genuinely hard to price.
Washington, up eight places to 22nd. They're still a bottom-ten team on our board, so going from the worst record in basketball to 22nd is a big move in points and a pretty modest one in standing.
Miami, up six places to 11th. They had the league's best offensive rating last season and they've now added a two-time MVP to it. But Tyler Herro, Kel'el Ware and Jaime Jaquez Jr. all went the other way and the backcourt looks thin, so we only moved them up two points. That's less than Antetokounmpo is worth on a blank page, and the reason is that he's missed meaningful time in each of the last few seasons and he turns 32 in December.
The Clippers, down five places to 24th, which is the largest downward move at four points. Kawhi Leonard went to Toronto, John Collins to Detroit, and Bogdan Bogdanovic is gone, while Brandon Ingram, Max Strus and Rui Hachimura came back the other way. They also traded Ivica Zubac away in February, so part of this is already sitting in last season's 42-41.
Memphis, down five places to 30th. They lost Jaren Jackson Jr. in February and Ja Morant in the summer, and Jerami Grant and Isaiah Stewart came in. We took them down a point and a half, which puts them last, and I expect to hear about it.
Milwaukee came down five places to 27th for the obvious reason, and Charlotte down four to 20th after trading LaMelo Ball and Miles Bridges. Our ratings had actually liked Charlotte rather more than their 45-39 suggested, so they had further to fall than the record implies.
The Four Teams Whose Record Lied
The hardest part of this wasn't the summer at all, it was the trade deadline back in February.
A trade at the deadline barely registers in an 82-game record. The player arrives for twenty-odd games, or he gets hurt and arrives for three, and the win total still mostly describes the roster that played the first four months. Our ratings read the record, and they have no way of knowing that a team is about to be different. Four teams were wrong for exactly that reason, and three of them were wrong by a lot.
Indiana finished 19-63. But this is a team that went to Game 7 of the 2025 Finals, and Tyrese Haliburton tore his Achilles in that game and missed the entire season that followed. So 19-63 is really the Pacers without their point guard, and he's healthy now, sixteen months out. They also picked up Ivica Zubac at the deadline and kept Pascal Siakam, Aaron Nesmith and T.J. McConnell. We moved them up five and a half points, and that figure isn't a guess: it's the gap between where they finished in 2025 with Haliburton and where last season left them.
Washington finished 17-65, the worst record in the league. Trae Young arrived from Atlanta on January 7 and played five games. Anthony Davis arrived from Dallas on February 4 and never played a game for them at all. So add Deandre Ayton, Khris Middleton and the first pick in the draft to those two and this is a completely different roster on paper. We moved them up four points and held it there rather than pushing further, because Davis and Middleton both carry long injury histories.
Utah finished 22-59. Jaren Jackson Jr. arrived from Memphis on February 3 and played three games before knee surgery, and he's back now alongside Lauri Markkanen and a very young group that includes Darryn Peterson at number two in the draft. But they also lost Walker Kessler to the Lakers, which cuts the other way, so the net on Utah is only up a point.
Dallas cuts against the pattern. Kyrie Irving sat out all of last season with a knee injury and then a calf setback, so his return reads like a clean gain. But Anthony Davis was traded out of Dallas at that same February deadline after playing the first half of the year. One arrives and one leaves, which is why Dallas is up a point rather than three.
So we now read the deadline just as carefully as we read the summer. Offseason transaction lists don't show February moves at all, and that's exactly how a team like Washington ends up mispriced by four points.
How Much of a Season Carries Over
If a team finishes a season a certain distance above average, how much of that distance is still there in October?
We answered that twice, two different ways, using our own stored ratings going back to 2007. First we lined each season's finishing ratings up against the next season's finishing ratings, one transition at a time, across thirteen usable season pairs. The lockout year and the two seasons the pandemic distorted came out of it. Then we took the first couple of hundred games of each new season, predicted every one of them off the previous season's final ratings, and measured how much of that predicted margin actually showed up on the scoreboard. Both came back around two thirds.
So each team keeps about 70 percent of its distance from the league average and gives up the rest. Oklahoma City won 75 games last season and still opens on top of the board. But they don't open as a 75-win team, because almost nobody repeats a year like that.
Carrying a finished season forward whole is the obvious alternative, and it's worse. We scored both against what actually happened over the opening stretch of every usable season, and the full-strength version misses by more. It calls games too confidently in October, which is exactly when it knows the least.
And this number belongs to basketball specifically. We ran the same two measurements for the NHL a few weeks ago and got barely half, because hockey forgets a season faster than basketball does. So the hockey figure would be badly wrong here, and ours would be badly wrong there.
One other thing came out of the margin measurement. It also reports how much the home team is worth without being told anything about it, and it came back between 2.2 and 2.4 points. We had already moved our NBA home edge to 2.2 points back in September after measuring it a completely different way, so two methods landing in the same place is worth a lot more to me than either one on its own.
