Category: NBA


  • Basketball Epistemology

    Basketball Epistemology

    Basketball’s best-kept secret is the manner in which it exposes the ways humans think analytically; the fluidity of its events and “optimal” thinking handle abstract concepts which are prone to reduction. Namely, during a Knowledge Revolution in the midst of exponential growth, the lack of agreement on what constitutes “knowledge” has complicated the acceptance and interpretations of new data. Thusly, the main conflict of this Revolution stems from a perceived diametric relationship between basketball’s two most prominent sources of knowledge: film and stats, visual and numerical reference points, respectively. The objectively illogical approaches to characterizing basketball knowledge prompted me to approximate the bases for “Basketball Epistemology,” or the theories of knowledge in basketball.

    NB: Theories of knowledge are based on the optimization of certain goals, with the focus of this article being the comprehensive, scientific analysis of “things” that happen on the court (referred to as “events”) and the resulting descriptive and predictive powers which assist in on-court decision-making.

    I. Philosophical Analysis

    The foundation of philosophical analysis in basketball relates to system thinking, the unique interactions of “smaller” systems (e.g. lineup combinations, head coaching and assistant coaching) and how they interact to form “larger” systems (e.g. teams). The nature of basketball makes it so that knowledge relating to events is always conditional, meaning systems are not the products of the sums of their parts, but the products of the interactions of their parts. Given the “goal,” such a foundation would be entirely relevant in player forecastings in which executives will interpret conditional information within the context of a separate system.

    There also exists a definitive absence of justified belief which stems from the intersystemic traits of optimization, meaning there is also an absence of conclusion to intersystemic problems. Similar “abstract” problems can be thought of as any questions which pose answers outside the scopes of qualitative or quantitative measurements, not to be confused with estimates. Examples could form as such:

    • Which player scored the most points during a time frame? (Non-abstract)
    • Which player was the best scorer during a time frame? (Abstract)
    • Which team managed the highest point differential during a time frame? (Non-abstract)
    • Which team had the best defense during a time frame? (Abstract)

    Namely, the inferences drawn from intersystemic analysis, regardless of confidence level, are definitively opinionated

    II. Sources of Knowledge

    “Numerical” and “visual” tools have been named the primary sources of knowledge under these branches, and both come with caveats and signs of caution when interpreting within the context of knowledge. Numerical and visual tools are interrelated and express events in different ways. Let’s start with visual tools, which typically involve an active viewing of a basketball game to identify patterns and trends to then be interpreted within the context of a goal. But unlike a crude, numerical value, visual tools have an incomprehensible scope in which all on-court events are visible to the viewer. But there are inferential limitations: memory and registering.

    Memorizing events which span a significant number of games (hundreds or more?) which characterize five players, five opponents, their actions, movements, body language, speech, coaching, coaching decisions, player decisions, communication between systems, among more… is evidently a fruitless endeavour, and any claim otherwise would be a gross overestimation of human cognition. Therefore, while visual tools are direct representation of reality, the ability to memorize and then register (i.e. interpret optimally within the context of a goal) all events does not exist.

    Variability exists among visual tools as similarly-labeled problems exist among numerical tools. Namely, how one viewer perceives the same film will vary from how another perceives the same film based on their memory and processing skills. (Note-keeping can be a valuable tool in such scenarios.) The result is that observations do not qualify as data outside of its scope (e.g. establishing an abstract or inconclusive cause-and-effect relationship based on an observation) regardless of confidence level. Examples could form as follows:

    • Draymond Green sets a ball screen for Stephen Curry around the three-point line. (Data)
    • Draymond Green setting a ball screen for Stephen Curry around the three-point line was a continuation of the Warriors’ patterned playbook. (Not data)

    Antithetically, it’s popular saying that “numbers don’t lie,” but such a proposition encounters similar problems. Numerical tools are similarly plagued with scopes and thusly can only qualify as data within the context of its scope. Examples could form as follows:

    • Stephen Curry led the NBA in points per possession and True Shooting percentage (as calculated by Basketball-Reference) in the 2015-16 regular season. (Data)
    • Stephen Curry leading the NBA in points per possession and True Shooting percentage in the 2015-16 regular season means he was the NBA’s best scorer during that period. (Not data)

    The overlap in the examples pertaining to visual and numerical tools is how the observations are interpreted within the contexts of their finite scopes, meaning a judgment ascribed to the observation does not qualify as data but rather an opinion, regardless of confidence. As a result, these numerical and visual tools are unreasonable tools to answer questions outside of their scopes. The theme of such connectivity carries over to the relationships between numerical and visual tools.

    As stated earlier, the tools are interrelated and are different mediums which express events. There are direct measurable relationships and abstract inferential relationships between these numerical and visual mediums. Namely, take the examples as follows:

    • Kobe Bryant swishes a two-point jump shot while crossing his legs with three defenders in his immediate vicinity.
    • Luke Kennard is intentionally fouled at the end of a regulation period and banks two free-throw attempts.

    The points column in the box score only registers that both Bryant and Kennard scored two points in those instances, which characterize the direct measurable relationship as both numerical values can be traced back to equal measurements observed visually. However, the inferential relationship between the tools which concerns abstract questions outside the tools’ scopes is not direct. This is another way of saying: “Not all points are created equally.” The principle applies to virtually every instance of cross-referencing between tools, as the contexts in which events occur will seldom possess significant overlap.

    III. Philosophical Skepticism

    The presence of abstract problem-solving relates to how sources of knowledge and questions deemed optimal disqualify the answers as knowledge. Therefore, skepticism is a natural byproduct in the classification of knowledge and the resulting judgments. The aforementioned example of player forecasting which estimates a player’s intersystemic qualities contains a hypothetical component in which the evaluator must estimate the transition based on conditional information. If the overarching question is posed along the lines of:

    • How does Player A in System A raise the point differential in System B in the following season?
    • How does knowledge of Player C in system C against opponent defense D imply changes in knowledge of Player C if he had played against opponent D in System E?

    The resulting abstractness poses threats to the concept of intersystemic knowledge and corroborates the deduction that contrary claims would rely on judgment. Namely, the skepticism of intersystemic knowledge serves to interweave the abstract nature of optimal problem-solving, the finite scopes of perceived sources of knowledge, and the abstract inferential relationships between tools. (Epoché)

    Listen (or watch!) to the companion piece to this post on YouTube!


  • 1954-55 | Retro Player of the Year

    1954-55 | Retro Player of the Year

    This is the first season for which I’ve done the Retro Player of the Year project, but it’s worth noting this isn’t the NBA’s inaugural season. That was actually the 1946-47 season, in which the league was known as the Basketball Association of America (BAA). For a myriad of reasons, but mostly related to the breadth and depth of statistical information, this was the cut-off I chose. Plus, although they’re uncommon, this way I had to deal with fewer instances like the Fort Wayne Pistons on November 22, 1950. I also could’ve chosen 1956-57 as the cut-off due to stabilizing talent distributions, but that felt too arbitrary. For now, let’s stick with the shot clock as successor to NBA “pre-history.”

    The state of the league

    First, let’s clarify a few things about the state of league, so that we’re in a better position to discuss the sport’s original greats in context. We’re still working with a primitive box score. Minutes weren’t tracked for teams. Steals, blocks, and turnovers weren’t tracked for players. Rebounds had also yet to be split between offensive and defensive boards. But otherwise, we’ve got a decent haul of points, shots, assists, rebounds, and fouls to work with.

    The other major sphere of disparity between the classic and modern games is the rulebook. I’ll highlight a few of the major differences, and will likely refer to many of them in later seasons. Here are the ones worth considering now, and how they affect (based on observation and theory) the strategy and style of the early shot clock period:

    • Illegal defense: The NBA outlawed zone defense in the winter following the league’s inaugural season, and wouldn’t overturn this rule for another fifty-four years. This had major ramifications for the spacing of the floor, how defenders operated away from the ball, and how teams strategized to score closer to the basket.
    • No threes: This is an obvious one. While primitive jump shooters etched an extra groove into their legacies with innovation, outside shooting and spacing weren’t as pressing back then as they are now.
    • Twelve-foot lane: The league wouldn’t re-widen the lane to sixteen feet until 1964, so for this nine-year stretch of basketball, there was even more emphasis on post play and crowded paints! Historians, anecdotes, and others seem to attest to this.

