Dr. Strangetriple or: How I Learned to Stop Worrying and Love the Three

In the spring of 2019, as the Bucks began their first and (so far) most successful playoff run with coach Mike Budenholzer at the helm (since 2001 too...), we started hearing the oft-used mantra “Let It Fly” more and more from the team. Bud brought it over from his Atlanta days; the players started using it in press conferences, the team produced an updated version of the early-2000s Light It Up video with it as the new refrain, slapped it on some t-shirts, and I’m sure there’s more.

It was obvious that Bud likes his players to shoot liberally from downtown, but he also didn’t mind if his opponents did either, as long as those threes came above-the-break. As we moved deeper into Year 2 of the Bud era, some of us began noticing how much an opponent’s prolificness from distance swung Milwaukee’s outcome at the buzzer. We usually point to a team’s 3P% in a given game, which is obviously a good indicator of winning vs. losing in today’s make-or-miss league, but 3PA and its associated differential is just as (if not more) important.

Piggybacking off Duhawk Steve’s recent and excellent FanPost (which I highly recommend if you find this sort of thing interesting), let’s set forth a few concepts that are supported by recent regular-season data. To paraphrase his findings:

1. For every 1% better your opponent shoots from three over an 82-game season, you win 4 fewer games, but this isn’t a strong correlation.
2. For every 10 additional 3PA you allow, you win 1 fewer game, but this correlation is very weak.
3. For every 1% better your opponent shoots from three, your defense allows 2 more points per 100 possessions, but this correlation is also quite weak.
4. For every 10 additional 3PA you allow, your defense is 6 points worse, which is the strongest correlation of these 4.

As he points out, some of these we naturally expect, particularly in regards to conversion. But here is where he noticed something odd: allowing fewer treys makes for a better defense but doesn’t mean more wins. This really sounds like the Bud-ball Bucks, who allow lots of 3PA but have still won lots of games. Meanwhile, allowing a higher 3P% doesn’t impact your defensive rating (DRtg) much, but it does impact how many games you win.

Like I did in an old FanPost, I took a look at the league’s 5 best and 5 worst teams by DRtg, opponent 3PA/game, opponent 3P%, and winning % in each season since 2017. These are the same seasons Duhawk Steve’s correlations are drawn from (note: lower rankings are better).

Though we’re only about a quarter-way through this season, we’re seeing a notable correlation between shutting down opponents beyond the arc and DRtg. Unlike in recent seasons, the top defenses aren’t giving up lots of 3PA, but their opponents still aren’t shooting well. This year makes the correlations suggested by Concepts 1 and 3 seem a lot stronger and matches common sense; teams whose opponents shoot worse from three generally win more games and have better defenses.

Opponent conversion rate is tough to control and, perhaps surprisingly, at least somewhat based on luck, however, and we know that impacting the volume at which your foe puts treys up is easier. So far, this year is making Concept 4 appear quite strong; teams allowing fewer threes have some of the best defenses, and the bad defenses “let them let it fly” more often. Concept 2 isn’t reflected as strongly with playoff contenders in Brooklyn, Dallas, and Portland giving up lots of 3PA, but if this is indeed the way the league is moving and these correlations become stronger, the best defenses will be the ones that curtail opponent attempts. Last season, unlike in other years, the best defenses were giving up lots of 3PA, but this looks like an aberration as top units from the two seasons prior usually limited the three-ball well. Bud’s Milwaukee-era defense exemplifies that aberration each year and is in some ways a polar opposite of Jason Kidd’s final defensive unit in 2017–18.

The Bucks’ DRtg has improved over the last week, and in their current five-game win streak, their opponents have attempted no more than 5.9 threes above their rolling per-game average. Visibly, the team is becoming more confident in its new defensive wrinkles, putting together several quarters of impressive play.

Though they’re still not a great switching team, opponents are becoming less trigger-happy. In the five games preceding their current streak (where the Bucks went 2-3) opponents shot between 3.4 and 17.2 triples above their average. But that regression came after five games (4-1) where the Bucks held opponents under those averages 4 times, so their progress as had its ups and downs. Overall this year, excluding the first two games, the Bucks are 6-2 when they hold teams to or under their 3PA/game mark and 14-5 when above it.

That seems like a step in the right direction, right? Not necessarily. In 2019–20, the Bucks only held teams below their rolling averages just 9 times (an 8-1 record), usually only barely. In 2018–19, there were 22 such games, but the Bucks actually lost 9 of them (13-9). That means in the other 124 games where the opponent went over—which happened 8 out of every 10 games—the Bucks were a stellar 48-16 and 47-13 each year, validating Concept 2.

