LAS VEGAS — Let’s rank my top 30 NBA prospects after watching them up close at NBA summer league.

This is essentially my updated Big Board, based on career projections for each player. This format is inspired partly by a Nik Oza column last year updating his priors after summer league.

As a reference point, here’s my final pre-draft Big Board along with more complete breakdowns of every top prospect.

And here are my new rankings after NBA summer league:

Read more

LAS VEGAS — Back at NBA summer league, where I stayed through Sunday night’s championship game, watching more of the 2026-27 rookie class up close.

You can find Part 1 here, including the top prospects:

Here’s what impressed me about this next group of prospects — and what left me with doubts:

Read more

LAS VEGAS — I’m here for the NBA summer league, catching the top prospects up close.

Here is what I’m excited about when it comes to this rookie class, and what’s concerned me most:

Read more

Earlier this week, I covered the 13 NBA deals I liked most — the steals of the summer.

I’m happy to report there are fewer deals I didn’t care for. But wow, did a handful of NBA teams make some regrettable decisions.

Here are seven of them, in order from worst to … just puzzling.

Read more

The piece stacks up numbers like these: In Brown's last four seasons in Boston, the Celtics went 47-10 (.825) without him — the equivalent of a 68-win team.

That’s what we do: careful, data-driven work that follows the evidence, even when it leads somewhere unpopular.

Jeremias’ analytics work lives here on 5x5, for paid subscribers — upgrade to get it all, at a special low rate of $30 for the first year.

Upgrade for $30

Thanks for reading!

With NBA superstar salaries soaring above $50 million and even $60 million per season, getting good value is more important than ever. (That’s one reason the Celtics traded Jaylen Brown.)

So I want to highlight eight teams — including contenders — that did amazing work this offseason acquiring players who can make a real difference at a great price.

Here are the 13 deals I’ve loved most:

Read more

The Lakers have undertaken a major makeover this offseason, getting a lot younger in the process.

So, should we expect an improved Lakers team this season?

The numbers are skeptical. Here’s why.

Read more

Jaylen Brown has finally been dealt — he’s going to the 76ers in exchange for 36-year-old Paul George and draft picks.

Did the Celtics gift the Sixers an NBA Finals trip, as Kendrick Perkins suggested on ESPN?

I have been critical of Brown’s performance and his contract:

On the other hand, the 76ers gave up less than the reported asking price of four first-round picks, and moved George’s giant salary.

Given the lower-than-expected price, is Perk right? Did the Sixers win this trade? Did the Celtics blow it?

I say no.

Here’s why.

Read more

Jaylen Brown is a divisive player: His supporters point to a Finals MVP trophy, deep playoff runs, and generally strong performance by the Celtics over the years.

On the other hand, his impact numbers are ugly, at least by the standard Brown is held up to, given his salary and status as a five-time All-Star.

But exactly which numbers are ugly, and why are they so bad?

As his name continues to circulate in trade talks and a debate roars on about how good Brown really is, let’s try to answer those questions.

Read more

Let’s examine the Giannis Antetokounmpo blockbuster from both sides, plus the fallout in Boston — and where things go from here.

Read more

Seven teams delighted, annoyed, and surprised me on Tuesday, based on my NBA Draft Big Board.

Let’s dig in.

Worst

Read more

Most NBA teams have very specific needs, given their current depth chart, long-term contracts and future title aspirations.

Let’s look at the players who (a) should be available and (b) best fit the needs of every team in the first round of this week’s NBA draft. That includes a Plan B and even Plan C for many teams.

For my Final NBA Draft Big Board, in which I rank the top 30 prospects according to stats and film work, click here.

Read more

Here’s how I see the top prospects in the 2026 NBA draft, based on advanced stats and film study — ranked from 1 to 30:

Read more

If you believe the analysts, the 2026 NBA draft is a strong one.

There is little doubt, for instance, that this year’s top 3 picks would’ve supplanted the entire top 3 in the 2024 draft.

But it doesn’t stop there: The entire first round grades out stronger than usual — the statistical evidence backs that up.

That said, this class really shines on just one end of the floor. And even there, the numbers might be overselling it.

Let’s see what the draft model says — and a reason or two it might be overrating this class.

Sky-high projections

Read more

In the stats world, measures of uncertainty are known as standard errors and confidence intervals.

Why would an analyst want these measures? Why would an NBA team?

