Nobody saw this kind of season coming. After an injury-riddled, dispiriting 15-win campaign the year before, Vegas had them pegged for 24 wins. This was going to be a slow build up from the bottom. And yet, on the back of a hyper-aggressive, swarming defense, the 2014-15 Bucks saw a massive improvement: from 15 to 41 wins and a feisty first-round playoff exit. They did it while withstanding a season-ending injury to #2 overall pick Jabari Parker, the sudden departure of defensive anchor Larry Sanders, and the midseason trade of starting PG Brandon Knight. They had a new coach, a young core and seemed like a potential future powerhouse. And it all started with that defense, a defense that ranked 4th overall, and 1st in forced turnovers (on a per-possession basis). Their offense lagged behind, ranking 24th, but that seemed like just a matter of time as the core grew and gelled. And if they could put both sides of the ball together — watch out.

Fast forward two years and the Bucks have, indeed, learned how to score: if their 2014 incarnation had sported an offense bordering on the top 10, including an eFG% similar to that of the Spurs and the Clippers, they would have been an elite team. But the defense hasn’t come with it. The contours remain the same — the same aggressive style, the same reliance on turnovers and quickness and length — but the results have not held up. Despite an initial uptick after the February return of Khris Middleton, Milwaukee sits squarely in the bottom half of the league defensively, after finishing last year’s disappointing season ranked 23rd.

Yet it’s not the results themselves that are so fascinating. Rather, it’s how those results have been achieved: with a unique defensive style that runs counter to many of the league’s trends. A style that, as you dig into it, reveals many of the underlying strategic tensions at the heart of the game itself.

Fear the Deer?

It started in Brooklyn. Jason Kidd was a first-year head coach, in charge of a team with high expectations. That offseason the Nets had pushed their chips into the middle of the table, trading a bushel of draft picks for Kevin Garnett and Paul Pierce. But the early results did not meet those expectations. When Brook Lopez went down in late December the Nets were 9-17, with the 29th ranked defense. Earlier in the season the Nets had amped up their defensive scheme to a more aggressive style, but the Lopez injury allowed them to fully lean into the change. After a little experimenting they decided to mostly play small, and from that point to the end of the season ranked 1st in forced turnovers and 8th overall on defense. When Kidd went to Milwaukee that offseason, he and defensive coordinator Sean Sweeney implemented the same scheme, the scheme that was responsible for the Bucks’ surprising 2014-15 season.

Lots of teams play small, though, and many play aggressive. What is it that’s so unique about this style of defense? It’s the extreme to which the Bucks take it. Milwaukee’s defense mostly comes down to this: either they’re going to frustrate you and turn you over, or they’re going to give up a shot from a high percentage location. There’s generally not much in between. In each of the three seasons Kidd has coached the Bucks, opponents have attempted the fewest midrange shots in the league against them¹. In his first year they were 27th in preventing shots at the rim and 29th in preventing threes; last year they were 30th and 27th respectively; this year they are 29th and 24th.

Despite giving up so many layups and threes, though, Milwaukee still ranked 9th in overall eFG% defense in 2014-15. They were able to hit that mark by forcing misses wherever their opponents shot: they ranked 7th in FG% defense at the rim, 7th in FG% defense in midrange and 8th in 3PT% defense. Because of the profile of shots they allowed, however, their margin of error was slim. Last season they ranked 13th in FG% defense at the rim, 15th from midrange and 16th from three and ended with the 19th ranked eFG% defense. This season has been similar, and so they currently clock in at 22nd.

What are the Bucks doing to allow these types of shots? This isn’t a result of laziness, poor defensive personnel or poor execution. In fact, the Bucks execute this style of defense, which requires an enormous amount of coordination, effort, and precision, remarkably well. That’s a testament to their coaching. But the scheme itself means that no matter how well it’s executed you will see these types of results. To understand why, let’s take a look at some video, starting with Milwaukee’s pick-and-roll (PNR) defense.

I’ve picked out two clips of PNRs: one from the Bucks and one from a team who plays the opposite style, the Utah Jazz². I deliberately picked some of the more extreme examples of their two styles of defense. The Bucks don’t always play this aggressively and the Jazz don’t always play this conservatively, but the contrast illustrates the point nicely. Pay attention to how high the screener defender starts, how aggressively he commits to the ball and the how much the weak side defenders leave their men.

Both of these styles are executed the way their teams want them to. The Bucks have their screener defender start high, at the level of the screen, then swarm the ball handler and pull the weak side “I” (so-called because of the “I” the defenders form in this alignment) way over to take away all of his options. Reggie Jackson has no choice but to retreat and swing the ball to Jon Leuer who is way beyond the arc, and by the time the ball is back in scoring position the Bucks are all recovered. And if Jackson doesn’t secure the ball or tries a risky pass to a teammate, it’s a turnover.