    The closest thing I could find to on-court footage from this season was this video, covering what looked like the highlights of the seventh game in that year’s NBA Finals. While it’s no basis on which to establish the tendencies of individual players, we can hopefully make inferences about how the game was played. Let’s turn to the tape:

    Based on this video, I’d infer the sport was barely evolved on the horizontal plane. You can see as early as the second clip (0:21), as the offense runs a traditional pick-and-roll, there’s one defender at the level of the screen. Not only is the screen poorly-set (watch the screener’s back fold when he makes contact with the defender), the defender’s instinct is to jab at the ball-handler (who’s several feet away) without freeing himself. (This isn’t to slander this particular defender, as he does a good job of chasing the opposing guard later in the possession.) We see this repeat in multiple possessions; although, if you’re optimistic, you could explain it as a reluctance to expand your defensive zone. There was no incentive to defend farther away from the basket without a three-point line!

    While the sport was obviously post-oriented, the breadth of attacks wasn’t as robust as it is today. I’d have said the second thing that stood out most to me was the lack of baseline activity. Much of these plays were pick-and-rolls and feed-ins to big men who could score with their backs to the basket. Lots of the swirling cutters—and the defenders they pulled in their paths—were attacking the basket head-on. This could relate to the absence of the corner threat, but it didn’t appear players were regularly exploiting lapses in backline rotations among defenders. The shape of the court in this manner, however, could just as easily be associated with the narrower lane. Passing also wasn’t as advanced of a skill, so it likely wasn’t reasonable to camp out in the dunker’s spot and wait for lobs.

    The most advanced (by both tactics and skill) pass in the video was thrown by Dolph Schayes, a behind-the-back pass out of a stationary post-up to a cutter. Re-affirming the primitive horizontal game was the defender of the other big stationed in the paint, who was inattentive to the cutter’s movement as it unfolded. You could say that’s a major theme of the older game’s spatial dimensions. The court was functionally half the size it is now, so “court coverage” and “roaming” were likely not as valuable, nor were they likely treated as valuable among the league’s defenders. But in the instances an off-ball cutter could attack (even if head-on) open space, both passer and receiver were generating easy points for their offenses.

    How valuable were the best players?

    There’s a pattern in NBA history in which the league’s talent distribution tightens during periods following expansion. For the most part, yearly standard deviations in Simple Rating System (SRS) are between 4 and 5.5 points; in the 1954-55 season, this was a measly 1.5 points. My guess is the likelihood of winning a title (prorated to a thirty-team league) this year to have been nearly twice as difficult than a typical non-expansion season. That means players who were able to separate themselves from the pack were even more impressive relative to season, and their marks in Ring Shares will reflect it! Based on team ratings, the implied spread of offensive or defensive value among players was not only compressed, but also fairly even, meaning we’ve yet to move into the period of disproportionately titanic defensive play.

    This version of Box Plus/Minus (BPM) uses estimators for the statistics that weren’t tracked (such as steals, blocks) as discussed earlier. This way, we could have a loose approximation—more robust than a metric like Win Shares—to understand how players provided value in previous decades. Listed above are the top-10 players in BPM impact per game (context-dependent, sums to team ratings) in the regular season. If you’re curious, the metric’s top offensive player (by far) was Bob Cousy (+3.0) of the Celtics, while the top defensive player was Chuck Share (+4.7) of the Hawks. Neil Johnston (+5.1) led all qualifying players in overall impact.

    Based on the distributions between the regular and postseasons, players weren’t drastically upping their impact in the Playoffs. Listed above are the top-10 players in BPM impact per game (context-dependent, sums to team ratings, relative to all playoff teams, not matchups) in the second season. Boston carried its league-leading offense in the regular season to the postseason, and sported arguably the two-best offensive players in basketball in Bob Cousy (+2.9) and Bill Sharman (+3.3), while the defensive side was dominated by Vern Mikkelsen (+3.4) of the Lakers and Harry Gallatin (+3.2) of the Knicks.

    Who were the standout players?

    For my money, Bob Cousy was the best offensive player this year. Known for his razzle-dazzle style, an innovative combination of dribbling and passing, he was likely basketball’s most prolific shot creator. He had a reputation as one of few playmakers in the sport capable of consistently hitting his teammates for layups; and he was called the Houdini of the Hardwood for a reason! His display and technique were almost magical with the ways he’d guide the ball while barely flexing his wrist. From later footage, we can also discern he had a mildly effective jumper, and his overall scoring efficiency was higher than average this year.

    Defensively, he was likely a negative, limited by size. Boston’s excellent offense was actually dragged down by mediocre defense to the point at which they didn’t even outscore their opponents in the regular season. But his estimated steal rates were better than average, and he was near the bottom of the league in fouls, which weakly reflects the superb hand-eye coordination that made him one of the game’s best passers.

    Perhaps because of the fledgling tactics of the time, and because fewer teams in the league meant more familiarity with one another, statistical changes weren’t as drastic as they are today. Cousy’s scoring and passing statistics in 300 playoff minutes were right in line with his regular season averages. His rate of free throw attempts dipped in the Playoffs by about 2 per 100 possessions, but it’s worth noting these came in two series against the two-lowest fouling teams (Knicks and Nationals) in the regular season.

    Dolph Schayes—who we just watched sling a behind-the-back pass nearly seventy years ago—was another player with a smoother stat retention in the Playoffs. He described his offensive approach as something similar to what you’d see nowadays: a big man whose threat to shoot would draw defenders out of the lane, after which he’d attack the rim and draw fouls. This is corroborated in his stat sheet, in which he was near the top of the league in free throw rate, and arguably the best at drawing fouls in that year’s Playoffs. Based on his efficiency, it was unlikely he was in the uppermost echelon of scorers; after all, you can only churn so much value out of primitive spacing, and a shot diet that’s heavy in jumpers could suppress True Shooting percentages relative to other bigs.

    His passing (as demonstrated on sparse film and via assist rates) and long-range shooting at his position could easily be considered futuristic, but I’m unsure whether that necessarily made him one of the very best offensive players this year. He was comfortably the best option on that Nationals team; and while I struggle with considering this point too heavily, the Syracuse team was a below-average offense in both the regular season and the Playoffs that year. So while you can point to Schayes for paving some of the way for the evolution of the big man, he was likely a generation too early for that offensive package to exert all of its futuristic value.

    There’s little to learn about Schayes’s defense. DBPM pegs him near the top of the league in regular and postseason; and even though rebounds were likely more reflective of defensive value this year than it is nowadays (Schayes was second in rebounds per possession in the Playoffs), I wonder if the numbers overstate his impact. He stood out the most among his bandmates, though fellows Red Kerr and Early Lloyd weren’t too far behind in dominating the boards. But since we’re working with so few measuring sticks here, and I see no reason to believe Schayes’s defensive portfolio is hollow compared to his teammates, my best guess was that he was a pretty positive on that end. And unlike offensively, Syracuse was the best regular-season defense and near the top in the ‘offs.

    Harry Gallatin is one of the lesser-known Hall-of-Famers in Knicks history, and stacked up well next to Schayes. They had similar marks in BPM as defensively-slanting and providing massive rebounding value. But it’s very likely that Gallatin was more limited as an offensive player. His career isn’t super well-documented, but it’s known his status revolved around his immense physical strength, which allowed him to ascend to top-of-league status as a rebounder; and he was sturdy enough to be known as one of the era’s iron horses. Next to Schayes, he was a notch below as a scorer, albeit on above-average volume and efficiency. He was drawing fouls at a high rate (8.9 free throw attempts per 100), which actually carried over to the Playoffs at an identical rate. His assists rates were also slightly behind; but unlike Schayes, we have no way to qualify his passes. Either way, he seems to have more positive than negative traits on that end.

    The Knicks were at the top of the league in rebounds and near the bottom (in the good way) in fouls, but weren’t clearly an above-average unit. Does this add to the suspicion that the team didn’t roster a true defensive stalwart? And to what degree does that contextualize Gallatin’s stats as the team’s leading rebounder despite not being the center? Ray Felix, the team’s center, struggled in the team’s playoff series and their overall efficiency plummeted. Is this evidence that Gallatin was something closer to a peripheral defender than the centerpiece that his box scores suggest? Either way, he’s clearly a strong value-add, and I think we can make educated guesses about his defensive makeup: helping off opposing fours, boxing out, and cleaning up the boards to allow the true big to focus on defending post-ups.

    Player DBPM change Team DRtg change
    Ray Felix +0.1 10.1
    Harry Gallatin -3.9 10.1

    Ray Felix, meanwhile, was an impressive player on both ends of the floor—a volume scorer on good efficiency was a monster on the boards. Based on height, positions, and the traditional style of play, I think we can reasonably assume he was analogous to the anchor of New York’s defense this year. However, as we discussed with Gallatin, these units were mediocre in the regular season and underperformed in the second season. (Unfortunately, I’m bound to emphasize team-level results with such limited tools to work with.) There’s limited documentation of his career as well, so we’ll have to base our inferences on his scoring, limited passing (based on assists), and infer his role from there.

