In Part One, we discussed biases and the value of efficiency, as well as a brief argument for the merits of tracking services like Synergy.
In Part Two, a broader look was taken at reasonable expectations of efficiency by shot location and the shifting demands of the college basketball market, and dealt with addressing common, situational arguments made by advocates of the mid-range.
So, where do I, from an analytical perspective, stand on the mid-range shot as a whole? Well, against it, in most current situations.
But, what situations am I for taking a mid-range, personally? For that, I have developed a criteria for when a mid-range attempt is likely to be an efficient option. The shot attempt should come:
In the final ~5 seconds of the shot clock
From 16 feet and in
Without a realistic way to forecast manufacturing a decent quality three-pointer or shot at the rim within two passes
On balance, both feet set underneath the player (this means no stop-and-pop pull-ups)
With at least three feet of space from the closest defender
It’s a lofty criteria, but it’s designed to only be reached at most 5 or so times a game, and that’s for an entire team. The fact that the situation described in this criteria would eliminate the vast majority of mid-range attempts currently taken is not a bug, but a feature.
Foremost, following this would lead to no more wasted possessions early in the shot clock (if the league even has one), when the team could’ve worked for a better shot, but instead took the first one that was even close to open. Everything else is designed to weed out attempts that would have no chance at being hit at even a 40% rate. When looking at a shot with an expected PPP of 0.80, that’s a PPP that 89.7% of DIII players were able to reach on guarded catch-and-shoot attempts during the 2018-19 season (min. 60 attempts). Are there outlier situations, like facing a shot-blocking 7-footer in high school play when none of the players on the other team are taller than 6’4”, when this criteria might need to be temporarily abandoned? Yes, of course, but those are exactly that: Outlier situations.
Considering that NBA studies have shown that the expected return on a possession only starts to drastically diminish in the final five or so seconds of the shot clock, heavily contested mid-range shots occurring with more time than that remaining on the shot clock are inherently very nonoptimal. As much as I would love to advocate for giving each player more freedom to decide when they should or shouldn’t take shots from the mid-range, the data shows that the vast majority of college players do not have a reliable idea of their actual efficiency from the mid-range. Why else would Division III players combine to take more than 70,000 cumulative mid-range attempts during the 2018-19 season, if they only shot a combined 37% on those shots? To put that 70,000 attempts number into further perspective, there were less than half the amount of possessions tracked by Synergy as occurring in the final 4 seconds of the shot clock during that same DIII season (34,041).
Ideally, the best shooters would get a little bit more leeway than the more average shooters with regards to how strictly they need to follow the above criteria. But, the fact is that almost every player needs to focus on cutting attempts first, and following it up with a more nuanced understanding of shot efficiencies as it relates to the game situation later. The players who would have a strong statistical case for not needing to heed any element of the criteria are rare, but they do exist. Kevin Durant immediately springs to mind as one of these players, but I’ve noted previously in assessing Durant’s play-type profile that there simply are not a lot of wrong answers for him, offensively speaking. Not to mention, to get Kevin Durant’s game, you just be willing to follow Durant’s regimen, one that famously involved literally sleeping inside the gym during childhood. It also helps to grow to be reportedly seven feet tall while retaining the ball skills of a guard.
And if the idea that a player might be the Kevin Durant of their competition level exists, track the data. The nature of percentiles is that somebody has to be the 1st percentile, where they are only better than 1 out of every 100 in a certain skill, for there to be a 99th percentile.
By definition, not everybody can score in the 90th percentile and above at a given time.
Still, the mid-range remains valuable across as levels of basketball as a counter, when used as a counter. The data shows however that players (and possibly coaches) are vastly overestimating how often the counter should be employed, as exhibited by the 37% shooting percentage. If the goal is to be efficient in all aspects of basketball, including the mid-range, how are so many players ending up taking a shot that is so massively inefficient when compiled as a group?
