Boundary and Nexus: Maximizing Inefficiency

It’s a well known fact: the long 2 is the least efficient shot in basketball. 

The reasoning is quite simple, really. The further you are from the basket, the harder it is to make shots. And since by going further than the long 2 range you gain an extra point for every shot, long 2s are the furthest, low scoring attempts in basketball. 

This (correct) notion has become ingrained in our analytical DNA. Sure, the long jumper is impressive when it goes in, but no basketball fan in their right mind would rather see Lebron James settle for a foot-on-the-arc fadeaway when you know that he can take it to the rim at will. 

However, there is more to basketball than just “long 2s bad, close 2s and open 3s better” (yup, that was an Animal Farm reference. Because at HP we are dedicated to bringing you both snot-nosed, self righteous, social-critiquing literature, and zombie movies). Long 2s are shot, and often, whether it’s because the players that shoot them know they can make them, want to keep defenses honest, or are just plain lazy. But how damaging are they anyway? 

Lets start by taking it back a few steps, and numerically conforming the statistical inferiority of the long 2. With the help of the life-altering Hoopdata.com, here is a table of the league average eFG% for different shooting ranges as far as Hoopdata’s records go (the 2006-2007 season), and the averages over the entire stretch (since the current season is just midway, I weighted the results under the assumption that the average NBA team has played 34.7 games this season, as was true of January 7th). 

The results: 

 

The number that jumps off the page is that the percentages for the 10-15 foot “in-between” range were not only in the same area of the 16-23 foot range, but actually worse for 4 of the last 5 seasons and over the entire 5 year stretch. However, that figure becomes much less surprising when you look at the figure right next to it – the % of made shots that were assisted. This makes sense – although we very often see shots from that range taken in ISO situations, you see plenty of spot ups from just inside the arc, and very few spot ups from 10 to 15. And in general, it’s much easier to make shots when your body is all lined up and a passer hits you than when you have to create on your own. 

In addition, shots from in between are usually tougher makes. Many of them are either of the floater variety (hello Derrick Rose) or come off the mid-range post game (hello Kobe Bryant). But Roses and Kobes are few, and these shots tend to be harder to make, especially since they are taken closer to the shot-altering reaches of various big men. So while I didn’t expect the two ranges to be similar percentage-wise, it does seem to fit in with general convention – in a vacuum, you’d rather be open from 13 feet than from 18 feet, but how you get there and how you’re being guarded is a factor. 

Moving on from the percentages, we reach the amount of field goals attempted from each range. Interestingly enough, there has been a baffling drop in attempts at the rim in 2010-2011 – 4.5 less a game. However, the corresponding, nearly matching jump in shots from 10 feet in (up 4 shots a game) makes me think that shot locations may be counted differently this season in contrast to years past, thus creating the difference. I may be wrong with this assumption, but since it’s not our topic of focus, we’ll keep it at 2-bit conjuncture levels. 

In what is our topic of focus, though, the “shots attempted” numbers are quite odd as well. The average NBA team since 2006 attempts 20.9 long 2s a game, which accounts for 26% of all shots. This exceeds the amount of shots from every other range except at the rim. Remember, this is the widely acknowledged least efficient shot in basketball. Are teams just stupid? Or is there more to this? 

Allow me to toss out two theories as to why this is the case. First of all, the evolution of the sweet shooting big men – initially just known as Euros, then acquiring the ever descriptive name “stretch 4s”, has surely helped in this regard. While some of these gigantic sweet-strokers make their hay behind the three point line, a substantial amount of them score their points off the 20 foot range. Guys like Dirk Nowitzki, Kevin Garnett, Chris Bosh, The Artist Formerly Known As Rashard Lewis and David West have made multiple all-star teams behind the premise that their high point of release combined with their shooting prowess makes them efficient from the non-efficient range. Nowitzki, specifically, is probably the best long 2 shooter in recent history, making a ridiculous 53% from 16 to 23 feet this season (by far his best since in Hoopdata’s records and unlikely to sustain but still darn impressive) while taking 43% of his shots from that range. 

The list of big men shooters goes on and on. Amar’e Stoudemire has been more than just a dunker for years now. Al Horford seemingly can’t miss a long jumper. Pau Gasol, Lamarcus Aldridge, Luis Scola, Kevin Love. All good shooters. In today’s NBA, even if you’re big, you need to make that shot. With so many more players capable of making the shot, it makes sense that more players take them, even if it’s not as efficient as getting to the line or spotting up a bit further back for an extra point. 

Another reason could be the annoying-but-ever-present pick up mentality that a swished jumper is more impressive than banging into your defender for two freebies. What’s more memorable, those Kobe jumpers that don’t even touch the net as they account for 2 more on the scoreboard, or when he fools his defender into fouling him on a 3 point shot that was never intended to go near the rim? Players want to show they can make those shots, so they attempt a disproportionate amount of them, even if their strengths lie elsewhere. Just look at Allen Iverson’s career. 

Of course, this still doesn’t excuse teams from focusing way too much on the least effective way to score points. While nobody would make the argument that good teams shouldn’t shoot long 2s at all – if you’re got a good shooter and you’re open, then by all means, knock it down – the stark contrast to conventional wisdom presented by these numbers makes one wonder. Do better teams try to lessen their long 2 attempts? Conversely, do they try to bump it up a few notches for their opponents? Does this even affect them? 

