Year Two: The Sixers Drastically Changing Shot Selection

Last year when Rich and I sat down with Sam Hinkie, one of the topics that came up was improving upon the Sixers shot selection.

During the final season of Doug Collins' reign, the Sixers had led the league in mid-range attempts, with nearly 1/3 of their shots coming from this inefficient part of the basketball court.  The Sixers had yet to play a regular season game under Sam Hinkie when Rich and I spoke with him, but the change in offensive philosophy, a change which was expected when the team hired an analytically minded GM, was evident at that point.

"Conscious," Hinkie said when asked to describe the change in shot selection. "Conscious. I don't have a good scale for degrees of consciousness but it's something our coaches have focused on. And in a tiny preseason, they've seen results."

A part of that is acquiring players who play the style of play they're looking for.  And, as the inverse of that, moving on from players who may not fit that philosophy.

"We acquired Shane Battier in Houston and didn't have to say ‘Stop doing this and start doing all the million small things that drive winning'," Hinkie said. "It was like, ‘You've always done that, just do you'."

The difference, whether because of Hinkie and Brett Brown drilling in an offensive philosophy to the players, because of the roster change, or both, was drastic.  The Sixers cut their mid-range attempts virtually in half, from it making up 32% of their attempts (most in the league) down to 16.7% of their overall field goal attempts.  That 16.7% would end up being the second fewest in the league.

(Related: It's probably easy to guess who had the fewest in the league, even without clicking on the link above.  That would be the Houston Rockets, Hinkie's former team, at an incredible 9.4%.  Just to display how ridiculous of a number that is, the Sixers were the only team in the league that had less than double the attempts of Houston: the Nuggets came in 3rd at 18.9%).

The folks over at Nylon Calculus take this one step further.  They scrape all NBA boxscores, categorizing shots into 5 locations: restricted area, in the paint (non-restricted), mid-range, corner 3, above the break 3.  They then find the league average efficiency from these attempts and try to predict how much a team should score based solely on the shots that they take.

The result is Expected Points Per Shot (XPPS).

Now, to be clear on this metric, this doesn't take into account the personnel on the team.  It simply tries to determine whether the shots that were taken were, in a general sense, quality shots.  It's taking "wow, that team took a lot of mid-range shots" one step further.  It's obviously not perfect, but it's an interesting number to try to quantify a shot chart.  It has data from all the teams in there dating back to the 2000-01 season.

The Sixers ended the season 3rd in Expected Points Per Shot last season at 1.071 XPPS, behind only Houston (no surprise) and Miami.  The site also tracks Actual Points Per Shot, which is essentially the teams ability to convert those shots into points.  No surprise, but the Sixers finished last in that regard at 1.020.

Which leads us to the next stat on Nylon Calculus: Shot Making Difference, which is the difference between Expected Points Per Shot and Actual Points Per Shot.  The Sixers Shot Making Difference was -0.051, by far the worst in the league, as only 4 teams in the league had a negative shot making ability.  In fact, if you take a look at all 14 seasons worth of data, 420 data points, only 5 teams had less of an ability to translate their shots into points: 2002-03 Denver Nuggets ( -0.074, which got them Carmelo Anthony), 2000-01 Golden State Warriors (-0.064), 2003-04 Memphis Grizzlies (-0.060), 2011-12 Cleveland Cavaliers (-0.055), and 2011-12 Charlotte Bobcats (-0.054).  Of those teams, only the Grizzlies took better shots than last years Sixers, with an Expected Points Per Shot of 1.080 and an Actual Points Per Shot at 1.020.

The overly simplistic conclusion: the Sixers have a sound offensive philosophy in place, but without the talent to actually convert that philosophy into points.

(Side note: The world champion Spurs were a relatively average team in terms of Expected Points Per Shot last season, as they were expected to generate 1.040 points per shot.  They actually generated 1.142 points per shot, good for 3rd best in the league.  They had a decent amount of mid-range shots in their shot chart, which accounted for 24.4% of their attempts.  It would be interesting to see how much being open / creating space impacts expected points per shot.  My hypothesis is that not only do the Spurs have personnel capable of hitting a mid-range shot at a higher expected efficiency than most, but that the shots they generate, in truth, have a higher expected efficiency than most mid-range shots because they're more open.  The Spurs Actual Points Per Shot exceeded their Expected Points Per Shot in 13 of the 14 seasons that Nylon Calculus has tracked.  I would guess that we tend to over-simplify "good" and "bad" shots based solely on their location, when in actuality there's far more to it than that.  That's not to say that the notion that mid-range shots are generally less efficient is unfounded, as I would guess that for 95% of the time, they are.  But what we see as outliers may not be outliers, they may be expected, but poorly quantified.  But, more or less, we're working with the data that we have available.  Perhaps SportVU will change that in coming years).

Looking at just the data the site has for the Sixers, 1.071 XPPS was the highest during the past 14 years, and by a considerable margin.  The second highest was 1.058 during the 2008-09 season, but even that only ranked 9th in the league that year.  More or less, the team has been in the bottom half of Expected Points Per Shot over the past decade and a half, which includes being dead last in the league during every one of Doug Collins' three years here (1.019, 0.998, 1.016).  In fact, that 2011-12 team, the one that saw the Sixers win that playoff series against the injury-depleted Bulls, then challenge the Celtics in the second round, was one of the 10 worst seasons in terms of Expected Points Per Shot in the last 14 years.  To the Sixers credit, they greatly outshot that expectation, generating 1.023 Actual Points Per Shot, to somehow form an offense that was only slightly below league average.

(Which brings me to the great question of the Doug Collins era: The Sixers always had a good ranking in Expected Points Per Shot allowed under Collins, maxing out at 3rd lowest in the league in 2011-12, which was a big reason in why that team was successful.  His defensive system never seemed to agree with his offensive system.  A part of that may have been personnel, and Hinkie's comments about Shane Battier, combined with moving on from Jrue Holiday being the first move that he made after taking over the team, seem to corroborate that he believes personnel certainly plays a part.  Still, last in the league for every season he was here should have been avoidable if his offensive philosophy matched his core defensive tenets).

Whether or not the Sixers XPPS remains towards the top of the league during the Brett Brown / Sam Hinkie era will be something to watch in coming years.  My guess would be that it will, and I would be surprised to see the Sixers not end up in the top 5 most years.

Now, if only the team could start acquiring talent that could actually convert those shots into points...