Cage-Free Centers vs. Incents

Plus an interview with a sharp web-scraping addict

This is The F5, a twice-weekly newsletter that helps readers be less wrong about the NBA. If you enjoy it, think its interesting, or want to learn how to make the visualizations you find in this newsletter (below), please consider a paid subscription!

---

An interesting trend in recent years has been the use of defensive roamers, or Centers that “guard” the opposing team’s worst shooter. It’s hard to say when the trend started, but one of the early flashbulb examples is Game 4 of the 2015 Western Conference Semifinals. Trailing 2-1 to the Memphis Grizzlies, the Golden State Warriors started the game with Andrew Bogut matched up on Tony Allen. The Warriors went on to defat the Grizzlies in six games and later won the NBA Championship, in part at least, because they left Allen open for a reason.

The idea behind the strategy is simple: reluctant shooters aren’t worth worrying about so leave them open and have your rim protector defend the basket against the remaining four offensive players. Occasionally, the strategy will backfire. Like when Terrance Mann has the best game of his life. But generally it’s a reliable way to improve your defense by shrinking the court and funneling the ball into the hands of the least dangerous offensive player possession after possession. Shots at the rim become harder and shots away from the rim become more likely to be attempted by worse shooters.

Ime Udoka deployed Robert Williams in this role to great success during the Celtics 2021-22 season. The Celtics had the league’s best defense that year and had Williams not gotten injured he might have won Defensive Player of the Year. Instead, Marcus Smart was given the award by default.

It’s no surprise then that Udoka has been using Alperen Sengun in similar ways this season with equally impressive results. The Rockets are 3rd in Defensive Rating with Sengun doing his best Time Lord impression.

Despite not being the same kind of athletic shot blocker that Williams is, Sengun still has the size and understanding of where to be to bother opposing would-be finishers at the rim.

In the clip above, Sengun is matched up against Kris Dunn who is bringing the ball up the court on the Clippers’ opening offensive possession. But Sengun is not really defending Dunn. He’s defending the paint.

The play unfolds as the Rockets had hoped.

The ball gets entered into Norm Powell, but Sengun is already in position to contest Powell’s shot attempt. Instead of shooting over the top of Sengun, Powell kicks it to an open Dunn who does what Dunn does and bricks it.

Everything that happens on the possession can be traced back to Sengun matching up with Dunn who is a Guard.

This season, Sengun has matched up with Guards 30 percent of the time when he’s on defense — nearly twice as often as he did last season. In fact, this season, he’s matched up with opposing Guards more often than opposing Centers.

The result is that Sengun is closer to the basket and contesting more shots than ever before. In raw numbers, he’s 3rd in the NBA in total shots contested within six feet of the basket. But more relevant to Houston’s defensive performance is that Sengun is allowing opponents to shoot just 53.6% on shots within six feet of the basket — that’s 3rd behind Victor Wembanyama (38.8%) and Chet Holmgren (43.6%)¹ and a sizeable improvement over his own numbers last season (59.6%). That’s helped Houston jump up to 2nd in Opponent Rim FG% this season — up from 11th last season.

Sengun’s story is not unique. All around the league, coaches are toying with their matchups to allow their Centers to roam around the court like cage-free chickens.

The chart above shows the percentage of matchup minutes that Centers across the league have spent defending opposing Guards. The trend line is clear: more and more coaches are putting their Centers on the Tony Allen’s and Kris Dunn’s of the NBA.

[I want to pause here and entertain the possibility that the reason the trend line is going up is due to the way the NBA classifies positions. For example, if a player is considered a Center today but would have been considered a Forward in the past, their matchups on Guards wouldn’t show up on the chart in past seasons thus making the line go up. But from the best I can tell, the number of players the league has classified as Centers and Guards has more or less remained constant since 2017. In short, unless the NBA has meaningfully changed the way they classified Centers and Guards I think we can have some confidence in the takeaway in the above chart.]

It’s easy enough to identify who is most often being used as a Defensive Roaming Cage-Free Center. All we need to do is rank every Center by the percentage of their defensive Matchup Minutes (MM) spent on opposing Guards, which you can find in the table below.

Jock Landale, the Rockets third-string Center, would rank 2nd on this list if not for falling just short of the minutes threshold. Meanwhile, the proximity of the two Mavericks Centers and the three Celtics Centers on this list further illustrates that this strategy is a coaching decision and not one dictated by a defender’s versatility.

