One feature I can’t read enough about from MIT’s Sloan Sports Analytic Conference was Sandy Weil’s presentation on motion capture technology. Zach Lowe did an excellent job covering it for The Point Forward. The basic premise is that 3-D motion capture cameras have been installed in five NBA arenas by the company STATS. The SportsVU cameras allow for the recording and analysis of data which previously had been too complicated to feasibly collect or tabulate by hand on a large scale. For example, a portion of Weil’s work looked at how field goal percentage is affected by the proximity of a defender. The proximity of the defender can actually be measured and recorded by the camera, something that would be wildly inconsistent and time-consuming if done by simple human observation.
Another of Weil’s findings was this (courtesy of Zach Lowe):
Catch-and-shoot attempts are much more efficient than other types of shots when you control for distance and the presence of a defender. A player’s shooting percentage jumps significantly when the last thing he does before the a shot is the act of catching a pass — and not the act of dribbling.
But if you catch a pass and hold the ball for about 2.25 seconds, whatever advantage you gained from catching the pass disappears. This makes sense, since holding the ball gives your defender a chance to catch up to you and prepare to defend your next move.
For mere mortals like myself, without access to this technology, one way to measure how often a player’s shots are catch-and-shoot opportunities is by looking at the percentage of their made shots which are assisted on. This does require somewhat of an assumption. To use assist percentage for this purpose we have to assume that the situation in which a player’s shots come is fairly consistent. That is that the shots a player makes from a certain location are fairly similar to the shots a player misses from that same location in terms of situation. This assumption is probably most true in the case of low-usage role players.
I was extremely interested when I noticed that the Portland Trailblazers, who have had the highest percentage of the three-pointers assisted on this season, are shooting roughly the same percentage on three-pointers as the New Jersey Nets, who have had the lowest percentage of their three pointers assisted on. The Trailblazers have had 93.0% of their three-pointers assisted on, while shooting 34.5%. The New Jersey Nets have had 79.4% of their three-pointers assisted on, while shooting 34.6%. Both teams are below the league average of 35.8% on three-pointers. The Nets’ performance dovetails perfectly with Weil’s findings but the Trailblazers seem to be bucking the trend, and not in a good way.
The player who seems most responsible for this trend is Nicolas Batum. The third year forward is shooting a, respectable, but career low 34.5% on three pointers this season, despite having 98.9% of his makes assisted on. In fact of players who have attempted at least 250 three-pointers this season with at least 94% of their makes assisted on, Batum ranks second to last in accuracy.
This performance seems out of character for Batum. He shot nearly 6 percentage points higher last season at 40.9%. It’s true he’s already attempted twice as many three-pointers as he did last season but this is almost entirely due to him missing 45 games last year. With the injuries to Brandon Roy you could make the argument that Batum has had to assume a more primary offensive role and might be taking shots he would have passed on in other seasons. However, the numbers don’t seem to bear that out either. Batum is actually attempting fewer three-pointers per 36 minutes this season; 4.8 compared to 5.2 last season. He’s also seen only a slight increase in his Usage Rate; 17.1% compared to 16.4% last season.
My next thought was maybe he’s making his catch-and-shoot opportunities but his misses are coming in different situations. That doesn’t seem to be the case either. According to Synergy Sports, 203, or 75.5% of Batum’s three-point attempts this season have come in spot-up situations. Unless I’m mistake spot-up shots are by their nature, catch-and-shoot. He’s shooting exactly 34.5% on those spot-up three-pointers, same as his overall percentage.
I’ve watched my share of Trailblazers games this season and I really can’t find an explanation. He’s being fairly selective and getting shots in the most efficient situations. He’s been a great shooter in the past but for some reason Batum isn’t knocking down his three pointers at a high rate this year.
I sent an email to Nathan Begley of Portland Roundball Society to get his thoughts. Here’s some of what he said:
I think that scouting has a lot to do with it. Batum blew up at the end of last season, so he’s no longer off anyone’s radar. When defenses are doubling [Brandon] Roy and hesitant to leave [Andre] Miller and [LaMarcus] Aldridge, an opportunistic scorer like 09-10 Batum can score very efficiently. Nowadays, Roy isn’t drawing the same defensive attention on the perimeter; Aldridge draws double teams, but his defensive “gravity” starts around 10 feet closer to the rim that Roy’s. In other words, Roy’s presence with the ball, even a few feet behind the arc, pulls the opposing defense over to get ready to stop him. Aldridge is almost never doubled at the three point line, he is even left in single coverage for long twos. Aldridge pulls a double-team down on the block, closer to the rim. This is a lower quality double-team from Nicolas Batum’s perspective because Aldridge has struggled with passing out of the double-team and just isn’t the creator that Roy is.
I also suspect that he needs to be in a good rhythm and that this season has been a very unstable one. Batum has had to shoulder a greater burden as well as play 3 positions (SF/PF and sometimes SG) and guard all 5.
Nathan’s explanations make a lot of sense. Batum is still getting good looks but they may not be the same high quality looks he was getting next to Brandon Roy. While his Usage Rate has only increased slightly Batum has undoubtedly assumed a bigger role for the Trail Blazers this season. That includes more defensive responsibilities and playing more minutes. Those factors aren’t adequately captured in the numbers.
*Nathan also pointed out after this post had gone up that Batum has been playing with an injured shoulder.
Which brings us back to the incredible potential of the data available from these motion capture cameras. They could theoretically offer us the information which may be the missing link in this discussion. When Batum is attempting a catch-and-shoot three pointer this season is his defender closer on average than he was last season? Is his defender closing out on him from a different direction, i.e. recovering from a low-post double team on LaMarcus Aldridge as opposed to recovering from the free throw line area after helping on a Brandon Roy drive? Because of the offense being centered more around Aldridge is Batum getting more three-point looks from the top of the key as oppposed to looks from the corner?
For someone who loves finding statistics to help to explain what my eyes see on the basketball court this tool could be mind-blowing. I’m sure some of you will groan reading this, but it could help identify statistics to fill in the gaps between other existing statistics. As exciting as the potential of this technology is, it’s equally depressing that the technology is proprietary. It may be years before widespread access to the information is free and open to the public. STATS does run a basketball blog which drops periodic statistical tidbits gleaned from their SportsVU cameras and the analysts who work with them, but it’s far from comprehensive. For now we’ll have to continue with the tools we have (which aren’t too shabby) and dream of what could be.