In order to answer these questions, I once again turned to last season’s shot location data per team (last season and not this season for the larger sample size). I plotted the percentage of shots each team took from from 16 to 23 feet (looking at the percentage of shots and not the total amount of shots from each range in order to adjust for pace) versus each team’s offensive efficiency, in order to measure just how damaging the “worst shot in basketball” is anyway. 

 

This graph clearly shows us that there is some kind of relationship between the relative amount of long 2s a team takes; however, this relationship is far from conclusive. As we can see, the scatter on this graph is fairly large, and the correlation between the two variables is only 0.0768 (running on a scale of 0 to 1, where 1 is a perfect relationship, and 0 being a Tony Allen-O.J. Mayo relationship), which means that many other variables have far stronger effects on offensive efficiency. However, some kind of effect is there, even if it’s minimal. 

Look at the two extremes on the chart, and you see exactly the teams you’d expect to find: Orlando took the least long 2s in the league by far, right on par with Stan Van Gundy’s “Dwight inside or everybody else from 3” strategy, taking it to second in the league in offensive efficiency; meanwhile, if you watched the Chicago Bulls last year, you know that Vinny Del Negro “coached” his team into an offense full of long jumpers and void of any offensive movement, resulting in the highest percentage of long 2s in the league and the 28^thranked offense. Right on their heels were the Washington Wizards, who spent the season featuring notorious gunners; early on, Gilbert Arenas and Caron Butler were the ones taking your contested long shots, only for the likes of Andray Blatche, Nick Young and Al Thornton to take over the inefficiency load once the two former all-stars left the team via suspension or trade. 

As for some outliers, the iso-heavy yet offensively efficient Hawks and Blazers expectedly broke the trend, ranking 9^th and 4^th in percentage of long 2s, and 3^rd and 7^th in offensive efficiency, respectively. The Bosh-led Raptors also efficiently built around long jumpers, ranking 6^thon both accounts. Underperformers were just as familiar, with a cluster of offensively terrible squads such as the Bobcats, Bucks, Pacers and Clippers appearing far below the trendline, with 09-10 laughingstocks New Jersey and Minnesota falling even further. All these 6 clubs were around the league average as far as taking long 2s, but were so atrocious elsewhere that it dragged them towards the lower regions of our graph. 

So good teams tend to take less long 2s, but it’s not a necessity for a good offense. How about good defenses? Do they corner opponents into long jumpers? We go for the same exercise, only with shots given up and defensive efficiency. 

The defensive graph seems to very easily break into three groups. The elite defensive teams of 2010-2011 – those in the 100 to 102 points per 100 possessions range – seem to have been all over the place as far as giving up long 2s. On one end we have the Magic, who seem to employ the same defensive strategy as they have on offense – have Dwight take care of the inside, everybody else close out strong on 3 point shooters, and thus force opponents to take plenty of 2 point shots from too far away. This plot ranked them 3^rd as far as long 2s conceded. On the other side we have the Thunder, whose opponents were 3^rd from the shallow end, presumably due to a lack of a true physical presence inside to deter opponents from stepping out for their 2 pointers. 

Looking past that elite group, we have an Atlanta-Portland-Dallas-Utah-Chicago clump, whose defensive efficiency ranges from 102.6 to 104 (ranked 10^th to 14^th), and whose opponents take long 2s 24.4% to 25.3% of the time. It’s hard to note any effect long 2s have on this particular group, since they are all very similar. 

The third group is the most interesting one as far as our research, and it shows the what pretty evenly breaks into the bottom half of the league defensively. In this region, there seems to be a pretty direct effect. In fact, the correlation for just this part of the chart was 0.381, by far stronger than that of the offensive graph. When we add the top 14 defensive squads, the correlation drops to 0.1019. In fact, if you take another glance at the trendline, it really does look as if it was tailor made for the top half of the graph before the evenly distributed bottom came and dragged it down in it’s original shape. 

The difference between top and bottom is very odd, and I find no explanation for it, seeing how there is no reason why something should only affect bad defenses and immediately stop mattering once you creep up to number 14. As such, we can probably dismiss it as a fluke, leaving a message very similar to that of the offensive graph: channeling your opponent to the inefficient long 2 gives you a better chance of being a good defensive team, but you can just as easily be a great one without it. There are many ways to skin a cat, this just happens to be an effective one. 

To get a good look at the complete effect of long 2s, I combined the two graphs by plotting differentials – offensive minus defensive. Since this is basically combining the two previous graphs, we shouldn’t be surprised that the results are pretty similar. The Magic again were stalwarts, with the absolute best differential in both categories, proving both that they strategically avoid the inefficient shooting range while funneling their opponents towards it, and that they are extremely successful in doing so. 

 

The conclusions here may seem trivial. However, they have analytical value. We know that long 2s are bad shots, but this gives us two seemingly conflicted truths: that it is generally smart to reduce those shots for yourself while encouraging your opponent to take them; and that is more than possible to fail while doing so, or succeed without it. While the long 2 is often considered to be inefficiency incarnate, it is not a death knell. It can be manipulated and molded into a prosperous ordeal.