    Bob Pettit made an immediate splash by capturing the league’s Rookie of the Year; and by my count, actually led the league in scoring rate. He was quickly integrated as a power forward and sported another one of those innovative jumpers. Later in his career, his cerebral style of play included head-on cuts, slips behind defenders, but it’s an unknown how much of this he could have done, given his dramatic weight changes. Paired with strong rebounding and an above-average team defense, there’s no indication he was a below-average defender. He doesn’t have a reputation as a defensive stalwart, but his eagerness to rebound and the sole footage we have demonstrates a keen awareness and activity in the post that’s likely providing positive value. My question is whether he’s on the level of, say, Felix or Gallatin, who both appear to be quite large positives on that end.

    Neil Johnston was in the upper crust of the league’s scorers and a considerable defender by available measures. His Warriors didn’t make the Playoffs this year, but based on postseason changes among big men with similar reputed styles and statistics, it’s unlikely Johnston would have been a playoff “faller.” He had airspace between himself and the next player on my BPM leaderboard; so based on that, he comes across as an early candidate for the season’s best player. The only cause for concern in his profile is the Warriors’ mediocre efficiency, but the surrounding roster doesn’t justify much penalization. Paul Arizin was the team’s next-best player, and he clearly needed time to break in after returning from military service in the previous two years.

    The title of the league’s best scorer this year might have belonged to Larry Foust. Despite eight All-Stars and two All-NBA’s during his career, he surprisingly missed the cut as a Hall-of-Famer, so this is another primitive star for whom we have to infer style from how he seems to stack up next to teammates. He was an efficiency monster in the regular season, but dropped off in two playoff series against teams with strong low-post presences (Mikkelsen and Schayes). During this time, his assists increased significantly, but on more true shot attempts. Fort Wayne’s offense improved in the Playoffs, so it’s worth entertaining the idea that more of Foust was a good thing, although you’d have to distribute that kind of credit unknowingly among teammates Frankie Brian and Dick Rosenthal for similar reasons.

    We haven’t even gotten around to the Minneapolis Lakers yet! Based on my BPM, they actually sported two players with cases near the top of the league in the aforementioned Vern Mikkelsen and Clyde Lovellette. Lovellette’s offense was known to shift between the style of a wing or a big, with his signature move reputed to be the one-handed set shot. That scoring versatility—despite a virtual absence of playmaking—made him a clear offensive value-add. Minneapolis’s defensive stud, however, was Mikkelsen. Like Gallatin, was known for dominating strength and rebounding ability, to the point at which opponents used stalling tactics to keep the ball out of his domain. If we’re to put the duo into modern terms, Lovellette was the rover to Mikkelsen’s centerpiece.

    The Ranking

    Ah! Time for the fun (excruciating) part. I’m obviously far less confident in how I’m going to arrange these players than if I were to do it, say, for the most recent season. Due to our measuring limitations, this is pretty much an anecdotal, box list—not to say that it’s completely devoid of an analysis sufficient to rank. But I want to emphasize how much wider the confidence intervals are. My original plan was to go down to tenth, but the lines were eventually too blurry. Based on everything I just discussed, here’s my top-5:

    1. Neil Johnston [0.18 Ring Shares]
    2. Dolph Schayes [0.14]
    3. Bob Pettit [0.09]
    4. Vern Mikkelsen [0.09]
    5. Harry Gallatin [0.08]

    I’m slightly more comfortable with Pettit and Mikkelsen higher than Gallatin; but realistically, anything past the third spot is up for grabs. Foust and Cousy were actually tied with Gallatin for the fifth spot (meaning Gallatin is my gun-to-the-head pick). I could easily be talked into sliding Clyde Lovellette into the back of the list. I’d also entertain an argument for Milwaukee’s Chuck Share, but his career is so under-discussed it’s difficult to establish a position for him.


  • Using Numbers to Predict the NBA MVP

    Using Numbers to Predict the NBA MVP
    Premise

    This year, I was curious if a player’s box scores, his team’s record, and miscellaneous factors could accurately predict how likely he is to win the MVP. Wielding the results from past award voting—supplied by Basketball-Reference—I got around to establishing patterns among these types of indicators to forecast who’s most likely to win the regular season’s most prestigious award.

    Voting considerations
    • Player value: This should be obvious—the building block of MVP voting. If there’s no indication you’re providing a high degree of value, you’re not in the running for the Most Valuable Player.
    • Team strength: Some voters have argued from the perspective that if you’re not contributing to winning basketball—in other words, if your team is low-seeded or doesn’t look like a contender—you don’t have a good case for MVP. There are more than enough anecdotes and testimonials through the years to support the idea that team record is a relevant factor.
    • Team weakness: Conversely, there are situations in which players are credited when playing with poorer teammates. I’d say it’s likely the ideal MVP in the minds of voters is able to carry teams to extreme heights without another All-Star teammate, thus reducing the error in assuming he’s an incredibly valuable player. Think of Stephen Curry and Kevin Durant in 2017. Both of them could’ve easily been called MVP-caliber players on clearly the best team in the league (67 wins), but received 5.1% and 0.2% of shares, respectively.
    • Historicity: Russell Westbrook’s triple-double averages were a major factor in his winning the 2017 MVP. He proceeded to do the feat an extra two times before being traded to Houston; and yet, he wasn’t an MVP candidate in those seasons. Westbrook earned 88% of voting shares in the 2017 race, a (comparatively) measly 7.5% in the 2018 race, and a negligible 0.8% in 2019! What does this suggest? The rarity of certain events and/or statistical achievements are factors at least somewhat independent of the achievements themselves.
    • Voter fatigue: Part of the spirit of the award—as an outsider observing voter tendenciesis not to simply reward the best players in each season. Otherwise, players like LeBron James and Michael Jordan might’ve racked up even more than they already have. It’s a little more interesting to see new faces and names at the tops of ballots, and makes for interesting story-telling when discussing the legacies of these players.
    Modeling factors

    There are obviously real factors in MVP voting that can’t be quantified. This year, Ja Morant is sixth in these projections despite it being unclear whether he’ll make an All-NBA team. Historicity and voter fatigue—while quantifiable and thus potentially relevant factors in the modeling process—were tested in the construction of this model and were concluded to be insignificant predictors in relation to the other variables. (When it doubt, went with simplicity.) But otherwise, a player’s MVP likelihood was estimated with the following:

    • Box scores: While raw box scores probably account for more of how voters intuit the process, adjustments were made based on how common each statistic was in a season. (This is analogous to “inflation-adjusted” statistics.)
    • Team record: This biases the results toward players on winning teams in such a way that predicts MVP likelihood with more accuracy than if it weren’t included.
    • Yearly adjustment: Because season strength varies, a player’s likelihood is adjusted based on who the other strong candidates are in a given season.
    Historical accuracy
    Season Pred. Winner (Actual Rank) Actual Winner (Pred. Rank)
    2021-22 Nikola Jokic (1) Nikola Jokic (1)
    2020-21 Nikola Jokic (1) Nikola Jokic (1)
    2019-20 Giannis Antetokounmpo (1) Giannis Antetokounmpo (1)
    2018-19 James Harden (2) Giannis Antetokounmpo (2)
    2017-18 James Harden (1) James Harden (1)
    2016-17 Russell Westbrook (1) Russell Westbrook (1)
    2015-16 Stephen Curry (1) Stephen Curry (1)
    2014-15 Stephen Curry (1) Stephen Curry (1)
    2013-14 Kevin Durant (1) Kevin Durant (1)
    2012-13 LeBron James (1) LeBron James (1)
    2011-12 LeBron James (1) LeBron James (1)
    2010-11 LeBron James (3) Derrick Rose (2)
    2009-10 LeBron James (1) LeBron James (1)
    2008-09 LeBron James (1) LeBron James (1)
    2007-08 LeBron James (4) Kobe Bryant (4)
    2006-07 Dirk Nowitzki (1) Dirk Nowitzki (1)
    2005-06 LeBron James (2) Steve Nash (10)
    2004-05 Tim Duncan (4) Steve Nash (11)
    2003-04 Kevin Garnett (1) Kevin Garnett (1)
    2002-03 Tim Duncan (1) Tim Duncan (1)
    2001-02 Tim Duncan (1) Tim Duncan (1)
    2000-01 Shaquille O’Neal (3) Allen Iverson (2)
    1999-00 Shaquille O’Neal (1) Shaquille O’Neal (1)
    1998-99 Karl Malone (1) Karl Malone (1)
    1997-98 Shaquille O’Neal (4) Michael Jordan (3)
    1996-97 Michael Jordan (2) Karl Malone (2)
    1995-96 Michael Jordan (1) Michael Jordan (1)
    1994-95 David Robinson (1) David Robinson (1)
    1993-94 David Robinson (2) Hakeem Olajuwon (2)
    1992-93 Michael Jordan (3) Charles Barkley (2)
    1991-92 Michael Jordan (1) Michael Jordan (1)
    1990-91 Michael Jordan (1) Michael Jordan (1)
    1989-90 Michael Jordan (3) Magic Johnson (2)
    1988-89 Michael Jordan (2) Magic Johnson (2)
    1987-88 Larry Bird (2) Michael Jordan (2)
    1986-87 Magic Johnson (1) Magic Johnson (1)
    1985-86 Larry Bird (1) Larry Bird (1)
    1984-85 Larry Bird (1) Larry Bird (1)
    1983-84 Larry Bird (1) Larry Bird (1)
    1982-83 Moses Malone (1) Moses Malone (1)
    1981-82 Moses Malone (1) Moses Malone (1)
    1980-81 Julius Erving (1) Julius Erving (1)
    1979-80 Julius Erving (2) Kareem Abdul-Jabbar (2)
    1978-79 Moses Malone (1) Moses Malone (1)
    1977-78 Bill Walton (1) Bill Walton (1)

