Brandon Knight and the (mythical?) point guard vs. shooting guard debate (Part I)

It was one of the recurring debates of the Bucks' 13/14 season, and one that promises to bug us through the summer and start of next season: Is Brandon Knight a point guard or a shooting guard?

As simple and straight forward as it might seem, I'd argue it's also a rather loaded question, brimming with assumptions about what point guards and shooting guards should do, burdened by the false need to categorize players with a single position, and colored by competing views of how good he is in the first place.

So spoiler alert: I don't think the answer to this question is as simple as one or the other, and it's probably still too early to really be sure no matter what we think we've seen to date. But that's not to say Knight's fit isn't an important question and one the Bucks should be pondering carefully. Though he's still just 22, the Bucks' leading scorer is now only a year away from restricted free agency, and he figures to find himself playing with a number of different types of guards--both old and new--this coming season. So how good is Brandon Knight, and how might he best be used going forward? Let's take it piece by piece.

Forgetting position for a moment, how good has Knight been in the first place?

Let's start with an observation: the better you are at basketball, the less people worry about labeling you one thing or another. And while Knight is far from position-defying greatness, he's at least trending in the right direction.

Having only turned 22 last December, Knight made major strides across virtually every offensive statistical category a year ago, increasing his scoring and assist rates, lowering his turnover rate, getting to the line 40% more, and upping virtually every efficiency metric in the process (PER, TS%, WS, ORtg) while cutting an ugly -1.2 offensive xRAPM in Detroit to a more respectable +0.70 in Milwaukee. According to MySynergySports.com, His pick-and-roll, isolation and transition finishing all jumped notably, indicative of his improved ability to make plays with the ball in his hands. The only notable offensive decline came in his three point accuracy, where he dropped from 38.0%/36.7% in his first two seasons to 32.5% last year on comparable volume.

Defensively it wasn't the same story, as the Bucks were worse defensively with him on the court, his opponent stats went from good to mediocre (12/13 vs. 13/14), his Synergy stats were largely below average, and his defensive xRAPM tanked (from -0.4 to -1.9) under the weight of the Bucks' flimsy defense. Then again, he was also playing on a mostly rudderless Bucks team with little semblance of a consistent defensive scheme, so I'd caution against drawing strong conclusions about any of the young Bucks defensively (for now).

Still, to Knight's credit both he and the Bucks also became more effective offensively as the season wore on. After suffering a hamstring injury in the season opener and coming off the bench as he worked his way back into shape, Knight averaged 19.7 ppg on 53.4% true shooting to go with 5.0 apg and 2.6 turnovers in 35.4 mpg over the season's last 50 games. The Bucks were also mostly awful in that span, but their offense quietly turned around after the all-star break as well, ranking 14th in the league efficiency-wise over the final 30 games. So while we should be wary of mistaking causation and correlation, it is nevertheless a positive that Knight carried the load while both his team and personal efficiency increased.

I'm not sure we can say the same about his defense, but overall he was the team's "best" player essentially by default. That also poses the question of whether Knight's numbers are inflated by opportunity (i.e. the "gunner on a bad team" thesis) or hindered by having so few weapons (i.e. the "point guard with no one to pass to"), an issue that was similarly raised about his predecessor Brandon Jennings. Likely it's a bit of both: Knight would see his shots and scoring numbers decline on a better team, but we might also infer that his scoring efficiency and assist/turnover rate could benefit from a more complementary role. It's a test case we'll hopefully see borne out this season.

Still, Knight's value remains more about what he could become rather than what he is now. Compared to the rest of the league, Knight ranked 18th among qualifying point guards in PER (22nd among all guards), 36th in win shares (58th among all guards, mostly due to his poor defensive numbers) and his "real plus minus" ranked comfortably below average among point guards. And while he also became much more dangerous as a pick-and-roll finisher (0.81 points per play vs. 0.69 in Detroit), like Jennings he was mostly looking for his own shot in P&R situations. He's yet to flash the sort of change of pace and playmaking instincts that you look for in a top point guard, an area of frequent criticism in any debate over his long-term future. All told, the numbers put him squarely in the slightly below average grouping of starting lead guards, though he was still a generally good player and at the age of 22 his window for improvement remains open.