Utah, meanwhile, is clearly trying to take away Houston’s threes (and they were very effective doing so in this particular game). Usually their weak side will stray a little further from the perimeter and play a little closer to the paint to help Gobert — without that they risk giving up an easy layup on a pass to the roller. By playing it this way the Jazz are putting a lot of pressure on Gobert to be able to keep the ball handler guessing and effectively defend both the ball handler and the roller, as he does in this clip. Notice that by playing Gobert this way they can still attempt to protect the basket without the weak side defenders having to leave their men much, which takes away kickouts to threes. In exchange they put much less pressure on the ball and concede decent looks in midrange.

The responsibilities of the weak side defenders have a lot to do with the forced turnover and midrange numbers we’re looking at. And, as seen in these clips, those responsibilities are in part determined by what a team chooses to do with its screener defender. How high up does he start and how aggressively does he commit to the ball? But what happens with the two defenders of the PNR itself is not the full story. Teams still have a menu of options for what to do with their weak side defenders regardless of what the screener defender does, and those choices make a big difference.

Milwaukee is very aggressive with their weak side defenders, and that can lead to the turnovers they want, as they start in, force a pass, and then suddenly close into the passing lane to get the steal:

The problem is that this style of defense doesn’t leave much room for error. Make a mistake and you give up a layup or a three. And there are a lot of ways to make mistakes:

Watch those first two clips again. In the first clip Middleton is maybe a step too far in, a beat too late changing direction, and it ends up in an open three. In the second clip he’s maybe a step too far out, a beat too late getting to the paint and it ends up in a dunk. That’s life in the NBA — 18 inches and half a second is the difference between a turnover and a bucket — and it’s why all of these small decisions can add up to something big.

But it also presents a different way of thinking about NBA schemes. It’s not just about what happens when those schemes are executed perfectly, but about how likely they are to be executed perfectly. And as well as the Bucks execute most of the time, this scheme demands such precision that there are still frequent slip-ups.

PNR is a good place to start to understand Milwaukee’s defense. But this same style is on display against most basketball actions. In the post they’ll often try to deny the entry pass with a full front. This requires bringing over a weak side defender into the paint to protect against the lob over top. And if the ball is entered to the post, even with the defender recovered, they’ll still often bring someone over to the strong side to deter the post player attacking. This makes it much harder for teams to score directly out of post ups, and, like the Bucks’ PNR defense, puts pressure on the ball. But, because it means multiple players are defending the ball, it opens up the possibility for passes to open teammates.

It’s a similar story against isolations. The Bucks will frequently flood the strong side, having a player pre-rotate and guard the entrance to the paint instead of their man to deter the drive. Once again, this is a tradeoff between stopping the player with the ball and opening up teammates.

Lastly, against off ball screens the Bucks tend to have the screener defender play the ball more aggressively than other teams. Once again, this puts pressure on the ball but potentially opens up passes.

Let’s take a look:

Again, this isn’t how the Bucks always defend these actions — they’ll change things up based on personnel and situation. But it’s a style they play more often than most teams, and in total it’s responsible for the high-turnover, low-midrange results.

Shot Selection vs. Turnovers

The Bucks, then, are on one end of the spectrum for aggressiveness. As we just saw, a team like the Jazz (6th in midrange shots forced, 25th in turnovers forced) are close to the other end. And really, we can think of many defensive strategic decisions as belonging to some place on this spectrum.

Let’s start broadly with something you don’t see too often on the NBA level: full court pressure. An amped up, trapping press is probably the extreme version of an aggressive defense. It’s a gamble that you can create enough turnovers to make up for the times where your press is broken and you are at a numbers disadvantage.

In the halfcourt, one of the biggest determining factors of your style of defense (at least in the NBA) is how you defend the PNR. How aggressively does your screener defender play the ball handler? Is he trapping, is he hard showing, is he playing at the level of the screen and committing to the ball handler, or is he playing further back and retreating and absorbing? What are the responsibilities of the three defenders not directly involved in the PNR? How far are they coming off their men? Are they playing the passing lane for steals, or sticking closer to shooters?

We can continue this way through the other typical basketball actions. Are you fully fronting the post? Double teaming? Or just playing straight behind? Are you pre-rotating to overload the strong side against isos? How far out does your screener defender show against off ball screens? All of these options are somewhere between the most aggressive you could play and the most conservative, and they do a lot to determine where on this aggressiveness spectrum you land.

