Best Publicly Available NBA Projections

With the surge in popularity of sports betting, NBA Win Totals have gone mainstream. It’s a useful exercise—a macro look at an NBA season which is long and filled with the variance of injuries, drama, and shot luck. While people love to predict sides, rarely do they want to set lines themselves, and even more rarely do they want to pony up their own money on it. Well, these are the best publicly available win projections. And I’ll bet on it [1].

Professionally [2], I live by a fairly easy motto: ABQ. Always Be Quantifying. So this year I went through every NBA team, aggregated minutes projections, added an aging curve to some of my favorite available one number metrics, and threw in a splash of subjective adjustment when necessary [3].

Often we think we know something the market doesn’t, but we don’t. The market is vast and its incentive structure (making money!) is robust. In this case, I know I don’t know anything that the market doesn’t. I’m just a guy, armed with some numbers and a bias toward liking NBA players that can pass, shoot, and try on defense.

This process surprised me (as good quantifying often does!). I didn’t expect to like the Hawks, or the Knicks, or the Raptors nearly as much as I do. I was shocked that I couldn’t find more wins for contenders like the Clippers or Bucks.

Most of these win projections don’t need much clarification: The Wolves, Mavericks, and Nuggets are 50 win teams. The Trailblazers, Wizards and Bulls are exceptionally mediocre. And the tankers are tanking [4].

This blog, of course, is not a suggestion that you should try to bet the penny stakes limits that cowardly sports betting sites will allow you to bet on season long win totals. But were you to be able to bet on such a thing, I’d be confident that you’d receive even more pennies back 6 months from now when those pennies will be worth even less due to inflation. But I digress.

The Good

Last year, the Pelicans had a neutral luck adjusted point differential without the best player on their team. Everyone on the roster except CJ McCollum is on the positive side of the aging curve, and one of the back end of the rotation guys is bound to pop as a positive contributor. Zion was the hardest player to project. Without a full season last year, do I give him a two year bump on his aging curve? Or should I assume he’s just been eating jambalaya and won’t improve at all without time on court? Luckily, Vegas is basically projecting them as if Zion won’t play, and I’ve got him projected for 2/3 of the season or so.

Though I don’t see the Raptors winning a playoff series with this version of the team, they’re one of the youngest teams in the league, play hard, and are well coached. They get very thin after their top 9– luckily Nick Nurse wouldn’t dream of going that deep into his bench! This is a 50 win team, with some upside from a larger than expected improvement from Achiuwa or Barnes.

The 76ers are no longer a niche pick to challenge for the top seed in the East. They’re good, deep, and not as reliant as you think on having Harden be an All-NBA level player. I have my reservations in the playoffs, but in the regular season this team will be a dynamo, provided the big fella stays healthy.

The Hawks, Knicks, and Kings are the three teams that surprised me the most. I don’t think the Hawks have much chance to win a playoff series, and I don’t think the Knicks or the Kings have much chance of making the playoffs, but the math is the math: they’re decent teams that go 10 deep. And Immanuel Quickley is one of the most underrated players in the league [5].

I have to write a quick paragraph about the Magic, the darlings of my basketball heart. Given my love for Franz Wagner, I tried as hard as possible to be conservative on my projections for this team. But I still like the over. While they're lacking the future resources of some of the other rebuilds, it's noteworthy how much better the Magic are right now than the Pistons, Rockets, and Thunder. Thankfully it looks like Jalen Suggs will be back for the start of the regular season, as a serious injury would have been a nightmare. If you moved 1000 of his minutes to RJ Hampton (one of the worst players in the league), it would be worth two wins!

The Bad

I would put the Bucks in my top tier of contenders this season (along with the Clippers and Celtics), but the regular season numbers just don’t show how the Bucks should be projected to win over 50 games. Wesley Matthews, George Hill, and Serge Ibaka are ready for taxidermy; Jevon Carter, Jordan Nwora, and MarJon Beauchamp are deer in the headlights. They may get a buyout guy, and Bud’s regular season preparation is as good as anyone’s, but even squinting only gets them to 50, and I see their line above that in most places.