The question isn’t, and never should have been, whether or not it’s possible to shoot 55% in a single-season from the mid-range on certain types of attempts (it is, regardless of statistical noise’s potentially helpful influence to generate outlier performances among a large population), The question is whether or not each individual attempt is an efficient use of a possession, with the entire situation considered, and the answer is that the vast, vast majority of mid-range attempts are not.
I have one last graph to share, it’s also from the 2018-19 season, when I consulted heavily with St. Thomas in St. Paul, Minnesota:
This is the team-wide, mid-range efficiency of every team in Division III to attempt at least 152 total shots from the mid-range in the 2018-19 season. The red bar at the very right of the graph, the most efficient team in all of DIII from the mid-range by a wide margin, is St. Thomas, in Minnesota. The team had the talent to make the shots due to a lineup stocked with shooting, the coaching from coach John Tauer to put the best players in the best position to make shots, and only took just a shade over 5 mid-range shots a game. And even with all this working in St. Thomas’ favor, the team mid-range PPP of 1.03 was actually less than the PPP of their catch-and-shoot attempts tracked as guarded, with a PPP of 1.06. Because efficiency is the most important thing in basketball, St. Thomas naturally gravitated towards more shots on guarded catch-and-shoot jumpers than mid-ranges, attempting the former at a rate almost two-and-a-half times more often than the latter on the way to a 24-5 season. It’s also worth noting that this season, with different roster construction, the team has taken 31% fewer attempts from mid-range as the team has been much more efficient from three, and therefore prioritizing those attempts to a greater extent. Some of the situations that led to promising late-in-clock mid-range attempts last year have now turned into a more deliberate effort to manufacture a three-pointer, or an attempt to draw a foul at the rim, leaving behind desperate mid-range attempts at a greater proportionality.
To tie this all together with one final story, recently I had a lengthy conversation about the mid-range jumper, shot selection in general, and the process and practice habits behind it all, with Dion Lee, longtime skills trainer for D’Angelo Russell and Romeo Langford. The conversation centered mostly on Russell, who is currently shooting 58 of 110 (53%, 1.05 PPP) from 15-feet and further on two-point attempts so far in the 2019-20 season, a massively impressive number when considering the league-wide PPP on mid-range shots was 0.83 in the 2018-19 NBA season. Watching Russell’s attempts so far this season, it became clear that Dion and I probably viewed the mid-range, and shot selection in general, in similar ways. It’s also worth noting that Dion isn’t a stranger to the value of the three-pointer, he once tied the Big Ten record for most made threes in a game with 10 made in a game against Minnesota.
Talking on the phone, that was immediately confirmed, as Dion mentioned that everything that he and D’Angelo have worked on has been aimed at making Russell as efficient as possible within his role, skill set, and physical constraints. In many ways, D’Angelo is a great example of a player who punches well, well above his weight when it comes to his physical gifts, and represents an attainable goal for any young player who is willing to work on their games. Russell works hard to get to the rim as often as he can, even for a player who isn’t particularly explosive relative to his peers, is a fantastic catch-and-shoot three-point shooter (38% on catch-and-shoot threes so far this year), and takes over 5 times as many two-point shots from the 15 to 19 feet (94) as he does from 20 to 24 feet (16). 53% from the mid-range is incredible, but it falls short of Dion’s lofty personal goal when he was a player, which was to be as efficient from 15 to 19 feet as he was on contested, at-rim finishing attempts.
The argument from Dion was, and has consistently always been throughout the course of multiple conversations on the topic, simple: With good habits and years of repetition, Dion and D’Angelo had anticipated that teams would over-commit to taking away higher-efficiency shots from threes and shots at the rim (areas where Russell has twice as many combined shot attempts from, compared to mid-range attempts), and having a counter from the mid-range was essential as a primary ball-handler and scorer. I don’t know exactly how Dion and Russell put together workouts to turn D’Angelo into one of the most efficient mid-range shooters in the NBA, but the outline was clear: Years or repetitions down to the smallest of details to build muscle memory, and constant self-critiquing to weed out the least inefficient habits and attempts. Dion was adamant, that with the right training, any shooter can shoot 55 to 60% from the mid-range… on certain attempts, with the right training.