What might not be obvious at first, but what studying the Bucks’ defense makes clear, is that these decisions entail a real tradeoff between generating turnovers and forcing worse shots. It’s very hard to do both of these things at the same time³. This is a fundamental aspect of the game of basketball. Defenses try to take away the best shots — if you want to get high efficiency shots you generally have to work for them, and that means risk. Difficult passes. Forays into the paint through the arms of defenders. Passing up bad shots early in the shot clock to extend possessions and look for something better. You can reduce your risk, but only if you’re willing to settle for lower value shots.

The James Harden-era Rockets have been a great example of this tradeoff⁴. In these last five years they have tested the limits of how few midrange shots a team can attempt, ranking 1st in that category in each season. Yet that has come at a price: coughing the ball up. This year is actually the first year in this five year stretch that they have been outside the bottom 3 in taking care of the ball. An opposite example is the Doug Collins-coached Sixers. In his three years at the helm, Philadelphia ranked in the top 2 in avoiding turnovers each season — and in the top 2 in attempting midrange shots as well.

These teams illustrate the general trend. In the last 10 seasons (including this season), of the teams like the Harden Rockets who ranked in the top 10 in avoiding midrange shots, almost half were in the top 10 in turnovers committed and a full quarter of them were in the top 5. Meanwhile, only one-fifth of those teams ranked in the top 10 in taking care of the ball and only 9 were in the top 5. Getting good shots comes with risk.

It’s the same thing on the defensive end, perhaps to an even greater degree. The more pressure you employ the more compromised you are if your opponent beats it — you’re going to give up something good. In the last 10 seasons, of the teams like the Bucks who ranked in the top 10 in forcing turnovers, only one-sixth ranked in the top 10 in forcing midrange shots, and only 3 ranked in the top 5⁵. On the other hand, of the teams who ranked in the bottom 10 in forcing turnovers, over half ranked in the top 10 in forcing midrange shots and 30 ranked in the top 5. This is especially surprising: you’d think teams that were that bad at forcing turnovers were bad defenses, that not forcing turnovers might go hand-in-hand with allowing good shots. But it actually turns out that more often than not, it’s a result of defensive style. A style that, as the league has increased its focus on the value of layups and threes, more and more teams are adopting.

Bucking the Trend

Like many aspects of the modern NBA, the Spurs got to it early. In the 20 years since Gregg Popovich and Tim Duncan first teamed up, the Spurs have ranked in the top 5 on defense in all but three of those years (and were still in the top 11 in those three “off” years). In each of those 20 seasons they have not once ranked outside the top 6 in forcing midrange shots, but in 14 of those 20 years they ranked outside the top half of the league in forcing turnovers.

Jeff Van Gundy was hired as head coach of the Rockets in Yao Ming’s second year. For the 4 seasons that Van Gundy coached Yao, Houston ranked in the top 7 on defense: never outside the top 4 in forcing midrange shots and never outside the bottom 8 in forcing turnovers. Van Gundy’s replacement, Rick Adelman, continued the style: the Rockets ranked in the top 5 in forcing midrange shots in each of the years Adelman coached as well (though the overall defense fell off once Yao went down with injuries). In fact, Adelman’s adherence to this general philosophy pre-dates his time in Houston. In the 16 seasons for which we have shot location data and in which Adelman was a head coach, only once (his final season coaching) did any of his teams rank outside the top 10 in forcing midrange shots.

Tom Thibodeau was an assistant on Van Gundy’s Houston staff. After Van Gundy was fired, Thibodeau went to Boston and teamed with Kevin Garnett to create one of the rare defenses that forced midrange shots (top 10 in each of Thibodeau’s three seasons there) and forced turnovers (ditto). In 2010, Thibodeau got his first head coaching gig and built the Chicago Bulls into an elite defense around, you guessed it, forcing midrange shots: a top 5 defense in each of Thibodeau’s four seasons in Chicago, they ranked in the top 5 in forcing midrange shots each year but in the bottom 10 in forcing turnovers in two of those seasons.

Steve Clifford, meanwhile, was also an assistant on those Van Gundy Houston teams. When Clifford took the reigns in Charlotte four years ago they had come off back-to-back seasons as the worst defensive team in the league. The next two years they ranked in the top 10 on defense — and yes, top 10 in forcing midrange shots, bottom 10 in forcing turnovers both years⁶.

Clifford was also an assistant in Orlando with Jeff Van Gundy’s brother, Stan, who, while all of this was going on, was coaching teams of the same style. In his 10 full seasons as head coach of the Heat, Magic, and Pistons, Stan’s teams have ranked in the top 5 in forcing midrange shots in 9 of those years, and in the bottom 10 in forcing turnovers in 7. Sound familiar?

