The NBA Win More or Less Test
A pseudo-scientific look at who will win more or win less in 2025-26 than they did in 2024-25
One of the best barometers of a season’s success is the good old-fashioned, “Did you win more?” test. Unfortunately, people, especially the 21st-century variety, are extremely impatient and want to know right now who will win more games. Since AI is still months away from pretending it can predict the future, we’re left to wonder, imagine, and dream over NBA win totals. That is, unless you have a super-pseudo-scientific-extra-terrific prognostication equation, which I do.
My NBA “Win More or Less?” test is a modified version of the Bill James’ indicators. In the baseball version, the six indicators are Pythagorean record (point differential), the plexiglass principle (more on it below), the law of competitive balance (teams tend to drift towards .500), age, AAA performance, and late-season performance. The plexiglass principle, at its simplest, is that teams who improve in one season are likely to decline in the next, and vice versa. The bigger the swing, the more likely it is to happen.
I find the first four indicators to have obvious parallels to basketball, but I’ve excluded AAA performance and late-season performance. The G-League isn’t a reliable pipeline of impact talent like MiLB, and late-season performance in the NBA is significantly impacted by schedule and jockeying for playoff seeding or lottery odds.
To replace the discarded indicators, I added opponent 3-point efficiency, top-five player availability, and offseason player flow. Opponent 3-point shooting has a dramatic impact on team performance, but is largely outside the control of a team, unless you’re the Celtics. Teams that underperform or overperform in this regard, relative to the league average, should experience some form of regression to the mean and see their record dragged along for the ride. I used a blend of opponent 3-point percentage and corner 3-point percentage.
Unlike in baseball, one or two players can make or break an NBA season. To capture this, I wanted to find which teams had their best players the least and most available, and who potentially added or lost the most talent in the offseason. Remember, these are meant to be relatively binary distinctions, which means there’s no weight being placed on player quality, outside of my arbitrary distinction of placing them within a team’s top five. I’m not trying to predict how many wins a team will add or lose, just what direction they’re headed.
To gauge top-five player availability, I took the total games played by their top five players and weighed them against the league average, which was 308 games or 61.6 games per player. As a note, for teams that swapped high-level players, I combined their games played for either team. Instead of saying the Lakers only got 28 games of Luka Doncic and 42 games of Anthony Davis, I counted it as 70 games played from Lukony Davic.
For offseason player flow, I only counted players with three or more years of NBA service (with a small exception for veteran foreign players) who averaged at least 10 minutes per game in the prior season as additions, and players who appeared in over 50% of their team’s games in the prior season as departures. This eliminates rookies, as they are generally ineffective players, and does a decent job of capturing who added the most NBA-ready talent for the upcoming season.
After collecting all seven indicators for all 30 teams, I found their averages and standard deviations, calculated Z-scores, and then added them together for one simple figure. Each category is weighted equally, which they almost certainly should not be, but I wanted to keep this simple.
To make sure my final results weren’t completely off the wall, I took the difference between teams’ Vegas over/under win totals for this upcoming season compared to their prior season win total, and then found the linear correlation between my simple Z-score method and their projected swing in wins. The correlation came in at 0.658, but the primary disagreements, as you’ll see, stem from variables outside of the indicators’ purview.
The 13 Most Likely Teams to Improve Their Win Total
You might be wondering, “Why 13?” And the answer is simple: that’s how many teams, based on my seven indicators, project to improve their win total. What’s immediately evident is that really bad teams are much more likely to improve. This makes sense intuitively. Most bad teams are young and have experienced some level of unrepeatable misfortune. Plus, they’re guaranteed to get a strong boost in the competitive balance category.
Out of this lucky group of 13, the most interesting team, in my opinion, is the Denver Nuggets. The Nuggets are an older 50-win team coming off a season where they were comparably healthy. That sounds like a recipe for decline, but their massive influx of new players, on its own, pushes them into the black. We’ll see if their offseason additions of Jonas Valanciunas, Cameron Johnson, Kessler Edwards, Bruce Brown, and Tim Hardaway Jr. offset the losses of Michael Porter Jr and Russell Westbrook. But the general bullishness over their offseason isn’t without warrant.
The Brooklyn Nets are a good example of the blind spots of the indicators. Vegas has them pegged for 21.5 wins, down from 26, and a big reason for that is their current organizational position. The Nets are in the middle of a rebuild, but owe their 2027 pick to the Houston Rockets, via swap rights. Since this season represents their last best chance, at least in the short to medium term, to land a top-five pick, there’s an overwhelming confidence that they will actively try to lose as many games as possible. That being said, there are enough positive indicators here that it wouldn’t surprise me if the Nets had to go the extra mile to hit the under.