    The model is right more often than it is wrong, correctly predicting 31 of the last 45 MVPs; and 96% of who the model predicts to win have at least been finalists. For the most part, there doesn’t seem to be a relationship between the year and the model’s accuracy, suggesting that these factors have been similarly relevant across time. The box score—as a traditional, unchanging snapshot of player tendencies, despite the analytics movement—has not lost popularity in higher circles. Team record and winning has always been important. 

    Looking through the past results, the only player to really “break” the model was Steve Nash in the mid-naughts. Not only are his MVPs some of the most controversial to date—with opposers citing the heroic efforts of Kobe Bryant and Tim Duncan in these seasons—but Nash is reputed as a player whose skills and value are impossible to capture in the box score. Otherwise, every MVP winner since 1978 has been projected to finish in the top-4, which means one of this season’s group of Giannis Antetokounmpo, Luka Doncic, Joel Embiid, and Nikola Jokic is highly likely to snag the trophy this spring!


  • Nikola Jokic’s MVP Case Is Stronger than You Think

    Nikola Jokic’s MVP Case Is Stronger than You Think

    While the NBA’s MVP has no criteria (and I guess this is a good thing), there’s definitely a set of underlying themes that circulate the ballots every season. There are countless examples from which to draw online. You’ve got those who prefer “the best player on the best team.” Some treat the premise of “value” literally, and support the player with the largest effect on his team’s win percentage. There’s even an originality component every once in a while. This post quotes an NBA.com staff member stating his support for Russell Westbrook as the 2017 MVP for the sole reason that he’d never seen a player average a triple-double before. This means a lot of contradicting opinions are represented in the voting results (and I guess this is also a good thing), which speaks volumes to the historicity of Nikola Jokic’s current run.

    Jokic is on pace to three-peat

    Regardless of your opinion, the safest bet to win the MVP right now is Nikola Jokic. Every major sportsbook has his odds in the -310 to -375 range, with the next-best odds belonging to Joel Embiid via Caesars Sportsbook at +375. Polls drawn from current and past voters all seem to be overwhelmingly in Jokic’s favor. While we’ve had not-so-competitive MVP races in recent memory (Stephen Curry in 2016 and Giannis Antetokounmpo in 2020), Jokic stands out as having narrowed out the competition in the previous two seasons; and as historical fans know, voter fatigue is a real thing. People don’t want to see the same player win the award over and over again. Refreshment is, well… refreshing!

    What else should be working against Jokic’s favor? He’s not a traditionally aesthetic athlete. Most of his prowess on the court is cerebral and decision-making. He’s a slow-footed, nonchalant-looking, formerly-chunky, foreign big man in an era that’s empowered the shot-creating guard and point forward unlike any point in NBA history. Like I said, to put this into a historical perspective is seriously mind-numbing. I almost forgot the league-wide consensus that the talent pool is supposedly the most top-heavy in… ever? What’s going on!

    Jokic is the best player on the best team

    I just lied to you, because technically, that’s not true. While you can take “liberty” with what best means, no one’s going to dismiss the idea that the best team has the highest winning percentage. That’s the Milwaukee Bucks at the time I’m writing this. So far all the ‘best player on the best team” proponents out there, get ready to rally for the Greek Freak’s third MVP in five years. If you’re a sucker for things like point differential (which are likely better indicators of team quality but arguably extend beyond the scope of the MVP), Donovan Mitchell could reasonably be your pick. And if you’re banking on a few decimals swaying one way or the other, Jayson Tatum could snag that title by the end of the year. 

    The Nuggets are certainly very good, and arguably the best team in their conference. (Memphis edges them out in point differential.) But not definitively by any measure. So why am I saying Jokic is the best player on the best team? Let’s reframe what it means to be on the best team for a second. (Trigger warning: I am about to emphasize the Player in Most Valuable Player.) People who say team record or point differential shouldn’t be factored into an individual award have a point, because those top-line measures are the products of supporting casts, coaches, other staff members, travel schedules, and then maybe the star has a share in there. So instead of looking at overall team efficacy, how about team efficacy only when each player is on the floor?1

    Player Team Net On-Court Net
    Nikola Jokic 4.4 13.6
    Joel Embiid 3.7 9.0
    Jayson Tatum 5.7 8.3
    Giannis Antetokounmpo 3.8 7.0

    W-O-W!!! While I won’t concede this as a catch-all remedy to contextualizing team record, Jokic is doing laps around the other candidates! For a little extra flavor in there, consider that Jokic is also leading the NBA in on-off differential (+24.4 per 100). The next-best player on any team that’s not the Nuggets (avoid collinearity) is Jaren Jackson Jr., who barely has half (+12.8) of what Jokic’s differential is! Again, this isn’t meant to end any conversation, rather spice it up a little bit. But when you look at winning from this perspective, it’s hard to argue that one name and one name alone doesn’t absolutely pop of the page!

    Jokic is the darling of analytics darlings

    Every year, it seems like Jokic’s BPM, his EPM, his RAPTOR, and his (insert acronym here) jumps to yet another unconceivable level. I’ve said many times before, as someone who’s studied these metrics with some intensity, that they do a fairly-good job at ballparking value for the vast majority of players. While acknowledging their imperfections, discussing modeling biases and techniques, and working through the analytics analytically is wonderful for evaluating players, that’s not the purpose of the MVP. Let’s straight-up compare Jokic’s ranks among the other MVP candidates.2

    Player BPM Rank LEBRON Rank RAPTOR Rank
    Nikola Jokic 1 3 1
    Joel Embiid 4 4 4
    Jayson Tatum 10 5 11
    Giannis Antetokounmpo 14 2 14

    Wait, hold on… Jokic is third in LEBRON? That doesn’t sound right… If that’s what you’re thinking, you’re right! While all the other ranks are among all players this season, Jokic’s ranks are among all player-seasons in history! While LEBRON and RAPTOR are hybrid metrics that (in full form) only have a decade’s worth of information, BPM goes all the way back to the shot clock (1955)! While this isn’t necessarily to say he’s definitively having the greatest regular season ever, let’s take some time to let those numbers simmer in our heads.

    What doesn’t count in the MVP

    There are exactly two points that have been made historically in MVP talks that I want to address, because they explicitly violate the premise of the MVP, not because I think they’re fallacious or illogical. So if you’re an Embiid or Tatum stan who’d like to white-knight on their behalf in the comments, I recommend checking back to the next two bullet points for a guide on which criteria I consider irrevocably irrelevant:

    • Playoff translation: Bam Adebayo was recently quoted for saying Rudy Gobert shouldn’t have won DPOY because his game didn’t work well in the Playoffs. This is irrelevant. The MVP is a regular-season award, blah blah blah… But seriously! Take it seriously if you care enough to rip people in the comments!
    • Voter fatigue: Here’s one way to think about voter fatigue. By not choosing a player because they’d previous won the award, you are by definition taking previous seasons into account. The MVP (and all awards like it) are exclusive to the one regular season. (For example, the upcoming 2023 MVP will only be considering the 2022-23 regular season.)