NBA decision makers can use them to decide whether to swing for the fences in acquiring a player or take a more conservative route. These measures also allow us to calculate the probability that one player is truly more impactful than another.

But while we sometimes see standard errors reported for NBA metrics based purely on lineup data, nobody reports them for metrics that include box-score data.1

While it’s not an easy undertaking, it is possible to compute these standard errors for every modern NBA metric.

Here’s how it’s done:

How it all started

Among modern NBA metrics, there are essentially three main types:

• Lineup metrics, such as Adjusted Plus-Minus (APM and RAPM2).

• Box-score metrics, such as Box Plus-Minus (BPM)

• Hybrids

For the original APM, it’s very easy to come by standard errors and confidence intervals. The method used by APM, known as ordinary least squares regression, spits all that information out together with the estimates.

But APM is too heavily influenced by noise, leading to weak predictive power, so the analytics world moved on to its more accurate cousin: RAPM.

RAPM uses penalization, which makes it more accurate. But it also cost us the ability to calculate standard errors easily.

At the 2011 NESSIS3 at Harvard, a statistical analyst for the Spurs told me that he preferred box-score metrics because we didn’t know the level of uncertainty in RAPM. Of course, what he didn’t recognize is that standard errors also exist in box-score metrics.

But now we’ve come full circle. While we were eventually able to estimate standard errors for RAPM — through my own simulation method, or Justin Jacobs‘ — no one has successfully computed standard errors for box-score metrics.

How box-score metrics work

Here’s a simplified overview of how most box-score metrics create player ratings:

(1) Use multiyear RAPM as the target variable, and (2) use multiyear player stats as independent variables, in order to (3) generate weights (coefficients) for all stats.

So far, the setup might look something like this:4

Then (4) multiply the found weights with more recent player stats, and (5) sum up to arrive at a final rating.

For more, see this article:

The different causes of inaccuracy and uncertainty in box-score metrics

Box-score metrics have uncertainty attached to them implicitly, just like other metrics. But that’s not an obvious fact, so let’s start with a simple example:

1. All the inputted player stats are uncertain

Box-score metrics multiply the player’s stats — such as 3-point percentage — with a weight.

Let’s say a player starts his career by making 5 of his first 10 3-point attempts.

For his 3-point percentage, box-score metrics plug in either 50% or a number that regresses to the league-wide mean.

But we don’t know this player’s true skill, and we’re just making a best guess. The reality is that many possibilities exist at the same time: The player might be a 37% 3-point shooter, but there’s a (slightly lesser) chance that he’s actually a 35% or 39% shooter.

This problem exists across all player statistics.

For example, let’s put Victor Wembanyama in an alternate universe, with a basketball environment exactly like the NBA.

His stats would be different, even if his skills were exactly the same. The numbers would change simply because of random chance.

In other words, if you replayed the entire NBA season, everyone’s stats would be a little different than they are right now, just by random fluctuation.

Generally, we are more certain about the stats of a player who has played more minutes and taken more shot attempts from the various distances. In the example above, the range of possibilities is narrower for someone who made 10-of-20 3s rather than 5-of-10.

Here is a chart displaying the uncertainty around player’s assist rates:

How to determine these estimates and bounds is described here:

2. The model weights are uncertain

Box-score metrics multiply the players’ stats with a weight.

Above, we’ve established that stats have uncertainty attached to them. And the same is true for the weights.

The weights might, for example, imply that an offensive rebound is worth twice as much as a defensive rebound.

But, once again, that represents only a best guess. Instead of a weight of exactly 2.0, it’s very likely the relative weight of an offensive rebound is 1.9 or 2.1.

So, what we need to do is compute standard errors for all of our model’s coefficients. This can be done through a technique called bootstrapping.

We essentially run the model a thousand times over, with the following differences:

• For each run, we choose randomly which players are part of the model, and which ones aren’t. Players can be chosen multiple times.

• From run to run, the model weights will turn out to be slightly different. We can deduce the standard errors of the weights from those differences.

3. The target variable is uncertain

Let’s recall how (almost) all modern box-score metrics generate their weights: A machine learning algorithm finds the best combination of weights so that the resulting box-score player ratings most closely resemble their RAPM ratings.5

But as we noted previously, the RAPM ratings are also uncertain.

In 30-year RAPM, for example, Wembanyama is rated at a +5.2 defensively, with a standard error of 1.1.