After all, Bam Adebayo and Nic Claxton have reputations for being capable of defending one-through-five, but are matched up with opposing guards less frequently than most of their peers.

We can also use NBA matchup data to reveal the reverse. That is, we can look up who are the Tony Allens of today (other than Kris Dunn). The table below shows the Guards that have been defended by opposing Centers most often this season.

Unfortunately, this list doesn’t look quite right.

The problem is some of the Guards on this list are matched up with Centers against the defense’s will. For example, Steph Curry often gives Centers no choice but to switch on to him or risk an open three.

What we’re really interested in is players like Kris Dunn, or Guards that are involuntarily defended by Centers.

Let’s call them Incents.

To identify Incents, we can filter out anyone that qualifies for the Synergy pick and roll tables². Removing the pick and roll maestros from the table should leave us with Guards that are involuntarily defended by Centers most often.

Ben Simmons is an Incent since he rarely runs pick and rolls these days and gets guarded by Centers 43% of the time that he’s on offense. Other than Kevin Huerter and Isaiah Joe, everyone else on this list profiles as a 3-and-D wing that’s light on the 3. To steal a bit from Defector’s Patrick Redford, just about everyone on this table is Tony Allen with Wi-Fi.

Note that I’ve added a column to this table showing how often each player was defended by Centers last season. With few exceptions, the frequency with which each player has been defended by Centers has increased — further suggesting that this is a tactic that more coaches around the league are embracing.

For offenses, the counter to this strategy is to play someone that defenses can’t leave open. The problem is that those players often come with their own set of weaknesses that are even more exploitable than the Incents above. As a result, teams will live and die by whether Incents like Josh Hart and Peyton Watson can take and make enough open shots to justify having them on the court.

---

A Refreshing Interview With Canzhi Ye

I first met Canzhi Ye at Vegas Summer League a few years ago. He had just handed a mutual friend of ours a winning ticket for a futures bet he had made on the eventual champion of the National Women’s Soccer League. Canzhi was flying out early and didn’t have time to cash the ticket and asked our mutual to do it for him and Venmo him the winnings.

Since then Canzhi and I have shared beers, dinners, and the same basketball court in pickup. He’s one of the sharpest people I know and I learn something new every time I talk to him. In fact, the inspiration for this week’s newsletter came from a text I received from Canzhi a few weeks ago.

I asked him about his interest in sports gambling, why it makes sense to bet on Preseason and Summer League games, and his favorite NBA draft analyst.

This Q&A has been edited for length and clarity.

---

F5: I’m not against sports gambling, but I think if it disappeared tomorrow nothing in my life would change. What do you get out of it?

CY: First, if you're into sports analytics and really care about the science behind building predictive models for a sports, I think there's no better scoreboard than the betting markets.

I've had this experience quite a few times: I'll build what I think is this sick model with novel feature engineering, only to see that I still fall short of the market odds' RMSE. A good dose of humility.

But it's just something that I've come to accept: wisdom of the crowds -- the crowd is actually very wise. For me, actually seeing it hit a lot deeper than just reading about this concept in the abstract. So after I thought I just did some sick modeling and was rudely told NO by the market, I go back and continually try to make improvements on the margins.

That mindset - the constant search for marginal improvements - probably translates really well to other areas, like working for a team.

Second, gambling is legit entertaining! In our circles, we value quality basketball analysis, but most people just want entertainment. We can rage about the Inside the NBA folks for not knowing ball, but they're popular for a reason.

No one is watching a random Heat vs Wizards game on a Tuesday, but if someone has a sick 5-leg SGP they cooked up, they're so much more likely to be invested in that game... and probably future ones too. Give away a little monetary EV and be entertained? Seems fair. The league needs eyeballs - this stuff brings eyeballs. 

You started off working for the Brooklyn Nets shortly after college. How did you go from someone that ostensibly wanted to work in the NBA to someone that bet on the NBA?

Say I build an amazing all-in-one player metric with all the super granular player tracking data that teams have and I find a handful of undervalued players. There's so much that has to happen between developing this insight and acting upon it to benefit the team.

There's organizational dynamics, convincing decision makers, and many things outside of my control. Perhaps I was just short sighted, but I thought that that I'd never be able to consistently substantially affect decision making.

On the other hand, if I build a good betting model, I can unilaterally act on the insights...and I can do it repeatedly. 

Are you worried at all that the league has been too quick to embrace sports betting. I don’t think it's crazy to think that the Jontay Porter scandal is just the tip of the iceberg. 

