Is there a FG% difference between teams that take more/less of these on average?

An obvious question arises as a final piece when attempting to research these questions: if you’re a high-percentage mid-range team, won’t you simply take more of these shots? To determine the answer, I looked one final time towards the North Star of Hoop-Math.com. What I got back from attempting to see if there’s any correlation between more of a certain shot and a higher FG% in that category was…interesting.

Average Correlation in Field Goal Attempts to Field Goal Percentage in Rim, Mid-Range, and Three-Pointers, 2012-2021

• Rim: -0.24 (Weak Negative Correlation)
• Non-Rim Twos: +0.12 (Very Weak Positive Correlation)
• Threes: +0.24 (Weak Positive Correlation)

In general, teams that took less shots at the rim were a bit more efficient; that’s not a huge surprise, because those shots were likely wide-open layups or easier rim attempts than the ones hammered in by teams crazy about getting to the paint. Also predictable: teams that took more threes made a few more in general. Coaches generally have a good idea on if their team is any good or not at shooting from downtown, so those that don’t have many shooters probably aren’t going to take many deep shots.

What was also predictable: there’s almost no correlation whatsoever between taking more mid-range attempts and making more of them. The best non-rim two-point field goal percentage by any team since 2011-12 is 2018-19 Liberty, who hit 47.9% of their mid-range shot attempts. If what held true on the other two categories held true here, you would expect them to take more of those shots. The reality: only 14 teams in college basketball took fewer non-rim two-point attempts than Liberty did that season.

So, no, even if you’re generally pretty good at hitting these shots, there isn’t much statistical backing for taking more of them. In fact, it could just be a push to either get closer for higher-percentage looks or to take a step back for a shot that counts for an extra point and goes in at nearly the same percentage.

Is anyone consistently good at it?

The second half of our previous question, though, is a fun one. From 2016-17 to now, 85 teams averaged 30% or more of their shots from the mid-range, with 26 cracking the 33.3% barrier. (30-33% seems to be the “we like these shots but don’t love them” area.) In the same five-year sample, just 28 teams hit 40% or more of their mid-range attempts on average. The list of teams that had 40%/30% crossover:

1. BYU (43.6% FG%, 35.8% of all shots)
2. UCLA (40.7% FG%, 31% of all shots)
3. Denver (40.5% FG%, 30.6% of all shots)
4. Missouri State (40.5% FG%, 38.8% of all shots)
5. Southern Miss (40.5% FG%, 39.9% of all shots)

That’s five teams out of what will be 358 in the 2021-22 season. It’s not very long, and there’s only one Big Six representative on this list. The amount of teams that shot 40% or better from mid-range every season is also minimal; only three programs were able to sustain the bare level of shooting necessary to make taking 33.3% or more of your shots from the Murky World of Other Twos a good idea. (None of these three programs took more than 26.1% of their shots from long two land on average.) The shooting average of the 26 schools who got at least one-third of their shot attempts from Other Twos was 37.1%; the shooting average of everyone else was barely any lesser at 36.8%. I think you get the point.

What other questions can be answered?

Here’s a few scattered ones I came up with during the research process that deserved their own section.

Do teams with one or more volume mid-range shooters have a higher team-wide non-rim FG%? For this, we’re looking with players who took at least 140 mid-range attempts in any of the last five seasons. (This seems like a strange number, but it was where the true mid-range heads and volume shooters began to separate themselves from the pack.) The answer here, perhaps unsurprisingly, was a minor yes. However, it wasn’t as resounding as some might guess. Here’s those stats, year by year:

• 2020-21 (36 players): 33% of all shots (+6.9% above national average), 39% FG% (+1.8% FG%)
• 2019-20 (124 players): 30.5% of all shots (+4.3%), 37.6% FG% (+1.4% FG%)
• 2018-19 (129 players): 29.2% of all shots (+3.8%), 37.3% FG% (+0.7% FG%)
• 2017-18 (149 players): 31.2% of all shots (+4.0%), 37.8% FG% (+0.6% FG%)
• 2016-17 (181 players): 32.3% of all shots (+4.0%), 37.3% FG% (+0.6% FG%)

In some fashion, the difference between the haves and have-nots is growing; these teams are taking more long twos on the whole and are showing a very modest advantage against the rest of the field. However, the difference is showing more in overall volume of attempts than it is in how good a team is or isn’t at making them. Outside of 2020-21, a very small sample, there’s a pretty consistent +4% trend in teams with a volume mid-range shooter (or shooters) taking more of these shots than the average team. However, it’s very important to note that these teams are not necessarily making more shots. 

Let’s narrow it down to the better cases now that we’ve set our base.