    With that all said, those are the main points I wanted to hit to try to explain the historicity of Nikola Jokic’s run for his third consecutive MVP, and it’s looking like it’s going to happen! I’d love to talk about it in the comments with people who disagree (or people who agree who have other points not referenced in this post), so have a go at it!


  • These Skills Are Most Valuable According to Analytics

    These Skills Are Most Valuable According to Analytics

    Analytics are imperfect, and we’re all aware of it. Proponents refer to them as ballparking measurements. I’ve used the “one size fits all” sock as an analogy, in which certain metrics tend to “fit” players better or worse depending on things like playstyle and modeling techniques. For my money, they’re pretty decent estimates of player value—good to fill in gaps that our eyes miss when we watch games. The other thing I really like analytics—and specifically impact metrics—for is understanding which types of players provide the most value. Basketball isn’t a clear-cut game. Binary outcomes are fueled by fluid process. We don’t always know where to assign credit. Impact metrics help us fill in those gaps.

    The Goal

    How can we use impact metrics to identify valuable skills and traits? That’s what I’m setting out to learn in this post. To do so, and for the sake of simplicity, I’m consulting the LEBRON metric. (Read LEBRON primer here.) It’s no exception from the myriad of alphabet soup we see in the analytics community today. LEBRON is a catch-all metric that blends statistics and spits out a one-number player rating. The ingredients of LEBRON, which were the focuses of its appeal, are:

      • Traditional counting stats: The box score—points, rebounds, assists, etc. LEBRON adjusts a player’s box score based on how early in the season it is to avoid jumping the gun on streaky performances. (Read about LEBRON’s padding technique here.) The drawbacks include playstyle bias and incomplete tracking.
      • Plus-minus: The top-down approach to analyzing players statistically. This is unbiased toward playstyles and only cares about the scoreboard. The drawbacks include sensitivity to team circumstance and sample size.

    LEBRON has one of the nicest mixtures of statistics as inputs, drawing from the two major types (bottom-up and top-down) of statistics. The other appeal to LEBRON comes with its provider, BBall-Index, which has wonderfully supplemented its LEBRON database with models for “offensive archetype” and “defensive role.” [1] (Read primers on Offensive Archetypes and Defensive Roles here, here.) Here are examples of today’s stars and how they’re categorized:

      • Stephen Curry: Primary Ball-Handler (offense), Low Activity (defense)
      • Kevin Durant: Shot Creator (offense), Helper (defense)
      • Nikola Jokic: Post Scorer (offense), Anchor Big (defense)

    These archetypes and roles—like analytics—are imperfect. For example, we know Nikola Jokic is one of the game’s best playmakers, so calling him a “Post Scorer” is reductive. However, these estimates are still really good, and the “objective” nature of these roles adds to consistency and rigidity when we classify players.

    The Method

    Because it’s a plus-minus-based metric, LEBRON is only available from the 2013-14 season onward. Since then, there have been 2,406 player-seasons with more than 1,000 minutes played in the regular season. What we can do is compile these statistics—offensive archetype, defensive role, and LEBRON—and plot out which roles correlate with the strongest impact on both offense and defense. Based on those plots, we can point out key trends, uncertainties, and hopefully draw rough conclusions about which “skills” (isolation scoring, passing) provide the most value to basketball teams.

    The Results

    Let’s start with O-LEBRON and Offensive Archetypes. Below is a grouped box-plot that illustrates the O-LEBRON distributions for each Offensive Archetype. Underneath are the descriptive statistics of the O-LEBRON distributions, also grouped by Offensive Archetype.

    There’s a mountain of evidence that O-LEBRON views Shot Creators [2] are the most valuable offensive players. They’re worth +1.56 points per 100 possessions on average, and interestingly post the highest standard deviation in its values among all archetypes. Not only do these players tend to provide lots of value, but the best of the best Shot Creators are also able to separate themselves from their peers unlike any archetype. Primary Ball-Handlers are the most similar, with the difference in their archetype being a lower rate of isolation scoring.

    Movement Shooters and Off-Screen Shooters sound similar, yet the latter has a visible advantage. What’s the difference? Movement Shooters are mostly identified by their three-point rate, while Off-Screen Shooters have a higher proportion of scoring coming off screens or hand-offs. Interestingly, Movement Shooters cost their teams 0.3 points per 100 on average, while Off-Screen Shooters add 0.2 points on average.

    Athletic Finishers and Slashers seem to have overlap. Yet, once again, the latter has a visible advantage. So what’s this difference? Athletic Finishers are characterized by their activity on cuts and putbacks, whereas slashers are determined by their three-point rates and drives per possession. Athletic Finishers were the least valuable archetype on average, costing 0.9 points per 100. Slashers, on the other hand, were the fourth-most valuable, adding 0.4 points per 100.

    Post Scorers and Stretch Bigs are on opposite ends of the basketball timeline. The latter is seen as more progressive, yet Post Scorers are still more valuable according to LEBRON. They’re worth about 0.3 points more than Stretch Bigs per 100, and their distribution reaches higher. Meanwhile, Versatile Bigs tend to be more valuable than both! They can post up, shoot threes, roll to the basket, and are supposed to have more well-rounded skill sets.

    Like Shot Creators for offense, Anchor Bigs appear to be the most valuable archetypes on defense. They are seen as less mobile, and tend to drop against shooters to defend the pick-and-roll. They also tend to not have very versatile matchups compared to other archetypes. Basically all of the all-time great defenders in the LEBRON era have been Anchor Bigs. Mobile Bigs trail as the “second-most valuable,” and these players tend to hedge screens and switch more heavily compared to Anchor Bigs. They’re worth 0.6 points per 100 on average.

    POA’s and Wing Stoppers are both perimeter-oriented defenders. POA’s often defend opposing Ball-Handlers, whereas Wing Stoppers defend Shot Creators. POA’s have higher rates of matchup versatility, whereas Wing Stoppers have more responsibilities off-ball, switching, et cetera. Both archetypes are about neutral defenders on average, with the best Wing Stoppers reaching higher than the best POA’s.

    Chasers and Helpers are both archetypes that involves lots of movements and court coverage. Chasers stick to shooters and cutters near the perimeter while Helpers are savvy rotators who provide resistance at the rim. Helpers are observably more valuable on average, and have reached far greater heights than the best Chasers.

    Low Activity defenders are best known for their communication abilities when engaged, but otherwise provide little value in defending actions and matchups. They are, unsurprisingly, almost always value-losses.

    Interpretation

    Based on the differences between Primary Ball-Handlers and Shot Creators, there may be an indicator that (at least, for guards) isolationism is more valuable than pick-and-roll actions. As observed, Primary Ball-Handlers can still be some of the very best offensive players in the league, but the evidence suggests isolation scoring is typically more valuable, and that pick-and-roll ball-handlers likely have to be very, very good at what they do to reach those heights.

    There’s also an absence of the assessment of passing quality in these archetypes. I’m compelled to believe passing bleeds into both of the above archetypes. Isolation scoring on its own is not very “valuable,” since it rarely adds more than 1.1 points per possession on average. For teams to exceed the league average of 1.13 in these years, they need higher-quality shots. As has been observed, isolation scorers who can pass to open teammates provide a dual threat that is unmatched by most any players. Perhaps that is why Primary Ball-Handlers can still reach the top of the O-LEBRON leaderboards. They create a ton of shots at the rim in the pick-and-roll without a massive scoring punch.

    The differences between Movement and Off-Ball Shooters could be a commentary on “teamwork.” Perhaps the quality of a player’s looks when his teammates are involved in getting him open. Teams employ screens for obvious reasons: to stunt defenders. Players who make timely cuts also exploit defensive inattention. Movement, while it stretches the defense, likely isn’t on its own a large value-add. There needs to be some predication, an action that prohibits defenses from guarding tightly.

    Athletic Finishers are clearly less valuable than Slashers despite similar damage at the rim. There’s something to be said geometrically, where players who pressure the rim with drives rather than scoring on second-chance opportunities are greater advantage creators. They can pass out of these spots, and their on-ball threat to score draws defensive attention in a more looming way than cutters. This is another instance in which the value of the scoring-passing combination seems to bleed through.

    Big who stretch the floor are “less bad” than the worst Post Scorers, but they don’t provide the higher levels of value that the best Post Scorers do. This could have something to do with scoring around the rim, whereas bigs who draw attention to the perimeter are “always” providing a geometric effect that Post Scorers often don’t. But it’s clear that a combination of the two is best, and that bigs pretty much need to score at the rim at a decent clip to ascend offensively.