The Rockets are part of an interesting phenomenon. When you project some of these particularly young teams who have a half dozen guys that are barely old enough to drink, a few of them will have outsized results, and those players will earn more playing time. But outliers are outliers because they’re outliers, and there isn’t much I can do beyond try to capture that by giving Jalen Green a little more of a bump than is expected by his aging curve, or pretending Josh Christopher might be able to play merely replacement level basketball instead of league worst type basketball. My raw win total on them without those adjustments was only 16! I do think with the rubberband effect along the poles this is a much more reasonable win projection, but I still don’t see how they get to the mid 20s (or why they'd be motivated to do so).

There’s actually a little bit of that same phenomenon happening with the Warriors. In my opinion, the Warriors young players aren’t particularly interesting. And despite that, one of them will likely be better than expected, and that one will earn rotation minutes alongside what’s left of the Death Lineup. The previous year’s NBA champion has gone under in 5 of the last 7 82 game seasons [6]. And teams with a win projection over 50 have gone under ⅔ of the time in the same sample (though teams with a low win projection have performed about the same as the field).

The Ugly

When incentive structures diverge, it gets hard to model. Some of these teams start this season with the mindset that a loss is a win[7]. It’s difficult to figure out the lengths that a team like the Jazz or Thunder might go to to get Wemby or Scoot. Front offices are probably already having serious conversations about the most damaging young players who you could pretend were prospects but would actually only be on the roster to drive the tank.

The Lakers are another such team whose win projection could best be described as a shrug emoji. They’re a .500 team as constructed, but a trade of Russell Westbrook for Buddy Heild and Myles Turner is worth about 4 wins, which is almost exactly their present Vegas total.

This exercise taught me the value of depth. It showed that where there are serious drop offs in rotations, injuries will provide the projections more fragility. The Lakers, Cavs, Raptors, and Pelicans might all have Vegas totals in the 40s, but they get there with drastically different roster composition. Although fit is overblown, I did dock situations I thought would have diminishing returns in Cleveland, Minnesota, and Atlanta. At the same time I'm also ecstatic to see those situations play themselves out, because there's magic behind the chemistry that goes into winning basketball games, and I'm here for it. Let us NBA.

Atlantic		Central		Southeast
Celtics	56.4	Bucks	48.4	Heat	49
76ers	55	Cavaliers	46.6	Hawks	48.4
Raptors	49.9	Bulls	40.4	Hornets	33.3
Nets	48.1	Pacers	25.6	Wizards	33.2
Knicks	41.3	Pistons	25.1	Magic	30.4
Northwest		Pacific		Southwest
Nuggets	50.8	Clippers	50.3	Mavericks	49.1
TWolves	49.8	Warriors	49.3	Pelicans	49
Trail Blazers	38.9	Suns	49	Grizzlies	48.4
Jazz	26.1	Lakers	41	Spurs	23.6
Thunder	19.2	Kings	36	Rockets	18.1

[1] I’m happy betting a lot on ones that I’m confident in, but don’t mind me if I don’t want to bet on team that has a lot of variance coming into the season like the Lakers or Jazz, or one where I’m only a win or two off the market.

[2] Insofar as I have a job

[3] Machine > man, but if machine doesn’t have to go it alone, it’s even more powerful.

[4] This was, of course, the toughest (and stupidest) part of the exercise.

[5] Hoping to do a deeper dive on IQ early in the season.

[6] And also 5 of the last 6 seasons, but I looked at the last 7 full seasons in my dataset and I wasn’t about to cherrypick for clicks.

[7] As they should be, I estimate about ¼ of the league should be tanking, but only ⅙ of the league is. I mean seriously, you really think Charlotte, Portland, and Washington are kidding anyone? It’s embarrassing, they have nothing on the roster that suggests the ability to win a playoff series in the next decade.

Big Bored

A lot of your favorite Twitter follows are probably posting their final big boards right about now. It’s fun, it’s draft szn! But I hate big boards, they’re an exercise in futility. Since we aren’t able to clone ourselves, no one (including some high level decision makers) is really watching a dozen college games of the dozen players available in their draft range, let alone the dozens of other prospects that might be available if they were to trade up. And no one has access to medicals, or has interviewed anyone, or has talked to a prospects’ high school teacher. So instead of doing some incredibly thin slicing, I’d rather ruminate about some macro draft thoughts.

