What Have You Done for Me Lately?
The NBA is a right now league. Each game can change in a moment. Every season’s trade deadline brings about substantial team restructuring, and the offseason free agency and trade market has become an event unto itself. Change happens quickly, and the team that wins the championship is often the team that becomes the best version of itself at just the right moment.
Many of the most important questions deal with which player or team is the best right now. “What have you done for me lately?” is the unspoken question on the minds of everyone in and around the league. In previous posts, I’ve used the defensive matchup data at stats.nba.com to create a model for defensive performance since 2013-14. The entire dataset is now features prominently on the homepage, and you can pull any player card for any season from the Google Sheets tool.
The Race for Defensive Player of the Year
Ben Simmons jumped out to a huge lead in the Matchup-Based Defense leaderboard, and got a lot of buzz as a candidate for Defensive Player of the Year. In fact, he was able to save Philadelphia nearly 460 points before suffering an injury. Simmons’ versatility and indispensability to the jumbo-sized Sixers help him stand out as an elite high-load stopper.
With Simmons out for the foreseeable future, however, he seems unlikely to finish the season as the league’s top defender. Who might able to supplant him as the statistically-informed candidate for Defensive Player of the Year?
In the time frame since 2013-14, only Draymond Green has been the rightful Defensive Player of the Year in two seasons. Jrue Holiday, who deserved the award in 2017-18, has also risen to the top of the ranks this year and looks poised to join him. Holiday has saved the Pels 548.4 points this year at a rate of 20.8 Points Saved per 100 Possessions. He currently has a significant lead in Total Points Saved, and ranks fourth among high-load defenders in Points Saved per 100 Possessions. He is also one of only six high-load rotation players in the league who are saving their teams at least 20 points per 100 possessions.
The best defender in the league on a per-possession basis is Gary Harris, the leader of Denver’s cadre of two-way wings. Harris is the only player in the NBA besides Holiday to have saved his team 500 points already. Harris’ teammate Torrey Craig is also saving over 20 points/100, but limited playing time has kept him to a total of 222 Points Saved.
Among recent All-Defensive selections, Patrick Beverley (358 Points Saved, 20.7 PS/100), Marcus Smart (419 PS, 18.8 PS/100), and Eric Bledsoe (376 PS, 18.7 PS/100) are performing at similar levels this season. Avery Bradley has saved the Lakers 19.7 points per 100 possessions this year in a bounce back campaign. Dejounte Murray (308 PS, 16.6 PS/100) also seems to be recovering nicely from his injury.
Who’s Up Next?
One of the best ways to use advanced statistics is to detect up and coming prospects before the competition. Both the eye test and statistical analysis require sufficient observations in order to reach conclusions or make predictions. That means that scouts need to see a lot of film in order to recognize a good player. Coaches need a lot of time in practice to form reliable valuations of a player. Can statistical models get ahead of the curve by identifying undervalued players before they can be recognized by the eye test?
Matchup-Based Defense identifies Luguentz Dort as a high-impact defensive prospect. In his NBA minutes, Dort has saved the Thunder 177 points at an outstanding rate of 22.6 PS/100. Of course, although it’s no hot take, Matisse Thybulle (239 PS, 16.8 PS/100) also looks like a standout defender in the minutes he’s played. Recent callup Juan Toscano-Anderson (66 PS, 19.9 PS/100) has made a good start on the defensive end, though we probably don’t have enough data yet to know if that’s for real or not.
Second year players Terrance Ferguson (292 PS, 18.8 PS/100) and Josh Okogie (378 PS, 18.4 PS/100) are making roles for themselves in the league based in large part on their defensive performance. Shaquille Harrison continues to work his way into a fringe spot in the league with 98 Points Saved at a rate of 18.8 PS/100. Hamidou Diallo (167 PS, 18.4 PS/100) and Dwayne Bacon (169 PS, 18.3 PS/100) have both been decent role players in their team’s defensive schemes. Any of these players could be the next defensive menace.
Who’s the Worst?
Now that we’ve highlighted the best defenders in the league and some of the best prospects, you might well be curious who the worst defenders in the league are. Among players who have played at least 500 possessions this year, the weakest defenders on a per-possession basis are:
- Denzel Valentine (the worst … just ask Jim Boylen)
- Timothe Luwawu-Cabarrot
- Michael Porter Jr.