Frankly, it’s hard to disagree with Dion, he has the proof, and the proof is paramount in an ecosystem where hundreds of trainers and coaches claim to have the secret sauce, few have the high-efficiency proof of concept to back it all up, as Dion and Reid have shown. Not to mention, my own life experiences have shown me that it’s correct, both as a player, and now as an analytics consultant whose job entails an endless stratifying of possessions for my clients.
Many say that analytics versus the eye test is a false dichotomy, and the conversations I have on a daily basis, like the ones with Dion, have led me to agree. If analytics is like a compass, then the eye test is like a map.
Relying solely on the map is asking to be blown dangerously and needlessly off the path, and no map is so detailed on a granular level to replace the aid of proper and objective alignment when all immediate terrain is roughly indiscernibly alike. Those who refuse to crack open the compass are betting too heavily on their natural ability of alignment when objectivity is readily available. Simply, there is no real, only imagined, risk in consulting objective data to confirm their suspicions drawn from many personal experiences, even if they have honed their abilities to reach a naturally high success rate.
Meanwhile, those who refuse to use a map, solely because they have a compass, make life harder on themselves with their incessant rigidness. Without the aid of a map, analytics may have the objectively correct alignments, but success never occurs in a linear fashion. Without consulting a map, many will find that their self-contrived parameters of success do not align with those of actual, real-world successes, and in times of legitimate need, will find they’ve blown their social currency among the group long ago on pedantic and petty infighting. At times an overreliance on objectivity, especially one that clearly does not align with their immediate surroundings, will be painfully apparent to all in the group, save for the compass-holder. Worst of all, the compass-holders will often struggle to meaningfully merge their most substantial information with those among them who do not have the same instrumentation or acumen as they do. While many compass-holders figured an informational war of attrition would naturally win others over to their process, instead of actual social outreach of ideas, this has not been the case. Their objective knowledge is inherently meaningless to those who do not share their (often perceived as blind, by others) faith in the compass, regardless of how objectively correct it proves to be.
One is an instrument that relies on the laws of nature to work, the other is the result of many generations of trial-and-error (or, in some cases, no trial-and-error, just personal, dubious intuition), that has been honed into a mostly accurate representation of reality over time. Neither single-instrument system precludes reaching the end goal, but both will undoubtedly make life harder on the one-dimensional traveler, and the journey longer, regardless of the truism that the shortest distance between two points is a straight line, which both methods seek to follow.
So, the question remains the same, but the answer is less binary than either side would hope.
Because an exclusive stance on either side of the artificial dichotomy is wrong: Arguing that the mid-range shot can’t, even in favorable situations, be efficient is as ridiculous as arguing that mid-range shots can, in every (or even most) possible situation, be efficient. The secret is to know what the proportionality between these two types of mid-range shots currently is. How many of the ones taken fall into the first category, and how many fall into the second, and how can data help us understand that relationship better? Remember, if 55%, or 50%, is the goal, the current marks in the high 30’s are falling considerably short, and on a massive scale.
Because to make the mid-range an efficient shot, a player has to be incredibly committed, and constantly probing for better looks. To sum it up in three words, “Planned, purposeful, deference.” By working for the best shots, planning for those to be taken away, and yet still seeking them out aggressively on a per-possession basis, that’s how you get the best, most-efficient mid-range attempts.
And if a player thinks they can cut corners in this process, to rely heavily on the mid-range without either a consistent three-pointer or the ability to finish at the rim, they’re bound to find themselves somewhere in the 30th to 70th percentile range on their mid-range attempts. And landing in the 30th to 70th percentile range of efficiency on those shots will invariably land them on the bench if they’re determined to continue taking these inefficient looks at a consistent rate.
To tie it all together, let’s go back to the question I started with, “How much should mid-range shots be emphasized, in both practice and games?”
Well, how resigned are you to taking them, and how nuanced will your approach to the shot quality be?
Comments