In the NBA Finals, over two-thirds of the minutes were played by players who were drafted by the Warriors (65.7%) or the Celtics (73.1%). Although the success of those organizations’ drafting and player development may be outliers, the draft is a demonstrably important team building tool.

You’ll hear people suggest that NBA draft picks, especially late first round picks, aren’t worth much. They say these things when they’re throwing the kitchen sink in for Jrue Holiday or Nikola Vucevic. Although there’s some truth to it, it’s largely indicative of the kind of short term, win-now thinking that plagues front office personnel on 3 year contracts, or every team Lebron has been on in the last decade, or the Sacramento Kings.

What the draft is really: high variance. Whoever trades up for Jaden Ivey will likely expect stardom, but mid lottery picks are only about a flip to become long term starters, and ~25% to be All Stars (I haven’t seen an updated draft value chart lately, would appreciate a link in the comments if there have been fresh posts!). The reality is that Jaden Ivey’s median outcome is closer to Jordan Clarkson than Ja Morant, but hopium is a helluva drug (and readily available in NBA front offices in late June).

From a front office point of view, I’d much rather look at the draft as an opportunity to combine three resources to amplify returns. First, to select a player better than the market; second, to use player development to make that player the best version of himself; and third, to use restricted free agency to squeeze a valuable second contract[1]. For the homegrown players in the Finals, it’s tough to argue that each front office hasn’t done all of those things impressively.

Seven years ago I did a “big board” on the 2015 draft. I wrote this about my draft philosophy and what was important to winning basketball games:

(1) versatility and ability to guard multiple positions (and not be targeted on D as we see in the playoffs), (2) ability to hit top X% outcome and be a transcendent player (I believe that everyone, including this board, still underrates outlier outcomes in EV-- i'd take a top 5 player 1/10th as often as the 50th best player every time, because of how the salary cap works), (3) elite skills (I believe a team is more made up of skills than players)

While I remain committed to points (1) and (3), I’ve changed my tune a bit on (2). Of course upside is still important, but I’ve come to think it’s a bit easier to peg median outcome and still have an understanding of where the long tail lies. The draft is an extremely high variance affair, but you don’t have to draft bad players in the name of upside (see: Kuminga, Jonathan).

Another thing I’ve noticed is that a person’s draft valuation is directly correlated with their player valuation. Maybe you thought Jalen Green was an easy choice over Jalen Suggs. Maybe you also think Zach Lavine is better than Jrue Holiday. Now we’re having a totally different conversation, and we’ll have to agree to disagree.

The worst thing about big boards, and player rankings for that matter, is that they’re ranked linearly. Poor human beings and our limited spatial abilities that allow us to only read from top to bottom. If I ever actually had decision making power I’d insist that my front office rank in tiers, and that all tiers be randomized. Of course the quality of prospects is a pyramid, not a list; this is the same way that talent in the NBA (and many other sports) is distributed. Preferring something more specific overrates human beings’ ability to evaluate basketball talent.

There is *no way* that you have a real opinion on whether Kendell Brown or Christian Braun (also pronounced “Brown”!) is a better NBA prospect[2]. They’re both in your fifth tier and you watched a 5 minute highlight reel of Brown standing in the corner, and then maybe you saw Braun get hot in the national championship game. They’re totally different ages. They’re totally different player types. And one of them is bound to get drafted by the Knicks and sign his second contract in Europe.

In the end, the draft is a petri dish of human cognitive biases. Remember that although there are infinite ranges of outcomes in the multiverse, in every one of them NBA front offices get to take credit for having the player that turned out good higher than consensus, while never taking blame for having the player that turned out bad higher than consensus.

[1] Restricted free agency is incredibly rigged for the incumbent team, and yet few teams use it to their advantage as much as they could. Most of this is probably pressure from agents to extend, with the carrot being their clients' happiness. I'd push the envelope in most cases, especially with the downside risk of a serious injury (re)appearing in year 4 (see: Porter Jr., Michael).

[2] To be clear, there are some people out there that have a good opinion on which of these prospects is likely to become a better NBA player, it’s just not anyone that reads this blog.