After all, remember that Knight was almost certainly the league's least productive starting point guard during his two seasons in Detroit, so there's been progress. And despite already having three full seasons under his belt, Knight still projects as the league's third-youngest starting point guard, behind only Kyrie Irving and Trey Burke. It's also worth mentioning that a number of solidly above-average point guards--Goran Dragic, Mike Conley, and Jeff Teague--experienced similarly prolonged development cycles while battling questions over their point guard "purity." Note that the only one of those guys who didn't lead his team to the playoffs this year was Dragic, and all he did was win 48 games, claim the league's Most Improved Player title, and earn third team All-NBA recognition.

But the paths of players like Teague, Conley and Dragic also underscore the somewhat absurd amount of talent currently at the point guard position; for all of Knight's progress, it's difficult to suggest he's any better than average compared to the rest of the league's starting PGs, and he'd need to become significantly better in order to claim a spot in the top ten. In other words: don't hold your breath for any meaningful talk of an early contract extension this summer, but that doesn't mean the Bucks should be giving up on him either.

How would we characterize Knight's positional tendencies from last season?

I normally subscribe to the "you are what you defend" theory whenever a player's position is debated, but with point guards the conversation is necessarily more complicated. Knight's size and tendency to look for his own shot has long led to discussions over his positional purity, both for good and bad, offense and defense. But there's also little doubt that Knight was the Bucks' nominal point guard from an offensive standpoint for the vast majority of the time he was on the court last season.

That's an important starting point, because I've seen some suggest that the Bucks' inclination to play Knight with other ballhandlers--namely Ramon Sessions, Nate Wolters and Luke Ridnour--could be equated with Knight already having played as a shooting guard in Milwaukee, and thus (by that logic) his improvement could similarly be traced to this change in role. But the data really doesn't support that line of thinking.

Aside from insisting that I saw Knight typically serving as the team's initiator and primary ballhandler--you know, the guy bringing the ball up, initiating sets and generally doing most of the quarterbacking--we can also point to SportVU data on number of passes per game, total touches per game and time of possession data, all of which are consistent with a primary ballhandler rather than a player utilized off the ball (just look at the names on the lists). It's not to say Knight was a great point guard during the 13/14 season, or that he's the Bucks' long term solution at the position. But we should agree that he did actually play the position last season, which suggests his improvement did not come from a sudden freedom to play off the ball. If anything, Knight was more point guard-ish last season than in Detroit, where his Synergy data showed he relied less on P&R (27% of plays vs. 37% in Milwaukee) and more on off-ball plays like spot-ups (20% vs. 14% in Milwaukee) and coming off screens (9% vs. 4% in Milwaukee).

So if a guy plays point guard and shows big improvement overall, does that mean he's now a point guard, through and through, and that his best complement in the backcourt is something other than another point guard? Not necessarily. Using the wonderful NBAWowy.com, we can look at both Knight's individual numbers and the Bucks' team performance when he was paired with different backcourt mates:

The data is more interesting than it is conclusive--fewer than 500 minutes per lineup on a team that spent most of the year thrashing around in irrelevance isn't exactly a huge sample--though it does offer some insight into a few important topics. Some comments:

a) When looking at this data keep in mind which part of the season might have correlated with which combinations. After an injury-plagued start of the season, Knight's season began to turn around December 18 with a then career-best 36 point night vs. the Knicks, which also happened to be Antetokounmpo's first start and the effective end of the Mayo era. Ridnour then cycled into the starting five on New Year's Eve, the night that Knight set another career high with 37 points against Marshall and the Lakers. That lasted for another month or so before Wolters took over his spot, and then Sessions arrived at the trade deadline in late February.