A few more highlights from this group are the Hawks and the Magic. The Hawks might have outperformed their point differential and improved their win total from a season prior, but their combination of age, offseason additions, injuries, and horrid opponent 3-point shooting luck should see them easily improve on last season’s performance. The Magic aren’t nearly as high as they ought to be, and it’s all down to how I judge player flow. They have a severely negative player flow z-score because they lost five players who played in over half of their games. However, I’d rather have Desmond Bane and Tyus Jones than Kentavious Caldwell-Pope, Cole Anthony, Caleb Houstan, Cory Joseph, and Gary Harris.
The 17 Most Likely Teams to Lose More
Unsurprisingly, the three teams that won over 60 games are projected to win fewer games. The Cavaliers, outside of marginally unlucky opponent 3-point shooting variance, experienced a charmed season. They won 64 games, a massive improvement on the prior season, are a bit older, were healthy, and actually lost more players than they added. They’re still the favorites in the East, but winning 60 games in consecutive seasons is rare for a reason.
The Celtics and Pacers are expected to decline, and that’s without any direct input regarding the health of Tyrese Haliburton and Jayson Tatum. However, both teams behaved like they were going to take a step back and let talent flow out the door. In either case, even with healthy stars, it was always going to be an uphill battle to improve on their win total.
If you pay attention to net rating or point differential, seeing the Lakers rank so highly (lowly?) shouldn’t come as a surprise. They outperformed their point differential by a league-high six games, are an older team, had about average health, and benefited from fortunate opponent 3-point shooting variance. Now, I’m not going to say that it’s impossible for them to win 51 games, but it’s very unlikely, and Vegas agrees, setting their over/under at 47.5. This may be pseudo-science, but at least it isn’t infected with Lakers exceptionalism.
Since I’m more of a basketball sicko than is probably healthy, seeing the Portland Trailblazers as the team with the fifth-highest odds of slipping in the win column is incredibly interesting to me. The only thing going in their favor is that they were a younger team that finished below .500. Every other indicator suggests they should win fewer than the 36 games they managed last season. I was highly critical of their trade for Jrue Holiday, and this is exactly why. This isn’t a team that’s poised to take a real step forward, and even if they do, 39 wins is not worth the $104.4 million left on Holiday’s contract.
Vegas vs the Lucky Seven Indicators
I am personally against gambling. Not for any religious reasons, but purely for financial and economic ones. Gambling provides next to no societal value, is highly detrimental, and siphons wealth from the poor to the rich. That being said, Vegas’ over/under win totals are a useful tool. When millions of dollars are on the line, people don’t do fuck all, unless you’re Kawhi Leonard.
The biggest gaps between my indicators and Vegas win movement are a matter of player talent. The Rockets added Kevin Durant. That makes them better. Vegas knows that. My indicators don’t. The Spurs are another example of this. Since they added De’Aaron Fox last season, he didn’t count as a new addition, but he also played 17 games for them last season. His lack of playing time and the fact that he didn’t start the season on the team led me to not include him in their top-five availability. He represents a massive talent upgrade, but that’s not being captured.
Now, if you want gambling advice, the team that I’m most confident Vegas is missing on and my indicators are nailing are the Heat. Their over/under is 37.5 wins. They won 37 games last season, underperformed their point differential more than any team in the league, and added Norman Powell and Simone Fontecchio for the cost of Haywood Highsmith, Duncan Robinson, and Alec Burks. They also won’t have to deal with Jimmy Butler loudly quitting, and probably won’t have the most catastrophic clutch season in league history. In 42 games across 195 minutes, the Heat had a net rating of -16.2 in clutch situations, with a league-worst offensive rating of 95.6, which was over five full points worse than the Toronto Raptors in 29th.
While Bill James crunched decades of data to build his indicators, I simply stole some of his work and added a few unconfirmed data points. There are far more sophisticated projection systems that I’d readily endorse over this one, but I think having easily understood and largely transparent methodologies has its benefits, even if they’re not the most rigorous. In all honesty, the reason a team will improve or decline is far more interesting than whether it happens at all. So, when the Charlotte Hornets eclipse 20 wins this season, you can say you saw it coming from a mile away.
For any inquiries about work, discussion, and the like, you can email me at nevin.l.brown@gmail.com.