- Tony Bradley
- Jeff Green
- Dzanan Musa
- Terrence Ross
- James Ennis III
- Rodions Kurucs
- Kelly Olynyk
These bottom ten players all save their teams less than 9.0 points per 100 possessions. Other notables are Jahlil Okafor (11th-worst), Darius Bazley (13th-worst), Davis Bertans (15th worst), and Kevin Knox (17th worst).
How to Look Up a Player
The table below is also on the site’s home page, and includes every player in the league for this season.
You can also get an idea of the shape of the data using the color-coded chart below. Colors reflect defensive load, while the size of each dot represents how much a player has played. The graph plots total performance against per-possession performance, with the second and third quartiles colored gray. Players in the gray areas are closest to average (both overall and per-possession).
Players in the top left white space and low-minute players with good per-possession data, while the bottom is the home of players with low totals and low per-possession values – the worst of the worst. The bottom right white space is usually going to feature scorers – guys who play a lot of minutes but are not very effective defenders on a per-possession basis. Finally, the top right corner is the home of the stoppers – high volume and high efficiency defenders. Mouse over any dot to find out who it is.
What Really Happened at the Trade Deadline?
A number of players with varying defensive representations were traded before this season’s trade deadline. Did it make any difference? Which teams increased their available defensive talent? Which teams got weaker on the defensive end? I’d like to briefly review a few of the more notable trades, starting with the trade that was ostensibly made for defensive purposes: the Clippers traded Maurice Harkless to the Knicks in order to acquire Marcus Morris (the self-proclaimed “LeBron stopper”).
Though the trade was intended to bolster the Clippers’ defensive performance, it is on the whole unlikely to have this effect. Maurice Harkless has been a noticeably better defender than Marcus Morris this season, proving more effective in preventing opponents from scoring despite carrying a far heavier defensive load. While it’s certainly possible that Morris might make the Clippers’ defense slightly better against the Lakers, it is not accurate to say that the Clippers beefed up their defense in the trade.
Which means that the Knicks … won a trade?
Another significant trade in terms of defensive impact was the four-way trade that sent Robert Covington to Houston and Clint Capela to Atlanta. The move was hailed in some quarters as a revolutionary commitment to small ball and defensive versatility by the Rockets. Robert Covington has indeed carried a heavier load than Capela this season, and done so more effectively, but the data suggests the gap between the two is rather slim (and Capela’s production rate per 100 possessions is slightly higher than RoCo’s).
One of the most talked-about moves at the deadline was the Warriors/Wolves deal centered on Andrew Wiggins and D’Angelo Russell. Looking at the trade from Golden State’s point of view, Wiggins represents a defensive upgrade over D’Angelo Russell. This result is not too hard to figure; Dlo has an obvious advantage on the offensive end, is younger, and is on a far more agreeable contract.
Andre Iguodala and Justise Winslow, both accomplished perimeter defenders, traded places in another trade. Neither has played enough yet this season to produce a reliable picture of their value. Below you can compare their performance this season as well as last season.
As you can see, Iggy was a substantially better defender than Justise Winslow last season. If he still has enough left in the tank, Miami could be a stingier defensive squad now than they were before the deadline.
How to Use Statistics
As a way of concluding this update, I’d like to provide some suggestions for using Matchup-Based Defense in real time. The first and most important guideline is to pay careful attention to the load each player carries. The table on the homepage can be filtered by Defensive Load and Relative Load (adjusted for team context). Comparing players who carry vastly different loads is irresponsible and will produce misleading results. The types of conclusions one might reach, based on the statistics provided here, are that a given player has performed at this level while carrying his current Defensive Load. Placing him on a different team or in a different scheme might impact the results observed, so it pays to bear in mind how much a player is being asked to do.
It is also fruitful to drill down in a player’s profile to determine how he contributes or fails to contribute, rather than simply using Points Saved as an isolated, all-in-one stat to help us prove that Player is X is a “good defender” and Player Y is a “bad defender.” Diving deeper can reveal insights as to how a player is most effective – perhaps in preventing his opponent form scoring, perhaps by forcing turnovers.
Even within the realm of forcing turnovers, it is important to pay careful attention to the types of turnovers a player causes. One player may be a pickpocket who piles up load of Solo Steals, while another may be a hard-nosed hustler who draws charges, and yet another player might be a pest that harasses their man into turning the ball over. Inspecting the columns on the right side of the table provide more insight on these distinctions for this season’s data. For prior seasons, please refer to my Tableau Story which contains comprehensive visualizations of league-wide defensive performance from 2013-14 through last season.