Arty Smith - The NBA's Pursuit of Efficiency - DATAcated Conference May 2021

Our next speaker is already Smith. He is a teacher and coach at 10 Denver school. And today, she's going to talk to us about the NBA's pursuit of efficiency artist. Mint is an award-winning teacher from Kent Denver School just caught every level of high school mad over. His 28-year career has developed an Innovative Project based analytics and visualization to high school seniors. Additionally. He has led the Sun Devils boys, soccer T25 Colorado, state champion, most recently in 2019. Alright, welcome to the

conference are excited to be here today. And he said, I work in the intersection of Education in sports. Analytics pacifically asked you to course the two seniors that focuses on data, visualization and storytelling. In the story begins in 1979, NBA added the three-point line and when they added the three-point line, I don't think anybody knew exactly how it would affect the game has taken us decades, to kind of figure out how to play the game of basketball. Most efficiently,

and the data set that, then I'm going to be using here, is over 5 million shots that were taken in the NBA over the last 22 years. Okay? For each shot. We know who took it, and we know whether it went in, we know the type of shot, it was, and we also know the location. Okay, and here's one shot that I'm going to share with you guys. This shot was taken by LeBron James on June 20th, 2013. If you're a basketball fan, you probably will remember this shot. Unbelievable about down game 7. 14224

Miami. And they're not going to for one. It doesn't look like. James. So that's the shot That Bass. The NBA championship for LeBron James and the Miami Heat in that here. And I'm going to show you guys is not just one shot. But rather I'm going to add all five million shots, the half-court here and what we get is, of course, a mess. All right, but each one of these shots has a story. There's a Steph Curry jumpshot from March 8th, but this is overwhelming. Okay, and then somehow, we need to make

sense of this data in the next route that I'm going to show you guys looks at the success rate of player shooting as a function of their did that. We would expect that when you're close to the basket, your probability of success. It is high in the further away from the basket. You get the lower your probability of success. Should be all my questions. You guys. Are you expected to be linear? We can easily get an answer to that question again. I've got distance on the x-axis here and shooting percentage on the y-axis. And what we can see is that when there's

are very close to the basket. So here at zero fee, they make 64% of their shots at 1:50 to drop the 63.7% 2 ft + 3 ft. What's really surprising here? I think, is that as we go from 3:50 to 10 feet and even all the way out to 22 ft. The probability of of making a basket doesn't change much. It's just a ticket below 40%. And that is really surprising. And I think is very counterintuitive. And I think it is taken teams long, long time to adjust to this back. Now, once we get

23-ft 24ft, we can see that shooting percentage, tends to fall off, kind of as we would suspect now. They're alerting variable in here course. I'm in a lurking variable, is that not all shots are worth the same amount of points? Okay, some of these shots are worth two points. Others are worth three points. So, the next graph that I'm going to show you guys will take that into account. Okay, and rather than looking at shooting percentage on the vertical axis. We're going to look at the average points per shot. So we can see

is 0 feet shots are worth about 1.3 points on average at two feet. It dropped about 1.09 and then from 3 ft. All the way on up, 21 feet shots, a word about .7, 9.8 points per shot. After that, though. We see a big increase. Now, these shots, of course, we're worth three points, right? And it leads to this surprising realization, that teams. Can't get a shot for very close to the basket. 0 feet, 1. Ft. Maybe 2 ft. Well, you're probably better off taking a 3-pointer and it's taking teams a long, long time

to a, to make this adjustment. This graph shows exactly that the time-series plot starting in 1997 through 2020. And it what we can see here. The orange crap is showing us the percentage of all shots in the league in a given year that were three-pointers in 1997. Only 15% of Scots were three-pointers. We can see that. That is risen all the way up to, about 39% done with that change really began was here in 2007 and 2070 Houston, Rockets hired, Daryl Morey to be their general manager, Daryl Morey, is a

self-proclaimed nerd out of MIT and a big proponent of analytics. In fact, Daryl Morey is the one who started the Sloan Sports analytics conference. That is held in Boston every year. And as of course the the percentage of shots that were threes has increased. Percentage of shots that were to have decreased. Now, we could be more specific. We could break it down into some subcategories and indeed I want to do that. Now, it turns out that some 3-pointers are actually more valuable than others. Take a look at the court here and we can see it at the three-point line.

It is part of us a circle here in our Circle, but it flattens out on that. You went on the two sides. Okay, more specifically the distance from right here from the left wing to the basket is 22 ft. Same on the right side, but the distance out to the ark is 23 ft 9 in to these three pointers from the corners are closer to the basket, then 3-pointers above the arc. And we can see the data tells us that indeed shooting 3-pointers from the corners or more valuable than shooting, 3-pointers, from above the break, and indeed. This is

informed a lot. NBA team strategies, I had a lot of teams. Now, what position players out in the left corner or the right corner and then try to get them open. Looks at the basket. Now, course that's easier said than done, and it helps to have somebody who can draw the defense into the middle and then kick the ball out. And, of course, nobody is better at that than LeBron James. Now, this is James, the game overtime and I don't want to show you visually a couple ways, how that's that's happened in. This is the shot map from 1980 1998, and 1999 is Messi.

