After looking at point guard purity, potential assists, and shot creation, one thing is probably clear: there's no simple way to determine what makes a great point guard statistically.
Purity doesn't necessarily equate to quality; shot creation ratios might demonstrate the relative importance of point guards, but might undervalue players who are asked to score on spot up shots by their teams; potential assists would be helpful and as soon as I set aside 272+ hours to watch every game this past season, I'll get back to you on the potential assist count of every point guard in the league.
But within the comments on all of those posts was a somewhat consistent observation: great point guards find some balance between setting up scoring opportunities for themselves and others that is hard to capture with any one number.
Who counts as a point guard?
Ultimately, what makes a great point guard is decision making - part of that is controlling the tempo, being able to initiate the offense, and figuring out how to keep the defense off balance to create shots for teammates. But to (over)simplify all of that, we can reduce what point guards do to the ability to effectively decide when to pass and when to shoot in the best interest of their team.
But in the WNBA this past year, even that oversimplified description creates a blurry line about who should "count" as a point guard. Players as different as Katie Douglas and Ticha Penicheiro spent some time this past season as a "lead ball-handler", literally the player who dribbled the ball over half court and made the first pass in the offense. But many people sort of sub-consciously think about "point guards" in terms of their "purity": players that exist closer to the right hand side of the point guard purity spectrum. And that makes a lot of sense - people commonly think about point guards as ball handlers and passers who distribute the ball rather than those who look to score.
Becky Hammon: Distributor, scorer, or scoring distributor?
For example, there are some people who still insist that San Antonio Silver Stars guard Becky Hammon isn't a point guard because she has a scorer's mentality. However, if you can score as well as Hammon can, there are times - more often than not - when choosing to score is the best thing you can do for your team and, by extension, being a threat to score can help set up opportunities to set up others to score. But Hammon is also a great example of how fluid the distinction is between distributor and scorer not only game-to-game, but year-to-year (which I've said more about here).
Keeping the matter of who counts as a point guard as an open question, Hammon had the highest usage rate of any starting "lead ball handler" in the league and is in the top 15 league-wide (24.78%); in other words, Hammon was looked to as a scorer for about a quarter of the possessions she was on the floor, which was above average for any player. But what makes Hammon unique as a basketball player is her ability to balance those scoring tendencies with an assist ratio (25.26%) that is well above league and point guard average. In other words, Hammon is able to both score more often than most players and create assists more often than most players.
Part of what allows Hammon to rank so highly among assist leaders despite shooting so often is that she has the ball in her hands more often than any player (53.26 touches per game). But what separates her from a player like New York Liberty guard Cappie Pondexter - about 50 touches per game and also responsible for bringing the ball upcourt quite a bit - is not only her assist ratio, but her overall style of play by the SPI styles framework, which sort of maps onto purity ratings.
You can click here and here to find plenty more about that player style framework, but the comparison of Hammon and Pondexter is a good place to start to illustrate this difference: they're about equally efficient, but have different overall tendencies.
Measuring point guard quality
As described at length in last season's point guard rankings, I think you can begin to approximate point guard quality by using two metrics:
Pure point rating (PPR): This is John Hollinger's creation that measures a player's quality as a distributor in terms of a player's ability to balance the risk of creating an assist with committing a turnover per minute. As mentioned in the potential assists piece, potential assists figure pretty prominently here. The question this metric answers is simple: How well does a player balance the benefit of creating a scoring possession for others with the cost of throwing away a possession? (Click here for more on this metric and potential assists)
Points per empty possession (pts/empty): This is a metric created by Bob Chaiken that I like as a complement to PPR as an approximation of scorer decision making. Similar to an aggressive passer, an aggressive scorer is bound to both commit turnovers and miss shots. The value of this metric - as compared to other efficiency metrics such as true shooting percentage - is that it tells you whether the points a player creates for the team was worth the empty possessions. This has to be taken by position - point guards actually have the lowest average pts/empty in the league - but overall a player who has a high ratio relative to their peers can be said to be a better scoring decision maker in terms of identifying opportunities that result in points.
You'll notice that this framework puts a heavy weight on turnovers, which might seem to amount to a form of "double jeopardy" - point guards commit more turnovers than most players simply because they have the ball in their hands more often. But the point is that above all else, a team can rely on a great point guard to commit turnovers less often per touch or possession than most players - in statistical terms, you want a point guard running your offense because you know they'll actually be able to initiate the offense without a turnover about 85-90% of the time.
So here's how Hammon and Pondexter compare on just those numbers, just to further illustrate a point about these numbers and what they do/n't tell us:
Going strictly by the two-pronged framework described above, Pondexter would definitely be among the top 10 "point guards" in the league. But in addition to the fact that she doesn't start there - meaning she does get a fair amount of assists playing from the wing - she has much higher scoring tendencies than Hammon, which makes her far more scorer than distributor despite her efficiency as a passer. We could debate it further, but that's a good starting point for this overall point: a rough cutoff for point guards is pretty much being in the top 20% of the league in perimeter tendencies (P%). Assist ratios help as well as 21% tends to be a pretty good approximation of who was in the top 20% - Pondexter is right around 19%.
So basically I limited this season's point guard rankings (forthcoming) to players who could be classified as "distributors" as follows (ordered by assist ratio):
So a few notes on players that sat on the border:
Erin Phillips: Phillips is a tough one statistically as her assist ratio is below average among that list of lead ball handlers and she had a SPI perimeter orientation of 80.2%, which would normally be a cutoff point. The thing is, Phillips played 24 games as the starting point guard for the Fever and actually played it more efficiently than most assume.
Tanisha Wright: I actually did put Wright as 10th on the list of point guards last season and I stand by that - she and Sue Bird alternate ball handling responsibilities quite a bit and she was rather efficient at doing it last season. What changed this season is that she wasn't so efficient this season and had an assist ratio below average for a lead ball handler.
Kara Lawson: On the other side of that Phillips border line was Lawson, who wasn't at all a high-usage player and had a reasonable assist ratio (22.10%). But arbitrary cut off lines tend to be arbitrary and Lawson had a perimeter orientation rating of 78.8%.
Noelle Quinn: Quinn is simply an example of a player whose numbers don't necessarily match observation - she's about as "pure" a distributor (37.42) as Sue Bird or Renee Montgomery and has a higher pure point rating (2.84) than either. So strictly by the numbers, she's squarely in the point guard discussion. However, if you watch the Sparks play, Penicheiro and Kristi Toliver assumed point guard duties most often and Quinn was often on the floor with them. Quinn was a very efficient distributor, but it's hard to say she served the function of "Sparks point guard" with any consistency.
Penny Taylor/Armintie Price: I won't even explain why they're not point guards, but it is interesting that they were among the most efficient distributors in the league. It was actually one of Price's biggest and probably most understated improvements this season, as she had an assist ratio (23.65%) right around average for a point guard. That clearly makes them very useful players to have on the wing - and I'd argue important parts of their teams' success - but they also demonstrate why even though I use numbers to help understand point guard play, they're not the sole determinant of a quality point guard.