    Defensively, the value of rim protection is Obvious with a capital “O.” There’s the narrative that shots at the rim are most efficient, so defenders who can not only sway them but deter them are most valuable. This seems true. Bigs who hedge screens may pose the risk of blow-bys and more shots at the rim, which might explain their decrease in value despite a more versatile skill set. Regardless, these defenders can still be really good by fizzling out perimeter actions while providing occasional paint resistance. But rim protection on its own seems to remain the gold standard for best-of-the-best defense.

    Because one-on-one basketball tends to not yield great results, POA defenders aren’t super valuable. (Especially if they don’t provide skills elsewhere in deterring high-quality shots.) It’s the defensive equivalent to decent isolation scorers who can’t pass. Unless they reach a certain threshold for efficiency and passing, they’re not really doing much. Therefore, as the data suggests, there’s more value in defending Shot Creators whose skills demand off-ball awareness. Helping, switching, and resisting on the perimeter are noticeably more valuable.

    Chasers are the antithesis to the “Movement” offensive skill, so it’s no surprise they are rarely positive defenders. (Again, especially if they don’t provide value in other areas.) But there have been some great Helper defenders in recent years. They are paramount paint defenders like bigs, but the positioning and switching that comes with the role does provide ancillary rim protection.

    Footnotes

    [1] Note that offensive archetypes and defensive roles aren’t based on “how well” a player carries out his role, but solely what that role is.

    [2] “Shot Creator: players in this offensive role are non-Bigs that we identify as have high rates of perimeter and interior isolation rates, creating their own shots within the offense as a key skill set. Examples from 2019-20 are Luka Doncic and James Harden” (BBall-Index).


  • MVP Power Rankings | Volume II

    MVP Power Rankings | Volume II
    Premise

    Historically, the MVP has been chosen arbitrarily—a mingling of analysis and intuition. This can be great by promoting varying styles of analysis. Broader conversations can launch new names into conversations. Different ideas challenge norms. For this list, I adhere to a strict criteria I’ve developed over the years—an amalgam of analyzing film, statistics, and value-theory. The overarching question I try to answer with these rankings is how well a player sets up a random team to win the championship. (This approach is derivative of the works of analysts like Seth Partnow and Ben Taylor.)

    This still limits consideration to the regular season. Expectations for Playoff risers and fallers is irrelevant. I solely care about how well a player sets the team up to succeed in the Playoffs (where “things matter most”). This means actual seeding is less important, insofar as home-court advantage doesn’t play a crucial role in later rounds. To balance these factors, I concocted a championship odds calculator that inputs estimates of player value and games played. (The impact estimates are based on analysis and interpretation.)

    The Ladder

    10. Damian Lillard (NR)

    This is dependent on Lillard not being a train wreck on defense, which the impact metrics seem to agree on. His offense has reached the level it once achieved, and his volume combination of scoring and playmaking is matched only by Luka Doncic. Major one-number metrics also seem to think he’s on the level of Curry, Jokic, and the likes.

    9. Donovan Mitchell (-)

    It’s been a few weeks since I’ve watched Cleveland, so I’m banking on a sustainable improvement in his defensive awareness here, which I imagine is feasible based on his situation. His offense is seriously defiant of the vertical nature of the game, though I do wonder a bit about his fairly low rate of scoring around the rim. Will restraint minimize his drive-and-kick playmaking? Not yet, that’s for sure!

    8. Jayson Tatum (-1)

    I don’t have much more to say about Tatum compared to last month. I’d like to see a bit more creation for teammates if I’m going to move him into serious MVP contention. But, wow… His combination of elite scoring and smothering perimeter defense is something you get from no one else on this list.

    7. Giannis Antetokounmpo (-1)

    If anything, I was expecting him to move up! His scoring efficiency has some catching up to do, though the remainder of his offensive skills seem to be on par with previous seasons. To my eye, his defensive awareness is slightly down, though I refrain from settling on that as a real thing since I don’t watch every game. Regardless, it’s hard to surpass an acceptable offensive number-one with Defensive Player of the Year potential.

    6. Anthony Davis (+2)

    This definitely isn’t my most confident pick. Davis is clearly better on both sides of the ball. LeBron James’s dominant style of offense has taken the backseat to Davis’s chops as a finisher and screener, and the result is mind-boggling statistics. His scoring output is at an all-time high. He looks like one of, if not, the best defensive players in the world. He also just looks better moving, manipulating low-post defenders with sly cuts and stampedes.

    5. Joel Embiid (-)

    Embiid is a fantastic two-way star. His scoring is the best it’s ever been and his playmaking is improving. He’s a beast around the rim on both sides of the ball, and his guard-like movements make him one of the most dangerous players in the game. He’s definitely playing like an MVP-level player, and if he were doing this two or three years ago, he might be a clear-cut finalist.

    4. Luka Doncic (-)

    I’m notoriously hesitant about gawking at volume statistics. Doncic is arguably the most prolific scorer and playmaker in the sport. Does that necessarily translate to championship offense when his teams are characterized by lackluster efficiencies and late-game fatigue? That’s why I’m not yet convinced. His defense is slightly improving however, and being one of the four-best players in basketball is nothing to shy away from.

    3. Stephen Curry (-2)

    Two main points move Curry down for me compared to last month: 1) the unsustainability of his finishing, which I explained was the deciding factor in his top spot, and then there’s the increase in competition that rivals his scoring output. If he were a strong defender, he’d possibly have the top spot again. But in this current defensive state with the Warriors, I moved him down until further notice.

    2. Kevin Durant (+1)

    There’s the aforementioned passing improvement from my last post, and his more solid defensive package pushes him over Curry for me. He’s now officially a member of the prestigious “30 points on +10% efficiency” club, and remains one of the three-best scorers and offensive players in the world for me.

    1. Nikola Jokic (+1)

    The best offensive player in the world, looking like one of the greatest offensive players of all time. He’s had an argument as basketball’s top scorer and playmaker for over a year now, but this season he’s dialed it up to eleven. There’s more than enough time to wait for his offensive rebounding to catch up. He also finally looks like a neutral defender in impact metrics, which to my eye is more reflective of his actual impact.

    Resources

    [1] Data from Basketball-Reference, DunksAndThrees, Thinking Basketball


  • 2023 All-Star Power Rankings | Volume II

    2023 All-Star Power Rankings | Volume II

    You know the deal: All-Star power rankings. Which players are clear-cut? Which ones perhaps have a little more to prove? Divided among the tiers below are the players I evaluate as worthy of All-Star consideration. Let’s do it.

    Tier 1

    These players are undoubtedly performing at an All-Star level or higher. The gaps in play quality in this tier are the highest of any succeeding tier.

      • Giannis Antetokounmpo (East)
      • Devin Booker (West)
      • Jimmy Butler (East)
      • Stephen Curry (West)
      • Anthony Davis (West)
      • Luka Doncic (West)
      • Kevin Durant (East)
      • Joel Embiid (East)
      • Shai Gilgeous-Alexander (West)
      • Tyrese Haliburton (East)
      • James Harden (East)
      • Jaren Jackson Jr. (West)
      • Nikola Jokic (West)
      • Damian Lillard (West)
      • Donovan Mitchell (East)
      • Ja Morant (West)
      • Domantas Sabonis (West)
      • Pascal Siakam (East)
      • Jayson Tatum (East)
      • Karl-Anthony Towns (West)
      • Zion Williamson (West)
    Tier 2

    Either because of talent-scaling (the league is more top-heavy than ever), a higher margin of error associated with an All-Star level of play, or general uncertainty about their game, these players are “doubtedly” All-Stars.

      • Bam Adebayo (East)
      • Jarrett Allen (East)
      • Desmond Bane (West)
      • Jaylen Brown (East)
      • DeMar DeRozan (East)
      • De’Aaron Fox (West)
      • Darius Garland (East)
      • Paul George (West)
      • Rudy Gobert (West)
      • Jrue Holiday (East)
      • Brandon Ingram (West)
      • Kyrie Irving (East)
      • LeBron James (West)
      • Brook Lopez (East)
      • Lauri Markkanen (West)
      • Myles Turner (East)
      • Trae Young (East)
    The Ballot

    You know the rules: 5 starters (2 frontcourt, 3 backcourt); 5 reserves (2 frontcourt, 3 backcourt); and 2 wild cards (position negligible). Rosters for both the Eastern and Western Conference.

