What does the Memphis Grizzlies' winning streak without guard Ja Morant say about Morant and the nature of on/off statistics?
This week's lead mailbag question asks what we should take from the Grizzlies' performance without Morant, who has missed the last three weeks with a knee sprain. During that span, Memphis has gone 9-1, including a victory over the Oklahoma City Thunder by the largest winning margin in NBA history.
Answering that question also helps us understand how volatile on/off statistics, or in this case splits of performance without a player in the lineup entirely, can be over small sample sizes -- a cautionary tale worth remembering for future situations.
Throughout the NBA season, I answer your questions about the latest, most interesting topics in basketball. You can tweet me directly at @kpelton, tweet your questions using the hashtag #peltonmailbag or email them to peltonmailbag@gmail.com.
This week's mailbag also addresses questions on the Dallas Mavericks' high-volume 3-point shooters regressing to the mean and on the accuracy of the multiplier we use to estimate possessions based on free throw attempts.
"I can't believe the Grizzlies are doing so well without Morant. Do you think they are better off without him and if so what should they look for if they trade him?" -- Kirby
This question has a short answer and a long answer. The short answer is no, Memphis is not better off without Morant and absolutely should not look to trade one of the NBA's brightest young stars before he even reaches restricted free agency.
The long answer is that Morant is yet the latest example in a long, long line showing that net rating differential -- something the statistical analysis community, myself included, admittedly popularized -- is dangerously flawed because of the role of shooting randomness over small samples.
When Morant went down with a knee sprain on Nov. 26, the Grizzlies were last in the NBA with a 115.9 defensive rating, according to NBA Advanced Stats -- the biggest reason they were languishing below-.500 at 9-10. As I wrote the following week, this was shocking given that Memphis had been a top-10 defense in 2020-21 and around league average two seasons ago. I also wrote that the Grizzlies were due to benefit from weaker opponent shooting the rest of the way, and boy have they since Morant went down.
In nine games without Morant, Memphis has been the NBA's stingiest defense, allowing just 96.6 points per 100 possessions. And much if not all of the difference can be traced to how well opponents are making jump shots. After shooting an effective 57% on attempts outside the paint (treating each 3-pointer as 1.5 field goals to account for their additional value) through Morant's injury, opponents are an ice-cold 44% on such shots since then -- going from the highest percentage any team has allowed throughout the season to the lowest.
Now, the quality of shots the Grizzlies have allowed has gotten somewhat more favorable since Morant went down. According to Second Spectrum tracking, Memphis' quantified shot quality (qSQ, based on the location and type of shot and distance to nearby defenders) has improved from 24th before Morant's injury to 14th after it. The Grizzlies are also leading the league in forcing turnovers on a per-play basis after ranking 14th through Nov. 26, which is probably more sustainable.
Remarkably, Memphis has also scored better without Morant, going from a respectable 12th in offensive rating when he was injured to fifth since then. This is surprising given that the Grizzlies were scoring 6.0 fewer points per 100 possessions with Morant on the bench before his injury. Again, I suspect randomness is the biggest factor. Memphis posted a league-high 155.1 offensive rating in its historic, 73-point rout of the Thunder. That and a 127.3 offensive rating Monday against a Philadelphia 76ers team playing without Joel Embiid have propped up the Grizzlies' offense, which has been slightly worse than league average otherwise after accounting for opponent ratings.
There are certainly elements of how Memphis has played without Morant that the team can hope to carry forward when he returns, in particular how effective Jaren Jackson Jr. has been. After struggling through the season's first 19 games, when he was averaging just 14.8 points per game and shooting a tad under 40%, Jackson has boosted that to 19.8 PPG (25.5 per 36 minutes) while hitting 49% from the field and 36% on 3s. And if more help from Jackson on offense helps Morant expend more energy at the defensive end of the court, that's a good thing for the Grizzlies.
Still, there's no good reason to believe based on a hot stretch built on fluky opponent shooting and a favorable schedule that Memphis is better without Morant or should be pondering trading him.
"At what point should we stop expecting regression to the mean from the Mavericks' top shooters -- Tim Hardaway Jr., Reggie Bullock, Kristaps Porzingis, Dorian Finney-Smith, etc.?" -- Oded
Including guard Luka Doncic, all five Dallas players who have attempted at least 100 3s this season are off their pace from last season -- many of them dramatically so. Collectively, this group shot 38% from beyond the arc last season, weighted by this year's attempts. That's dropped to 32% thus far, costing the Mavericks nearly five points per game. You can see why Mavs owner Mark Cuban is tweeting about the impact of the NBA changing suppliers for its game basketball.
Nearly 30 games seems like plenty of time for 3-point percentages to even out, but it's still a relatively small sample in terms of attempts. As I did a few weeks ago for team performance, we can compare how well 3-point percentages to date correlate with remaining 3-point percentages (for players who saw action in at least 75 games and attempted at least 200 3s in 82-game seasons since 2012-13).
As compared with the previous season, current 3-point percentage doesn't become as good an indicator until about midseason. And even then, previous 3-point percentage has a similar relationship until the very end of the schedule.
Now, regression doesn't mean Dallas' shooters are likely to shoot as well as they did last season. For one thing, they're not the only players experiencing a drop-off; the high 3-point percentages we saw in 2020-21 increasingly look like an aberration. Additionally, this year's cooler shooting still conveys information. We should adjust our expectations for the Mavericks' players downward to some extent.
Still, the odds are good that Dallas will shoot better as a team going forward than it has thus far.
"What is the biggest difference recorded in a season for a player between the 0.44 fraction used to correct for non-two-shot trips to the line in the TR% formula and the actual fraction if we take into account the and-1s and three-shot trips and technical FTs?" -- Chakad
This question refers to how we have historically estimated possessions from free throw attempts for players and teams. Because not all free throws are shot two at a time, the average across all free throw attempts is: each of them translates to 0.44 possessions. But that ratio can vary from player to player depending on the types of free throws they shoot.
Thanks to the now-defunct PBPstats.com, I was able to pull totals for those non-possession free throw attempts dating back to the 2000-01 season. (Note that flagrant foul free throws would also be non-possession, but PBPstats.com did not track them and there are probably not enough to make a substantial difference.)
First, the good news. The multiplier generally works pretty well:
For 57% of players in this sample, their actual possessions on free throws were within one of the estimates. We can see a few outliers, particularly at high free throw totals. Here are the players who most exceeded their estimated free throw possessions:
Big men tend to have higher free throw multipliers because they rarely draw three-shot fouls or shoot technicals. DeAndre Jordan and Dwight Howard also stand out as players on the receiving end of intentional fouls, which by definition are of the two-shot variety. Three of the next four seasons on the list are from Howard as well.
At the other extreme, we find players who draw a lot of three-shot fouls and regularly shoot Ts, with a recent MVP dominating the list:
James Harden's proclivity for being fouled beyond the arc breaks the multiplier. He drew at least 70 three-shot fouls in each of these four seasons; nobody else on record has drawn more than 55 (Lou Williams in 2016-17). Add in shooting technicals and Harden's actual multiplier was below 0.4 these four seasons.
Stockton got near the top of the list by shooting 84 technical free throws in 2000-01. That was the most in this database and good for 30% of his total attempts -- the highest rate for any player to take more than 150 free throws in a season.