Hoop Informatics
Analyzing rookies in Summer League can be a challenge. Some are 19 years old playing against older players with NBA experience, guys who already know what they can do on an NBA court, while the rookies are still testing the waters. Then there is the fact that Summer League has a stronger focus on collecting information than winning games, challenging players to focus on something that may not be a strength so coaches can get film they can use to teach, or just see what they have.
Which brings us to the stats. Surely we should throw all Summer League stats out the window, right? What can be gained from such few games in such a chaotic situation?
Luckily, someone has already done the research. Owen Phillips, now a Senior Basketball Analyst for the Denver Nuggets, wrote "Summer League Stats You Can Trust" last summer. The question is whether what a player does in Vegas tells you anything about what he does in the regular season. In his research, Owen looked at various Summer League stats since 2008 to find which ones held up. As Owen now works for an NBA team, I wanted to continue the tradition with a 2026 version.
The Research
Owen sorted the results of his research into three buckets: Sticky Stats, Other Stats, and Icky Stats. The measure is R², which runs from zero to one. A one would mean a player's Summer League number in a category perfectly predicts his rookie season number in that same category. A zero would mean there is no relationship at all. The higher the R², the more likely that number carries into the regular season. The following visual is from Owen's article last year.
If you are looking at +/-, or how well a player shot the ball in terms of efficiency, just stop. Those stats are close to meaningless. I've seen various research that suggested it takes 250 to 750 three point attempts for 3P% to stabilize and become meaningful.
I can find a five game stretch last season where Steph Curry shot 29.6% (16/54) on three pointers. I can find another five game stretch later in the season where he shot 32.7% (18/55). Nobody looks at those numbers and questions Curry's shooting ability. The sample gets dismissed because we already know the answer. With a rookie, we might not know the answer, so the same size sample gets treated as evidence.
It runs the other direction too. I can find a five game sample where Dyson Daniels, a career 30% shooter, hit 46% from three (6/13). Putting the ball in the basket is volatile, and how efficiently a rookie did it in Summer League is mostly noise you shouldn't spend time thinking about.
How often a player did or did not get threes up was the most predictive stat of regular season performance. A player who takes a lot of threes in Summer League will likely do the same in the regular season. A player who did not likely won't provide much spacing his rookie season. This, along with assist rate, block rate, and rebound rate, gives us the best read on how a rookie will perform for your team next season.
Three Point Shooting
The table below is sortable if you would like to click a column header to reorder it.
First Round Picks Who Are Going to Shoot It
Mikel Brown Jr. (BKN) | 10.8 3PA/36, 26 attempts in 87 minutes
Caleb Wilson (CHI) | 10.1 3PA/36, 31 attempts in 111 minutes
Cameron Carr (LAL) | 9.8 3PA/36, 48 attempts in 177 minutes
Christian Anderson (CHA) | 8.9 3PA/36, 33 attempts in 134 minutes
Bennett Stirtz (OKC) | 7.9 3PA/36, 30 attempts in 137 minutes
Alex Karaban (SAC) | 6.5 3PA/36, 24 attempts in 132 minutes
Zuby Ejiofor (ATL) | 5.9 3PA/36, 17 attempts in 104 minutes
First Round Picks Who Are Not Going to Space the Floor
Tarris Reed Jr. (SAS) | 0.9 3PA/36, 3 attempts in 121 minutes
Koa Peat (PHX) | 0.7 3PA/36, 2 attempts in 109 minutes
Hannes Steinbach (CHA) | 0.6 3PA/36, 2 attempts in 122 minutes
Aday Mara (OKC) | 0.6 3PA/36, 2 attempts in 120 minutes
The biggest three point revelation in Summer League was the fourth overall pick, Caleb Wilson, who shot 26% from three on only 27 attempts across 24 college games. In Summer League he shot 42% on 31 attempts in four games. He's obviously what you get if Steph Curry and Kevin Garnett have a kid, right?
As we now know, we can mostly ignore the efficiency as noise. Making them is better than not, sure. But the sample is so small it's not enough to say he'll be a plus shooter as a rookie. He may become a good shooter. He may become a bad one, or just an okay one. We don't know. The number that matters is how many he got up. Wilson went from roughly one attempt per game in college to 10.1 per 36 minutes in Vegas, and that willingness, not the percentage, is the best signal for Bulls fans.