    East
      • James Harden
      • Donovan Mitchell
      • Giannis Antetokounmpo
      • Kevin Durant
      • Joel Embiid
      • Darius Garland
      • Tyrese Haliburton
      • Jimmy Butler
      • Pascal Siakam
      • Jayson Tatum
      • Jaylen Brown
      • Jrue Holiday
    West
      • Stephen Curry
      • Luka Doncic
      • Anthony Davis
      • Nikola Jokic
      • Zion Williamson
      • Damian Lillard
      • Ja Morant
      • Jaren Jackson Jr.
      • Domantas Sabonis
      • Karl-Anthony Towns
      • Devin Booker
      • Shai Gilgeous-Alexander
    Thoughts

    I’m fairly pleased with how this ballot turned out. The East wild cards were the most difficult to land on (the player pool was the entire second tier), but it’s about time I showed Jaylen Brown some All-Star love. I’m perfectly happy with the West, save for a slight discomfort caused by the strength of the conference’s guards. Devin Booker and Shai Gilgeous-Alexander being wild cards feels unnatural.


  • MVP Power Rankings | Volume I

    MVP Power Rankings | Volume I
    Premise

    Historically, the MVP has been chosen arbitrarily—a mingling of analysis and intuition. This can be great by promoting varying styles of analysis. Broader conversations can launch new names into conversations. Different ideas challenge norms. For this list, I adhere to a strict criteria I’ve developed over the years—an amalgam of analyzing film, statistics, and value-theory. The overarching question I try to answer with these rankings is how well a player sets up a random team to win the championship. (This approach is derivative of the works of analysts like Seth Partnow and Ben Taylor.)

    This still limits consideration to the regular season. Expectations for Playoff risers and fallers is irrelevant. I solely care about how well a player sets the team up to succeed in the Playoffs (where “things matter most”). This means actual seeding is less important, insofar as home-court advantage doesn’t play a crucial role in later rounds. To balance these factors, I concocted a championship odds calculator that inputs estimates of player value and games played. (The impact estimates are based on analysis and interpretation.)

    The Ladder

    10. Ten

    The tenth spot is such a toss-up that I may as well treat it like an Honorable Mentions! Names considered for this spot include, as alphabetized:

      • Devin Booker
      • Jimmy Butler
      • Paul George
      • Shai Gilgeous-Alexander
      • James Harden
      • Ja Morant
      • Karl-Anthony Towns

    9. Donovan Mitchell

    Mitchell has upped his offense to a new level. The stylistic aspect of his game is consistent, but its efficiency is unbelievable. He attacks the rim like a madman and creates tons of offense for the corners. (Now we know Gobert wasn’t the problem.) There’s his shooting… my goodness. His defense has also been serviceable too; after multiple seasons of questionable play on that end, his anticipation and ability to clog passing lanes look better than last year. I don’t see evidence that he’s a clear negative.

    8. Anthony Davis

    Anthony Davis is… almost back? Not quite there. He’s somewhat close to his 2020 form, though it’s unlikely he reaches it. Any semblance of a jump shot he had during the bubble has recessed, which diminishes his spacing value. He’s also held back on the passing punch he had a few years ago. Regardless, he’s the Defensive Player of the Year in my eyes. Davis has an unmatched combination of valuable and scalable traits on that end, and… my gosh. Name any defensive skill and he’s got it.

    7. Jayson Tatum

    Tatum is having the most efficient season of his career despite a slight drop in three-point shooting! His shot selection and isolationism are both upgraded from last season. He looks better as an athlete, and that physical aid may help him ascend to “obligatory wing defender in DPOY conversations” territory! (No, seriously, his ball-pressure is some of the most impressive defensive work in the league.) Tatum has a chance to win the actual award because he plays for the game’s best team. On my list, he lands seventh.

    6. Giannis Antetokounmpo

    This is… weird. He’s still a prolific scorer and playmaker with Defensive Player of the Year chops on the other end. Those things alone make him a mainstay on the upper end of a list like this. However… his offense has come with a lot of opportunity cost so far. Antetokounmpo’s shot outside the paint is lagging hard to start the season, and it drags his overall efficiency below the league average. That’s enough for the major impact metrics to peg him in this range rather than finalist territory.

    5. Joel Embiid

    Embiid is probably not in the uppermost echelon of scorers in my book. (Even though he’s an incomprehensible isolation scorer.) But as his passing ability trickles in from more opportunities to create, his overall offense looks stronger. He’s not what I’d consider a really “good” passer, but he has enough range on his deliveries to be the primary force on a contending offense. Embiid is still a monster rebounder and low-post defender whose skills check more boxes than most other players in the world.

    4. Luka Doncic

    Doncic is probably (?) the favorite to win the award, so this may be a surprising placement. I’m not sold on “Luka-Ball” as the next model for great offense. The term “heliocentrism” is thrown around like pennies into a wishing well, but Doncic’s role in the current Mavericks roster demonstrates a drawback. Dallas tends to wear down late in games. Doncic is the most prolific decision-maker in the league, and it seems his relative lack of conditioning weakens the entire attack down the stretch. Regardless, he’s still fourth for me due to an unheralded mix of volume scoring and creation for teammates.

    3. Kevin Durant

    Durant’s relative True Shooting percentage is 7.7% ahead of the league; once his three-point percentage is up to speed, that number is going to skyrocket his name into contention for the best overall scoring numbers in the league. His un-guardable jump shot is still intact. He’s having the best passing season of his career to my eye—slicing out holes in perimeter coverages with more anticipation and stronger deliveries than ever. Add that to a positive defensive package, and you’ve got a strong MVP candidate.

    2. Nikola Jokic

    Jokic was the clear-cut MVP of the last two years. His scoring volume isn’t quite up to par yet, and voter fatigue is due to trickle in from the public and voters alike. That’s my estimation as to why he isn’t seen as an obvious finalist. Jokic is the most efficient of the game’s volume scorers, and he’s the best passer on the planet. He’s arguably the most complete offensive player in history. His rim protection as Denver’s primary backline defender still lags behind, which is the only thing holding him back from the top.

    1. Stephen Curry

    My, oh my… Where to start? Curry is “re-peaking” as the league’s top scorer and shooter, with more than a little airspace above the contenders. His leading spot is probably dependent on a permanent upgrade in his scoring around the rim, where he’s converting at a 76% rate. He has the craft and guile, but likely not the durability and positive aging signals to maintain it for long. (But I’m partial to streakier two-point than three-point improvements.) He’s otherwise Curry doing Curry things, and he’s at the top of virtually every major one-number metric.

    Resources

    [1] Data from Basketball-Reference, BBall-Index, Thinking Basketball


  • The Lauri Markkanen Corollary

    The Lauri Markkanen Corollary

    The Utah Jazz have received a lot of press in these past four weeks. The most surprising team to start the season, are they not? Perhaps. They are (as of the time of this writing) the fourth seed in the West with a positive point differential—worth a double take considering they were seen as a contender to land Victor Wembanyama. “But what does this have to do with Lauri Markkanen?” you ask. That’s a decent question—and the answer has everything to do with All-Star voting! Let’s take a dive into what in the Julius Randle is going on.

    The Premise

    Historically, All-Stars are selected through a mingling of their performances and their teams’ standings. This is great news for Lauri Markkanen. Why? Let me begin with an assertion—Lauri Markkanen is not an All-Star-level player, but… because of the Jazz’s record and preseason expectations, he might stand a chance to make the All-Star Game. “But… but…” you ask, “if Lauri Markkanen isn’t an All-Star-level player, why does he deserve to make the All-Star Game?”

    I’m so glad you asked.

    From the perspective that players should be rewarded for their ability or skill, the boost that some receive from being in the right situation is stupid. I’ve already fleshed this out (read here) so I won’t belabor any points. How this is relevant has lots to do with Lauri Markkanen’s case because, like I said, he doesn’t demonstrate All-Star skill but his name is littered in All-Star conversations. This was obvious when I published my first All-Star ballot of the season. My omission of Markkanen from either tier of All-Star consideration implied I did not consider him an All-Star-level player. This is correct. Resisting comments were quick to defend Markkanen’s case:

    I clearly disagree with this, don’t I? And what’s the best way to further a point in a caring and considerable manner?—Making the opposing argument as strong as could possibly be.

    To the best of my ability, I am going to lay out Lauri Markkanen’s All-Star case—but with a twist. I still only care about his efficacy as a player. Team, teammates, and attributes of the Utah Jazz’s system are irrelevant. I believe All-Stars deserve their recognition for playing like All-Stars, and Markkanen won’t receive a special treatment. So let’s get into it!