Wagner's Law

The 2021 rookie class is an outlier. It’ll have multi-time All-Stars, All-NBAers, and maybe even a couple Hall of Famers. But one of the reasons it's so good has been decidedly overlooked. It’s true, Cade Cunningham and Jalen Green are exciting, though overrated, prospects. And it’s also true, Evan Mobley is a generational talent, the quality of top overall pick that we only get every few years . What if I told you that if your goal was to win NBA basketball games right now, the best rookie this season has been none of those players?

Those of you familiar with Gödel, Escher, Bach[1] will know Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter's Law.

Today I present to you Wagner’s Law: No matter how highly you rate Franz Wagner, you will always underrate him, even when you take into account Wagner’s Law.

It’s relatively hard to figure out what wins basketball games. We have some reasonable ideas, but such a large portion of the game takes place without the ball. And this is where Franz is so good: moving into threatening space on offense, keeping the opponent out of it on defense. He’s already an incredible cutter, despite playing on a team without a positive passer at any position (except himself!).

One way to be threatening on offense is to move with the level of the ball and make the defenses' eyes be in two places at once.

And though his best skill is team defense, individually he snuffs out everything before it happens, and has the lateral quickness to stay with just about anyone

The video clips of Franz trying to pull a quarter point out of thin air or saving a tenth of a point with his footwork or gapping are nice, but the only way to truly approximate value for an NBA player is with a large sample. So for trying to figure out how good Franz has been, I looked around at a few one number metrics to get a better idea of the best rookies this year.

EPM is reliable one number metric that includes shot luck. There are only 3 rookies who are positive in this metric: Franz +1.2 (80th percentile), Evan Mobley +1, and Herb Jones at +0.5.

I don't usually agree with 538's basketball coverage, but they have access to the Second Spectrum tracking data. RAPTOR seems like their most referenced one number metric. Among all rookies, Franz is behind only Herb Jones.

And it’s reasonable to think that you don’t need black box sorcery to understand why he's been so good. His team is almost 10 points better with him on the court, playing something like league average defense luck adjusted. If that doesn’t sound impressive to you, you haven’t looked at the Magic roster lately.

Go looking for more stats, and you’ll be more surprised: He got to the free throw line better than Cade, despite less usage. And he’s one of the best free throw shooters in the league, shooting a hair better than Jayson Tatum from the line at 86.3%. How do you think that bodes for his future from behind the arc?[2]

If you scouted Franz, you knew he’d be this type of player. A connector. But what I wasn’t ready for is the versatility of his shot, his finishing footwork, and how good the passing already is at this level.

He's an exceptional passer who should've been tasked with even more creation responsibility for the Magic.

But this is the kind of stuff that'll make him an All Star, the footwork and patience to create shots near the rim (he shot a high volume from floater range per CleaningTheGlass, with above league average efficiency).

And here's The Full Franz Experience: starting with how ready he is in the gap, and finishing with a nifty pass.

His strength lies in his processing speed, the angles he takes, the efficiency he plays with. For those reasons, he will never be overrated. Among the included benefits to this are a discount on his next contract (see: Anunoby, OG; Bridges, Mikal).

With a little more on court time, Evan Mobley will be the far superior NBA player. When Cade Cunningham learns to dribble, or Jalen Suggs learns to shoot, or Jalen Green grows 3 inches and gains 25 pounds, lookout.[3] But by any reliable metric Franz Wagner is a league average starter right now. He's one of the top 100 or so basketball players in the world, at age 20.6, on a roster that isn’t doing him any favors.

I don’t blame you for not watching the Magic, you shouldn’t watch the Magic. They’re an awful team owned by a not particularly great bunch of people (even for owners) in a crappy city who likely chose the wrong head coach. But one day many years from now they’ll be in the Play-In, and on that day you’ll remember Wagner’s Law.

[1] Of course I haven’t read Gödel, Escher, Bach, but I’m cool enough to know people who have.

[2] I don't feel the need to say more about his shooting; at present you can't leave him open, in the future it may be a five alarm fire if you do.

[3]There are a lot of Raptors slappies out there, so I have to at least mention Scottie Barnes: Offensive rebounds don’t scale particularly well, and Scottie’s numbers are all buoyed by the Raptors sending more people to the offensive glass than any team in the modern era. The Raptors are four points better with Scottie off the floor, some of that is capable backups, a lot of that is bad footwork and his constant insistence on impersonating a traffic cone on defense. But I digress.