b) Generally speaking, decent point guards (Wolters, Sessions, and Ridnour) proved better foils than underperforming shooting guards (Mayo and Neal). The data suggests at minimum that Knight can play well with secondary ballhandlers, though we also never saw him play next to a "traditional" shooting guard who was any good. This might lead me to prefer putting Knight next to another ballhandler, as well as other players who can help facilitate an offense. But I'd stop short of saying that we know Knight doesn't play well with shooting guards. Giannis and Jabari Parker's potential emergence as shot creators could also lessen the reliance on Knight or any other point guard as playmakers, though we may be getting ahead of ourselves for now.

c) Knight probably found the best balance of individual and team success with Wolters, as the duo was barely negative in differential and Knight posted good scoring, passing and efficiency marks. His numbers then dipped a bit with Sessions, which seemed to make sense given what I remembered seeing at the end of the year. Sessions' own ball-dominating, slashing style just seemed a worse complement to Knight than the more deferential, defensively sound Wolters--and the numbers bear that out. The fact that Wolters played steady defense and actually started to hit three pointers starting in February (15/33, 45.5% over the last two months) likely didn't hurt. Interestingly Knight scored the most and passed the least with Ridnour, though it was still Knight who handled the majority of ballhandling and initiating responsibilities.

d) Knight's individual numbers with Giannis Antetokounmpo (21.9 ppg, 7.2 apg and just 2.5 to) are both the most impressive and the most "point guard-ish," though the team's defensive numbers are an obvious eyesore. Note that to capture the lineups where Giannis was nominally the Bucks' other guard, I eliminated all lineups involving any of the other players in the table above. Thus, you're left with "big lineups" featuring Giannis in addition to two other forwards and a big man. Not that Giannis was doing anything particularly special from a box score standpoint--his 51.5% true shooting was below average, he had a sub-1.0 assist/turnover ratio and scored under 10 pts/36 when paired with Knight in the backcourt. None of the other players in the lineup look phenomenal either, but for whatever reason (small sample size warning!) Knight looked really good and the Bucks scored a lot of points when he played with Giannis as his nominal backcourt mate.

So overall we're left with a somewhat muddled picture. Knight's own numbers improved greatly while having the freedom to lead an offense, but that offense also happened to look terrible for large stretches of the season. Then when he started playing more with another ball-handler the offense began to show signs of life, though his best individual numbers came while being partnered with a 6'11" teenager who was barely entrusted to do much of anything offensively. So maybe we just call him a combo guard and keep experimenting?

On the one hand Knight's positional ambiguity can be a selling point: his size and athleticism don't put him at a major disadvantage defending taller guards, and his ability to handle the ball, attack the basket and create his own scoring opportunities makes him more useful than both "game manager" points or run-of-the-mill shooting guards. You know, the kind who are often defined more by what they can't do (dribble a basketball, do stuff other than shoot) than what they can. In many ways that also makes him the polar opposite of the recently acquired Kendall Marshall, whose brief career to date suggests that being a "pure" point guard isn't a one-way ticket to job security either.

Still, the liminal space between the two guard positions isn't all upside and rainbows (well, unless you're Russell Westbrook?). Knight's mediocre assist-to-turnover ratios and below-average true shooting percentages belie more serious questions about both his decision-making and ability to win games, as he's not good enough as a scorer (yet?) to gloss over the other stuff. Knight optimists are thus still banking on potential improvement as much as immediate production, and on a bad Bucks team it was only natural to question the guy with the ball in his hands. So we're left with a few conclusions and one major open question: Knight did in fact serve as the Bucks' "point guard" for the majority of his minutes last season, and he did show major signs of improvement in areas that are decidedly on the ball skills. But does that mean he's the Bucks' point guard answer going forward?

Stay tuned for Part II, which will focus on how the Bucks can best use Knight along with the rest of the current roster.