There's too many shots here. We can't really make any Rhyme or Reason from them. So I'm not going to use a scatter plot of a density plot in this next crap is going to show us where the shots were most dense. Okay, where shots were taken from most frequent little cut through the Clutter if you will and MD. Densest areas are going to appear to be the brightest. So this is the 1998-1999 season. Okay, we can see more shots were happening on the basket. OK. Google Fairmount, mid-range

jumpers from the right wing and from the left wing, but other than that it is a fairly uniform distribution. People were shooting kind of all over, maybe not a ton from right inside the 3-point line, right? It's kind of dark area there and that makes sense. Right? Generally speaking. You'd be better off taking it one step back and in taking a shot at 3, rather than too. But I would say, overall, a fairly uniform distribution here. This is 19898, 99. This is before the NBA really started to play more efficient basketball. My next crap. I'm going to

compare this season 9899 to a more recent season to the 2019-2020 season. And I think it when you guys see this, you're going to see how the game of basketball has changed over the year, on the left. We see the old NBA on the right. We see the new NBI the efficient, and be a, okay? And on the way, we see that, if teams can't get a shot from underneath the basket while they're very likely to take a 3-point shot. Okay, and even more specifically, we can see that the mid-range jump shot his kind

of disappear. All right, at the game of basketball has changed because of data analysis. Now not only has has it changed the shot distribution for the league, but I also want to show you an example of how it's changed individual play. Okay, and in my next graph, we're going to look compared to scoring Champions were going to call. Look at Carmelo, Anthony versus James Harden on the left. We have Carmelo Anthony who won the scoring Championship in 2012, 2013. With 1764 points. Okay, we look at his shop map down here, bottom left. Each

green. Is a successful shot. Each red. Was a mess. And if we compare Carmelo to James Harden on the right, the shot map with very different. Okay, James Harden, and in the 2017-2018 season scores over 1,900 point and when we look at his shop map, it looks very different than Carmella's. Okay, James Harden with a much more efficient player. He took shots underneath the basket. He couldn't get those. He's going to take shots, from from three-point land. All right, and indeed, when we compare their efficiencies. Okay, points per shot Carmelo. Average

point nine eight points per shot. Where is James Harden? I was at 1.05 points per shot. A much more efficient player. I've heard it said that Carmelo Anthony was the best player in the league at those mid-range jumpers that he created. Mid-range jumpers, he's really good at making them. But unfortunately, over the next several years, the demand for mid-range jumper Shooters, really Stout. Okay, let's demand for people who had Carmelo skillset in essence. He was really good at something, that turned out wasn't that efficient? Okay. Now, a lot of basketball fan to argue that

this is the game of basketball. I got a quick quote, from coach, Popovich, from the San Antonio, Spurs. I'll give you a second to read it. So the game has changed for better for worse. Some people think we should move the three-point line, brain kind of more balance back to the game, and I'm no expert on basketball, but I think it's an interesting idea. I'm a few quick. Thank you. Thank you today to world for the day. Thank you to Kirk Goldsberry for the inspiration on this talk. And of course a special. Thank you to Kate from dedicated for giving me the so thank you guys, very much.

Absolutely already. That was, that was a great presentation. Thank you so much. I am simply want that video clip that you included her creative. I love that. And I I love beta visualization and and have low. So it's great to see all those days will come together. You have a lot of great feedback coming your way in terms of dust that used to happen. The way you visually represent it all of that. But we do have time for one question, will go ahead and jump into back from Elise. She's asking are outliers included in your assessments. For example, players, taking a shot with their sick, or

injured or rather. Is there a variable that accounts for the level of Defense fight while attempting to stop? That's a great question? No, I did not control for, for any of those things. It would be Anita. Richard Gator status at the time, to take a note to look at how closely the players were being defended. And indeed is keeping that date of the tracking data would give us a better. Look at how How close is the nearest Defender was to these shooters but that data is proprietary in the in the in the teams have that they're not sharing it with

with people like me unfortunately, but those are all valid things to look into it. And certainly, I could have filtered out and a last-second shots. Maybe they were taken. As the shot clock was expiring or at the quarter or half words firing. I can get rid of some of those which which would probably get rid of some of the noise and maybe give us a better look. So much for your time, and thank you for helping us close out. The sports industry. I will see you online. If you, if you want to have time. You can go hop in and answer some of those questions that have come in on LinkedIn.

Thank you.