It was also nice to see that he was doing it not only on catch and shoot, but pull up threes as well.
On the other side of things, Isaiah Evans of the Minnesota Timberwolves shot just 20.5% from three in Summer League, but that number doesn't carry much weight. In college he hit 38% on 429 attempts and shot 85% at the line, which is a strong indicator of outside shooting ability. At this point, that tells us far more about his shooting than a few Summer League games do.
The same goes for Christian Anderson. Charlotte Hornets fans shouldn't worry too much about the 21.2% he shot on 33 attempts in Summer League. Put more weight on the 40% he shot on 447 attempts in college. That's the better indicator. The more important thing is that both players kept taking a lot of threes.
Hannes Steinbach is a different case. During his draft availability, he said he wants to improve his shooting to help space the floor. He may get there someday, but his two attempts in Summer League suggest he won't stretch defenses as a rookie.
Takeaway: focus less on how many shots went in, and more on which shots a player was taking.
More Sticky Stats
The stats worth trusting describe how a player plays, not how well he plays. Passing, rebounding, and rim protection are habits. A rookie who looks for the extra pass in Vegas is going to look for it in November. A rookie who chases offensive boards against a Summer League roster will chase them against a real one. Blocks work the same way, because the instinct to contest at the rim doesn't switch off when the games start counting.
The interactive table below shows how this year's rookies performed in those categories.
Chris Cenac Jr. really stands out here. He has the highest block rate and the 10th highest offensive rebound rate. No other rookie ranked that well in both. I may need to go back and watch more Boston Celtics Summer League games.
The Celtics also have Dillon Mitchell, who ranked second in offensive rebounds per 36 and 11th in block rate. Both should translate to the regular season.
AJ Dybantsa did not have a single offensive rebound in his 51 minutes of play. This Wizards team is going to be interesting. Keaton Wagler posted good offensive rebounding numbers at Illinois but did not do the same this summer for the Clippers. Nate Ament finished with the 11th worst offensive rebound rate and the 19th worst block rate among rookies, not great for someone that is 6'10".
Three players really stood out with over nine assists per 36 minutes: Sergio de Larrea, Kingston Flemings, and Braden Smith.
A big question coming out of the draft was Darryn Peterson's playmaking. At Kansas he averaged a measly 2 assists per 36 and didn't play point guard. Some pointed to his high school tape and argued Kansas simply didn't use him that way. Others took the college numbers at face value, that a player who didn't create there wasn't going to start creating in the NBA.
Assists per 36 is one of the stickier stats we have, and Peterson finished fifth among rookies at 6.7. The handle was shaky and he turned it over a lot, but the question of whether he can run some point got closer to an answer this summer.
Takeaway: a rookie's habits tend to show up in NBA Summer League
Full Summer League Stat Heat Map
I decided to make a visual of the first and second round picks, with each stat ordered from most meaningful to least meaningful. The heat map runs lighter for better and darker for worse, so you can see how rookies performed in various stats while also knowing how much weight to give each one based on whether it sits on the left or right side of the graphic.
One thing to watch for is when a column is mostly purple with one bright cell, it means a single player separated from the field by enough to compress the scale for everyone else. Brayden Burries and his 5.80 assist to turnover ratio is the clearest example. He's the only yellow in that column because nobody else is close.
It works the other direction too. A mostly bright column with one dark cell means someone finished far enough below the field to stretch the scale on his own. I'm looking at you, Dailyn Swain, and that 25% true shooting. Lucky for Swain, that's an icky stat that tells us almost nothing about his rookie season. Less lucky is the rest of his row, where the columns that do matter aren't much brighter.
So if you're panicking because your rookie couldn't hit a shot in Vegas, relax. And if you're bragging because yours shot amazing from three, temper your expectations.
Summer League is chaotic and it's hard to take away much. But it can show us a rookie's habits. It may not matter as much how well a rookie is playing versus how they're playing. Descriptive stats like offensive rebounds, blocks, assists, and three point attempt rate can help tell the story of what to expect as we project a rookie into their first regular season.
Notes
I used NBAdraft.app for the Summer League stats in this article.
The interactive tables were made with datawrapper.de.
And again, thank you to Owen Phillips, who did the original research in his "Summer League Stats You Can Trust" piece last year. Without his work, this article would not have been possible. The Denver Nuggets got a good one. One of my favorites of his is this research on whether the Kornet Kontest actually works.