Do teams with one or more highly efficient (42% or better, minimum of 125 attempts) volume mid-range shooters have a higher team-wide non-rim FG%?

I backed the attempts down a hair simply so we wouldn’t have a very tiny sample size. In general, 42% on an individual basis is around 5-6% higher than the national average. That would be like shooting 39% on threes, which pretty much everyone would agree is very good.  It won’t shock anyone to hear this, but teams who have highly efficient mid-range scorers tend to have a better mid-range offense on the whole.

• 2020-21 (30 players): 32.6% of all shots (+6.5% above national average), 40.1% FG% (+2.9% FG%)
• 2019-20 (64 players): 31.9% of all shots (+5.7%), 39.3% FG% (+3.1% FG%) 
• 2018-19 (70 players): 28.4% of all shots (+3.0%), 38.2% FG% (+1.6% FG%)
• 2017-18 (81 players): 30.1% of all shots (+2.9%), 39.7% FG% (+2.5% FG%)
• 2016-17 (89 players): 30.4% of all shots (+2.1%), 39.6% FG% (+2.9% FG%)

Really, outside of 2018-19, you could say that these teams are consistently 2.5%-3% better than the national average at hitting these shots, and the ratio of more shots/more makes is at least better here than it is in the example above. However, one thing is obviously worth noting: in our last ‘normal’ season of 2018-19, 2,100 players averaged at least 16 minutes of game time across 25 or more games. Only 70 of them turned into what I’d call Green Light Guys. The average college basketball player has slightly over a 3% chance of developing into this; you need to make sure your guy (or guys) are taking the shots they need to take.

Do teams that take a lot of mid-range shots go farther in March?

The qualifier here: at least 33.3% of your shots need to be long twos. I’m measuring the success rate here based on PASE (Performance Above Seed Expectation), which tells you how many games a certain seed would be expected to win based on past performance by that seed. The results are…telling. There was no tournament in 2020, as a reminder, so I dated it back to 2016.

In terms of PASE, teams that qualified for our setup managed to win precisely 0.03 games more than the average team over the last five NCAA Tournaments. A total of 37 teams qualified, and while there were some obvious standouts (2016 and 2017 North Carolina, 2021 USC, 2021 UCLA), the vast majority either fell right in line with their seed expectation or underperformed horribly (hello, 2019 Virginia). Essentially, there is no correlation to more mid-range and more wins.

But what about more mid-range and a higher FG%?

Do teams that make a lot of mid-range shots and take a decent amount of volume (30% or more of overall shots) go farther in March?

Our qualifications here are simple ones: 40% or better on mid-range FG%, 30% or more of your shots from the mid-range. We ended up with a 26-team sample here across five Tournaments, and…yeah, there’s also pretty much no serious correlation here. The average PASE across these 26 teams was just 0.09 wins above expectation. 13 teams either performed to their seed line expectation or below it; 13 performed above. It might sound great as a narrative, but there’s no serious evidence that having the combination of volume and good mid-range efficiency actually helps you win more games in March.

If your assist rate is higher on mid-range attempts, will your FG% be higher? 

Here’s an easy one: no. The correlation of mid-range FG% to mid-range assist rate is +0.02, which is essentially no correlation whatsoever. As much as it pleases me to see a catch-and-shoot mid-range attempt versus a pull-up off-the-dribble, there is nothing statistical that backs my feelings up. Even stats freaks like myself have the capacity to feel disappointment sometimes.

Lastly: did teams who took a lot of closer mid-range twos show any sort of correlation with better mid-range FG%?

The beauty and the problem of Other Twos is this: depending on the source, they can also mean floaters, runners, post-up attempts, and even really long layups that are categorized as something else. To try as best as possible to separate actual mid-range jumpers from everything else, I went to CBB Analytics’ fantastic Shooting by Zone breakdown. Here, you’re able to break down shots by At the Rim (<4 feet from the basket), Short 2s (4-10 feet), Medium 2s (10-15 feet), and Long 2s (15+ feet). Because of the nature of our search here, we’re strictly looking at those latter two categories. You simply don’t see many runners from deeper than 10 feet, so we can assume that our data pull here is almost entirely jump shots.

What stands out here is both the averages of these two shots:

• Medium Twos, 2020-21: 36.1% average
• Long Twos, 2020-21: 36.7% average

And their correlation to overall mid-range FG%:

• Share of Medium Twos Correlation with Mid-Range FG%: +0.00

Literally no correlation whatsoever, I hate to say. But it did require exploring at the very least, so now we know: anything from 10-20 feet is valued at pretty much the exact same for the average team.

NEXT PAGE: Attempting to find key threads