The Jontay scandal was pretty bad, but I’m not sure the NBA is embracing sports betting any more so than like the NFL or MLB.

I'm hopeful that Jontay being banned from the league forever will be enough of a deterrent for future players.

Honestly, legalized gambling probably led to him being caught a lot faster. They bet $80k to win $1M on DraftKings -- of course that's going to get flagged!

What are some things you see that media (or fans) consistently getting wrong about NBA betting?

The thing that comes to mind first is all the pre-season content about win totals. Content creators are just picking over or under without any consideration of the price -- over 46.5 -110 is not the same as over 46.5 -160! It's really not a big deal, but I don't find a ton of value in even doing a pick against a line at the end of a team preview anyway. 

I know sharp bettors take an outsized interest in International Basketball Tournaments (especially the group stages), Summer League, and Preseason games. What’s the deal with that?

Infrequently offered stuff most likely has worse pricing. Why is the pricing bad? Well whoever is making the line can't just go to inpredictable.com and manually adjust a point or two for injuries and get a pretty good number.

I don't know exactly how, for example, a FIBA World Cup line would get originated, but likely some trader will just look at some past results and recent results and throw something up. They could just totally miss that Gallinari is not in the Italian squad and that KAT suddenly is now playing for the Dominican Republic.

If you have this info and can do some back of the envelope league translations, you probably find yourself with an edge. Good luck doing that little work and having an edge on a regular season NBA game. 

Meanwhile, NBA futures seem like a fairly soft market. The fact that Jokic can go from 5th in MVP odds to 1st in less than a week suggests to me that something’s fundamentally off there.

Futures are probably soft, but I don't think that Jokic moving that much in week is necessarily the proof that futures are soft. Part of his odds moving so much was probably Chet getting hurt, reducing OKC's 1-seed odds. 

Futures would be soft because there's not a lot of price discovery -- there's not a lot of money being bet to fix a wrong price when it gets a little bit out of whack. If you bet a price that you think is 5% off, you're basically locking up your money for months to maybe exploit that edge. Meanwhile, you could just park that cash in a Wealthfront account for a guaranteed 4.5% APY.

Without giving away anything, are there any edges that you used to take advantage of that have dried up?

Every year in the playoffs, there's some discussion of the zig zag effect which is the home team down 2-0 playing better in the first half of game 3.

The theory here is that the team down 2-0 is playing as hard as they possibly can to not go down 3-0 in front of probably a big loud home crowd. This is well-known and makes sense...and the market prices it in. On average this is true, and the market tends to apply this adjustment pretty mechanically. But, there's probably cases where the theory of the zig zag effect does not apply and there might be value in fading that effect.

Are there any strategies (in game or otherwise) you think teams should be trying to exploit more than they already are?

Not an on-court strategy, but teams still do some funny things on draft night. They don't draft best player available enough, they trade up too much, and they don't respect consensus enough.

I think a lot of that comes from overestimating their player evaluation edge. I get that teams put a lot of effort into evaluating prospects and have a ton of data that the public doesn't know about. (Actual draft slot is predictive of career success beyond public consensus draft board) Even then, when a team trades up or reaches for a guy they're assuming a massive edge when in actuality on average it is much smaller. 

Dean On Draft -- Overrated, underrated, or properly rated as a draft analyst.

Hugely underrated by the people on twitter who kind of chased him off. It seems like he expresses his views like someone who just watches the box score. I think that comes off as not knowing ball to the eye-test twitter scouts people who are tweeting out a bunch of clips of like "omg 6'10 guy just threw a skip pass out of a ball screen - passing upside !!"

Also, many people have a hard time separating some of his disagreeable non-basketball takes from his draft analysis.

Ultimately, I don't think Dean misses the forest for the trees in his draft analysis though. Like who cares if he just looks at the box score (not saying he only does this, but a lot of his writing is based on box score stats).

He has a solid mental model of what winning basketball looks like in the box score. Actually quite useful to be able to have a "brain BPM" model lol

What’s one thing you can’t live without during the NBA season?

Dunc'd On Prime Podcast with Nate Duncan and Danny LeRoux.

I like how serious Nate and Danny are about covering everything going on in the league. It's direct and informative. There’s genuinely nothing quite like throwing on “Mock Trade Deadline” and knocking out a few chores. That they succeed despite Nate and Danny's awkward chemistry and Nate's annoying arrogance tells you just how good the content itself is!

1. Among players that have defended at least 100 shots within six feet of the basket.

2. Minimum of 10 min/game and 10 possessions per play type to qualify