Thus, we are 95% certain that Wemby’s true rating exists in a [3.0, 7.4] interval.6

Subscribe now

Putting it all together

As we’ve now seen, every step of the process we laid out earlier contains uncertainties that are usually not accounted for. That’s what we’re here to address.

To then create confidence bounds for our metric, we execute the following Monte Carlo method:

1. For all multiyear stats and every player, we randomly draw new stats, in accordance with the player-specific and stat-specific standard errors we created earlier.

2. We randomly draw RAPM — our target variable — according to the RAPM standard errors.

3. We use bootstrapping to create standard errors for the coefficients.

4. Same as in Step 1, but this time for single-season stats.

We run this setup thousands of times to give us the confidence bounds for our box-score metric.

The results

While the ranking isn’t the primary purpose of the article, here are the top 10 players for the 2025-26 regular season in our box-score metric:7

The top four show no surprises.

Fifth-ranked Ty Jerome averaged a blistering 42/6/12 per 100, with a 62% TS% and 42% 3p%.

But Jerome played only 339 minutes in total. Subsequently, his estimate has the highest standard error attached to it. Note that we have standard errors for the offensive, defensive and total rating.

The lowest standard errors generally belong to players that have played the most minutes and have taken a high number of attempts in stats that denote percentages (FT%, 3p%, etc.).

For this last regular season, we are most certain about Amen Thompson’s, Immanuel Quickley’s and Donte DiVincenzo’s ratings. On the other end of the spectrum, confidence intervals are wide for players such as Walker Kessler and Dejounte Murray.

Future work

For the sake of simplicity, we assumed all of the statistics above follow a normal distribution when creating the appropriate standard errors. At least for some statistics, chances are high that this assumption is incorrect.

There are a number of distributions to choose from, several of which might fit certain stats better than the assumed Gaussian distribution.

In Game 4, the New York Knicks pulled off the largest comeback in NBA Finals history — roaring back to shock the San Antonio Spurs after being down 29 points in the third quarter.

In a related story, Karl-Anthony Towns barely played in the first half due to foul trouble, and then played 18 minutes in the second half.

Towns is now +48 in the NBA Finals when he’s on the court, while the Knicks are -40 when he’s off, prompting this question:

Is he the very best center in the NBA to go against Victor Wembanyama?

In other words, is KAT the Wemby antidote?

Let’s answer those questions, and look at (a) how Towns is doing it and (b) which other centers are best equipped to give Wembanyama trouble.

Read more

The NBA draft is just two weeks away, and it’s about time for teams to make their decisions — on what is considered an especially strong bunch of prospects.

Here is how I would make the top 14 picks:

Read more

The 2026 NBA draft class is a strong one, especially when you look metrics based on box-score stats, such as BPM.

But what does lineup data tell us about this year’s class?

Who are the standouts? Whose rating seems like a big red flag?

Let’s dive in.

Read more

In an alternate timeline, Jalen Brunson is still a Dallas Maverick.

“I really did want to stay in Dallas,” he would later say.

Instead, Brunson is headed to the NBA Finals with the New York Knicks.

From 2019 to 2022, I worked in the Mavs’ front office, providing a unique vantage point on the blunders that cost Dallas a future All-NBA point guard — how the franchise, then led by majority owner Mark Cuban1 and general manager Nico Harrison, fumbled Brunson away.

How exactly did we get to the point where the infamous Luka Dončić trade is in contention to be only the Mavs’ second-most disastrous decision of the past four years?

This is my version of that story — the story of how Jalen Brunson put himself in the position to become one of the greatest Knicks ever while making the Mavs’ blunders more and more painful.

Read more

Last year, Victor Wembanyama visited the Shaolin Temple in China for extensive training.

He was coming off a sophomore NBA season in which he had played just 46 games because of minor injuries followed by a more serious diagnosis of deep vein thrombosis, a blood clot in his right shoulder that ended his season early.

The journey bore fruit. He went on to have his best season of his young career, improving at a pace we rarely see among players already so impactful. This past regular season, Wemby surpassed Nikola Jokić in NBA impact, as measured by xRAPM — a metric that had been dominated by the Serbian for half a decade before the rise of Shai Gilgeous-Alexander and Wemby.

What Wemby did there, and why it worked, is what this piece is about.

As someone who has practiced meditation for more than a decade, I can address these questions with some specificity:

• What are the techniques that helped Wemby’s rapid ascent to the top of the NBA?

• What are the benefits?

• How can other NBA players — or any of us — take the same strides toward peak performance?

Read more