The Van Gundy brothers have something else in common: they both were assistants for Pat Riley (Jeff with the Knicks, Stan with the Heat). And though we don’t have shot location data for most of Riley’s years coaching, for the years we do, Riley’s teams mostly adhere to this same pattern: consistently forcing midrange shots without always forcing a high rate of turnovers.

The Riley-headed Heat, though, took a detour from this style after Shaq was traded during the 07-08 season. From 08-09 through 14-15 the Heat ranked in the top 10 in forcing turnovers in all but one year and were in the bottom 10 in forcing midrange shots in 5 of those 7 seasons, including in all 4 of the LeBron-era years. Those Heat teams used their speed and athleticism to fly around the court and wreak havoc, and, when fully engaged, devastated opponents. But after one season post-LeBron of trying the same thing without the same results (21st on defense), Erik Spoelstra has revamped Miami’s defensive style. Both this year and last the Heat have built a top 10 defense around (say it with me now) forcing midrange shots without forcing turnovers.

Not all of these teams had the same schemes to generate these results, there are other teams and coaches that worked to implement similar styles, and this all may sound a little like Charlie in the mailroom — but this history is important to highlight the trend. As this defensive style has had success and spread around the league, the Bucks’ style of defense has become increasingly uncommon.

Looking Ahead

The Bucks have played very well since Middleton’s return, going from wondering if they would make the playoffs to jockeying for the 5th seed in the span of a few weeks. But a late-season playoff push shouldn’t distract them from the important, long-term question of whether this is the style of defense to build the team around.

Perhaps the defensive dropoff we’ve seen in these last two years has been a result of injury and youth — maybe with more time and more health Milwaukee will be able to approach the results of 2014. Possibly it was personnel: replacing Larry Sanders’ and Zaza Pachulia’s minutes (and speeches) with Greg Monroe’s would likely result in a decline no matter what style is played⁷. If that’s the case then just slotting the right pieces in will get them back to where they need to be.

But it’s also possible that their 2014 results were not sustainable. Maybe to play this style of defense well everything has to click. You have to have the right players, health, and maybe get a little lucky with opponents missing jumpers. Perhaps a more conservative style doesn’t need quite as much to go right to succeed. And indeed, this seems to be the case, at least in recent years: over the last 5 seasons, of the teams who ranked in the top half of the league defensively and played an aggressive style of defense⁸, on average their defensive rank fell by almost 7 spots the next season. Meanwhile, of the teams who ranked in the top half of the league defensively but played a conservative style of defense⁹, on average their defensive rank fell by 2 spots the next season.¹⁰

Interestingly, and maybe most importantly for this time of year, there is some suggestion that teams that play the Bucks’ style don’t do as well in the playoffs. This makes sense: the advantages of aggressiveness are closely intertwined with the advantages of talent. Think about the different levels of basketball. Full court pressure and trapping defenses are much more prevalent in the NCAA, high school and AAU, where skill levels are lower and talent is less evenly distributed.

This would suggest that it would be hard to gain an edge with an aggressive style of play in the playoffs. In an environment where you’re playing teams closer to you in skill and athleticism and when opponents have the ability to prepare and scheme for you, your flaws might be more exposed¹¹. The data seems to support this conclusion: while the relationship isn’t the most powerful, playoff defenses have had more success in general and against top offenses in particular when playing a more conservative style rather than a more aggressive one¹².

The last example in the coaching breakdown, of Spoelstra’s change of the Heat’s defensive style, illustrates something that shouldn’t be overlooked. The Heat played one style of defense when they were small with LeBron and Bosh, and another now that they have Whiteside. Style of defense, like any scheme, is inextricably tied to personnel. Pop had Duncan; Jeff Van Gundy and Adelman had Yao; Thibs had KG and Noah; Stan had Shaq, Dwight and Drummond; Riley in Miami had Zo. The roster absolutely matters. A Bucks team that wants to play with Giannis and Jabari Parker at the forward positions might not be able to build a wall in the paint the same way a team with more traditional big men might, and could be better suited for a style of defense based on length and activity. It’s entirely possible that, as currently constituted, this defense is optimal for Milwaukee.

And so, as the Bucks head into the playoffs we’ll watch and wonder if this is the furthest they can go playing this style. Can Milwaukee get back to their 2014-15 level, or was that season a mirage? Do they just need to tweak their roster, or do they need to overhaul their scheme? Do they have to change how they get stops in order to keep going?