Maybe this trend of tamping down opposing 3PA will continue, though it remains to be seen if it actually will translate to continued wins for the Bucks. As we’ve seen, giving up many 3PA has actually translated to lots of wins and lovely defensive numbers during the regular season, but neither of those seemed to matter in the postseason, where it certainly feels like that the Bucks are be affected more by the opponent 3PA than other teams.

Let’s see how teams fared in the same categories above during the small sample-sized playoffs, replacing winning percentage with teams’ series outcomes:

Unsurprisingly, there’s a lot of noise here. A lot of these numbers depend on matchups, seeding, opponents getting on hot streaks (“Van Vleet-ing”), and so on. The regular season correlations seem to apply even less. Still, a lower opponent 3P% sometimes leads to deep playoff runs and naturally makes for a better DRtg. Also like the regular season, giving up a lesser volume from downtown doesn’t seem to affect how much you win either, nor does it impact your DRtg.

Obviously, you don’t win titles by just having a great DRtg. Having a better DRtg than each conference champ since 2018 didn’t get the Bucks much, in part because the teams they lost to took and made a lot of threes. Both of last year’s conference champs are conspicuously absent from the rightmost table: the Lakers and Heat had the sixth and ninth-ranked playoff defenses respectively. Each team did at least one thing well on defense, though: no team gave up fewer 3PA than the champs in LA, and Miami’s 3P% against was .347, which ties for fourth. Each of the last two title-winners in Toronto and Golden State also had an enviable 3P% against.

So back to Duhawk Steve’s quandary. Your defensive rating will be better (in the playoffs too, it seems) regardless of how many threes go up or how many go in, but winning does depend on conversion rate, and fewer 3PA doesn’t mean more wins. This seems apt looking at both tables; while a high DRtg does make for lots of regular-season wins, it is not necessarily conducive to playoff wins, and neither is 3PA allowed. Unlike the regular season, though, opponent 3P% doesn’t correlate as strongly: the last three winners ranked 1st, 3rd, and 9th.

So if the way your defense handles triples doesn’t matter in the postseason, how is it that the Bucks have flamed out twice because of how opponents shot them? To answer that, I point to what I call net three-point attempts.

With a tip of the cap to prolific Brew Hoop FanPoster and commenter retired janitor, who likes to look at 3P% differential, this is just the differential between teams in 3PA, but I’ll take it game-by-game. In regular season matchups where the Bucks have attempted more threes than the opposition, the Bucks are 66-26 (.717) and 60-19 (.759) when the other side attempts more. This season, those numbers are 10-4 (.714) and 5-4 (.556) respectively. We’re still not very deep into the season, but a 20% worse record when their opponent outshoots them by volume—a scenario in which they have historically won more games—is something worth watching from here on out.

One positive trend to note is the frequency at which the Bucks are “out-volumizing” their foes. Since Bud took over, the Bucks won the net 3PA battle in 51.3% of their games, and that number is at 58.3% this year. During the current win streak, they only were exceeded in volume once. Over the past 15 games, they’re 8-2 when they at least equal their opponent’s 3PA and 3-2 when they don’t.

While it didn’t as matter as much in past regular seasons, perhaps it’s now more important the Bucks shoot more threes than the opposition. Doing so has been critical to their successes in the playoffs. Instead of individual games, let’s expand back to series. In their series losses to Toronto and Miami during the 2019 and 2020 playoffs, the Bucks net 3PA were -36 and -1 respectively. As you’d probably guess, their 3P% differential/net 3P% was bad too: -.244 and -.377. In their three series wins, their net 3PAs were 2 vs. Detroit, 28 vs. Boston, and 6 vs. Orlando (net 3P%s: .143, .186, .074).

While 3P% is so volatile game-to-game, teams obviously have better control over how many they shoot themselves. Attempting more threes than the opponent is generally a good offensive strategy in the playoffs, even if you concede a lot on the other end. In the 45 playoff series since 2018, the team with the higher net 3PA won 32 (71.1%) of them. This leads me to believe these numbers aren’t really from teams chucking up threes when they’re falling behind, as we’ve seen in many regular season Bucks games. Out-attempting works just as well, if not better than having the higher percentage: the team with the higher net 3P% won 28 (62.2%) of those series.

As Mitchell noted one year ago, liberal three-point shooting is not a cure-all for the Bucks. But it is a proven key to their success. As we continue to flesh out (h/t burt snipes) The Brew Hoop Grand Unified Theory of 3 Point Shooting, we rightfully focus on defense. The Bucks are clearly trying new looks on that end and their labor is starting to bear fruit, they must remember that as they’re burnishing the defense, it’s equally important to maintain a healthy dosage of “let it fly” when they get the ball back.

All statistics as of February 8th.