How We Read the Offseason
Shrinking everybody toward the average handles the general case, but it obviously can't handle Giannis changing teams. So the second half of the job is reading what happened since June, team by team, and moving the ones that changed. We make those adjustments in points of point spread, which means moving a team five points moves the spread in every game it plays by five points.
This was a heavy summer even by NBA standards. Antetokounmpo and Bobby Portis went to Miami for Tyler Herro and a young package. LeBron James signed in Philadelphia for his 24th season, and Boston sent Jaylen Brown to those same 76ers and took Paul George back. LaMelo Ball went to Minnesota, Ja Morant to Portland, Kawhi Leonard to Toronto, Walker Kessler to the Lakers, and DeMar DeRozan to Denver.
Every adjustment is tied to a specific move, and the whole set is forced to sum to zero, since thirty teams in a closed league can't all improve at once. If somebody gained, somebody lost.
Five teams got nothing at all. Golden State ran it back, re-signing Stephen Curry, Draymond Green, Al Horford and Kristaps Porzingis without adding anyone meaningful. San Antonio added Tobias Harris and extended Victor Wembanyama. New Orleans brought back essentially the roster that went 26-56, and Houston swapped Dorian Finney-Smith for Marcus Smart and Bogdan Bogdanovic, which is close to a wash. The Lakers are the interesting one, because they lost LeBron James, Deandre Ayton, Rui Hachimura and Marcus Smart while adding Kessler, Quentin Grimes, Collin Sexton, Kevon Looney and Matisse Thybulle, and all that movement roughly cancels itself out.
And a team can move in the rankings without us touching it at all. Houston dropped a place and New Orleans dropped two, and neither one got an adjustment. Rankings are relative, so when Philadelphia jumps ten spots somebody standing near them in line has to come down. Shrinking on its own never reorders anybody, because if you pull every team the same fraction toward the average then the order comes out identical with the gaps squeezed. So every rank change between last season's final board and this one traces back to an offseason adjustment somewhere.
What Week 0 Is Actually For
These numbers aren't built to last. They're built to get out of the way.
Once games start counting, the model solves each team's rating from results. But two games into a season the results alone say very little, and a 2-0 team is not the best team in basketball. So the preseason rating acts as an anchor, and it weakens every week on a fixed schedule. By week three it carries half the weight it started with, and by late November it's gone entirely and the ratings are built purely from games that have been played.
That makes the job of this board a narrow one. It has to be good enough that our predictions in the first three weeks are worth reading, and then it has to disappear. So judge it on the opening month and not on April.
The NBA rankings update daily once the season starts, and the game predictions built off them update with every run. Projections may change as teams play.
Frequently Asked Questions
What are NBA preseason power rankings?
They are team ratings published before any regular season game has been played, which is why we label ours Week 0. Ours start from where all 30 teams finished the previous season, then pull every team part of the way back toward the league average, and then adjust individual teams for what changed over the offseason: trades, free agency, the draft, and injuries or returns from injury. Exhibition results are not used at all.
How much does one NBA season predict the next?
About two thirds of a team's distance from the league average carries forward. We measured it two independent ways across thirteen season pairs of our own ratings and both came back near that figure, so we keep roughly 70 percent and give up the rest. And it is specific to basketball, because the same measurement for the NHL gives barely half, so the two are not interchangeable.
Which NBA teams improved the most this offseason?
By our ratings, Indiana, Washington and Philadelphia. Indiana gets Tyrese Haliburton back after he missed the entire season with a torn Achilles. Washington's 17-65 record contains five games of Trae Young and none of Anthony Davis, both of whom were acquired midseason, and they have since added Deandre Ayton, Khris Middleton and the first pick in the draft. Philadelphia added Jaylen Brown and LeBron James while losing Paul George.
Who is the best NBA team going into 2026-27?
Oklahoma City, narrowly over San Antonio, with New York third. The Thunder and the Spurs both won exactly 75 games last season counting the playoffs, and both of them come down from those heights in our preseason ratings because a season that good is almost never repeated. The gap between the two of them at the top is small enough that it may not survive the first month.
Are NBA power rankings free?
Ours are free, and that covers team ratings, rankings and team summaries across every sport we cover. Game predictions and the game simulator are part of a Versus premium membership.
Please note that some links in this article may be sponsored or contain affiliate content.
Steve Pugh, Founder and CEO, Compughter Technologies, LLC
Steve Pugh, the founder of Versus Sports Simulator, is passionate about the world of sports analytics and statistical modeling. With a deep-rooted love for math and sports, and a commitment to innovation, he's dedicated to equipping sports enthusiasts with useful tools and insights to help them make informed betting decisions.