    Scouting Report

    Markkanen has demonstrated solid three-point shooting during his career. His three-point percentage of 36.5% is unspectacular, and right in line with his career average of 36.4%. To me, he provides value as a shooter away from the ball—a stretch four who can catch and shoot at an above-average clip (40.2%), mostly in pick-and-pop situations. On the other hand, he’s an infrequent and inefficient pull-up shooter—a skill crucial to a player’s ability to create offense for teammates in a spaced-out offense. The Jazz have methodically had him work from the corners, spots (in which his three-point percentage is 45.8%) that are valuable real estate. The corners are also outlets for his attacks to the paint.

    Markkanen is tall, strong, and sturdy with a fine-tuning of footwork that creates separation between defenders in the paint. He has spins, twists, and twirls that carve out floater-range shots. (Markkanen’s percentage of attempts that are floaters had nearly doubled from its previous high.) Farther downhill, he’s a solid finisher who can draw fouls and convert at high rates—driving about five times per game and finishing at a 76% rate. Markkanen does require “assistance” from teammates’ passes at times; he has burst as a big man, but not enough to get to the rim at will. (He often picks up his dribble only a few steps into the paint.) His bruising and jostling allows him to contend with formidable big men like Bam Adebayo and Anthony Davis close to the basket.

    This is where his All-Star case becomes tricky. His scoring punch has been good—not great, for all intents and purposes—without a strong isolation package or slashing ability. His style is suited to play alongside a more demanding offensive force, a truer “number one” who can leverage the pick-and-pop and make timely passes when Markkanen cuts baseline. That’s a pretty good scoring package, but where it falls short is in its ability to boost the value of his lackluster passing. He’s had flashes of range and accuracy, but nothing that—when sifted through—indicates he’s growing into the role of a playmaker. (For instance, the percentage of Markkanen’s assists that end in layups is 21% below the league average.)

    His defensive skills are slightly fuzzier to me. Markkanen defends a fair amount of shots at the rim, inducing misses without a high block rate. But any skills he has a rim protector have yet to translate to latent value, such as deterring shots at the rim. Teams are content to attack the rim with Markkanen on the floor, which could be a problem due to his nonexistent presence as a perimeter defender. He exemplifies the Jazz’s inconsistent switch tactics, so he doesn’t content many threes nor is he an avid helper. Without brushing up on his defensive range—which seems unlikely to happen in Utah—Markkanen’s argument as a clearly positive to strong defender seems weak.

    The Jazz offense is a suitable place for him to mimic his ideal offensive role: a pick-and-pop, bruising, floater-range specialist who can score at two levels. In Utah, he’s a semi-frequent but undesirable pick-and-roll ball-handler, which is an action teams would want to avoid to build a strong offense. Paired with weak passing, I see Markkanen as a solid third-to-fourth option on a contending offense. Defensively, he’s going to need backline help from a stronger, rangier rim protector; and if he’s the primary rim protector, his team will need to bank on strong defensive play from guards to prevent open three-pointers.

    Here are Markkanen’s ranks in high-level impact metrics [1]:

      • Backpicks BPM +2.6 (48th)
      • Basketball-Reference BPM +2.8 (36th)
      • Box RAPTOR +1.7 (80th)
      • EPM +4.7 (23rd)
    The Strongest Case

    What is Markkanen’s upper bound?—the highest extent to which I can evaluate his impact. If that estimate doesn’t match All-Star level, by the rules of this exercise, Markkanen has no case to be an All-Star. I’ll view each of skills through the rosiest of glasses, give him the benefit of the doubt in all reasonable areas (considering the trade-offs between skill interactions). To start, here are Markkanen’s strengths as a player, by my scouting report:

      • Floater-range footwork
      • Bullying in the paint
      • C&S and screening at the corners
      • Complementary rim protection

    I can’t reasonably upgrade his passing, nor is his off-ball package enough for me to say he’d be a “number two” on a contending offense. For that, he’d need rangy, connective passing to and from his corner spots. (In theory, these could lead to more layup passes.) His footwork and physical attributes can dismiss the notion that his scoring near the basket “will eventually cool down.” This version of his scoring—high volume on high efficiency—could be here to stay.

    Defensively, I’m still convinced he needs backline help. His opponents are finishing at a low clip when he defends at the hoop, but it doesn’t justify Markkanen’s sedentary defensive role. He could probably help keep a poorer defensive afloat—but without a flank of perimeter defenders or a better rim protector as a failsafe, Markkanen’s defensive package is neither good nor bad. This all goes to say I see a limited ceiling on how highly I can evaluate his defense.

    Markkanen’s impact metrics are inconclusive. EPM, which includes tracking data and plus-minus, pegs him at an All-Star level. But RAPTOR also includes these parts (in a varied fashion) and indicates he’s not close to contention! The box-score metrics both agree he’s outside contention. These metrics, in which force-fits to team performance can overstate players on teams that are greater than the sum of their parts (the Jazz), are reluctant to launch Markkanen into All-Star territory. This signal works against his “strongest” case.

    Conclusion

    Markkanen is not an All-Star.

    Returning to the article’s title—what is the Lauri Markkanen Corollary? To my estimation, it’s when a team (the Jazz) jumps out with unexpected success. That team, whoever it may be (the Jazz), plays an egalitarian style, and its success is the function of many contributions from solid or good players, rather than fewer contributions from great players. But this explanation is unsatisfying—it’s too long, takes up too much headspace to put all the pieces together. Thus, the instinct is to look to one player (Markkanen) for the majority or all of his team’s success—the Lauri Markkanen Corollary.

    Footnotes

    [1] Box RAPTOR is my preferred variant of RAPTOR for all players. Especially early in the season, the plus-minus component is unstable. Markkanen ranks 66th in total RAPTOR.

    [2] Data collected from Backpicks, Basketball-Reference, BBall-Index, DunksAndThrees, FiveThirtyEight, NBA, PBPStats.


  • 2023 NBA Power Rankings | Volume II

    2023 NBA Power Rankings | Volume II

    Every few weeks, I power rank teams based on their likelihood to win the 2023 championship, as decided by me! Last month’s edition was a success by my standards considering I didn’t lose sleep over it. But a lot has changed in the NBA landscape. A lot. Listed alongside each team is its change in rankings from the preseason edition. (NB: The gaps between teams means less in the lower ranks. After the “Good” teams, all odds are essentially zero.)

    Let’s Not Talk About It
    30. San Antonio Spurs (-3)

    Already forgot them.

    29. Houston Rockets (-)
    28. Detroit Pistons (-3)
    27. Orlando Magic (-4)

    Weird aesthetics.

    26. Charlotte Hornets (-5)
    25. Oklahoma City Thunder (+3)

    Giddy for Giddey.

    Average-ish
    24. New York Knicks (-4)

    Greek Yogurt.

    23. Washington Wizards (-1)
    22. Chicago Bulls (-3)

    Prototypically average.

    21. Utah Jazz (+10)

    They have a “salty” flavor to them.

    20. Atlanta Hawks (-2)
    19. Brooklyn Nets (-9)
    18. Los Angeles Lakers (-3)

    Deepest of the deepest sleepers.

    17. Portland Trail Blazers (+1)

    Called it I guess…

    Good
    16. Indiana Pacers (+10)
    15. Minnesota Timberwolves (-3)
    14. Miami Heat (-7)
    13. Sacramento Kings (+11)

    LOLOLOLOLOLOL.

    12. Toronto Raptors (+1)
    Pretenders
    11. Memphis Grizzlies (+3)
    10. Denver Nuggets (-2)

    Eh.

    9. New Orleans Pelicans (+7)

    Wow! Cool!

    8. Dallas Mavericks (+3)

    Strange. Very strange…

    7. Philadelphia 76ers (-1)
    6. LA Clippers (-2)

    They might be a contender. Kawhi. I don’t know.

    Contenders
    5. Cleveland Cavaliers (+6)
    4. Golden State Warriors (-3)

    Don’t ask.

    3. Phoenix Suns (+3)
    2. Milwaukee Bucks ()

    Two of arguably the three strongest DPOY candidates (Antetokounmpo, Lopez) defending the backline; and that doesn’t even begin to scrape Holiday’s defensive impact. The offense will hopefully stop lagging when All-Star Khris Middleton returns.

    1. Boston Celtics (+2)

    The best offensive team in the NBA designed for repetition and sustainability, poised for even greater success when all the elements of their defensive core return from injury. The clear-cut frontrunner for the 2023 NBA championship in my book.

    Nice job?