The NBA play-in tournament begins one month from Sunday, which means there isn't much time left for teams on the fringe of the playoff races to make a push.
Throughout the season, I will be answering 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 edition of the NBA mailbag includes:
• The importance of seeding in the play-in tournament
• The significance of role to player value from an awards perspective
• Whether picks can be too protected
Let the mailbag begin:
"How important is it to be seventh or eighth in the NBA's new play-in tournament rather than ninth or 10th?"
-- Austin, Los Angeles
I think the NBA media has been somewhat guilty of discussing the play-in tournament for the final two spots in the playoffs in each conference as a monolith: either you're in the play-in or you're not. In practice, the way the tournament is set up dramatically favors the two teams who would traditionally go directly to the playoffs as the seventh and eighth seeds.
Let's imagine a world where all four teams in the play-in tournament have the same underlying ability and all that matters is where they're seeded. The tricky challenge, during this unusual season, is to figure out how much value to assign home-court advantage.
Overall, games played with fans have seen the home team get an edge of about 1.4 points, equivalent to winning 55% of the time on average, but that's down dramatically from a two-point edge when I last studied this data two weeks ago. It's possible that's just random noise, but I do wonder if the novelty of limited crowds has worn off as nearly all teams have welcomed them back.
For now, let's assume the home team will win 55% of the time in the play-in tournament. That produces the following expected outcomes based on seeding:
There is a massive, massive difference between entering the play-in tournament in seventh or eighth -- and needing just one win out of two games to advance -- and coming in ninth or 10th and needing to win twice, at least once on the road. And in practice the chances for the seventh and eighth seeds to make the playoffs will probably be better because odds are these teams are in fact stronger than those in ninth and 10th.
Certainly, the chances are still not as good as the 100% odds teams in seventh and eighth had of making the playoffs before, the source of the frustration Dallas Mavericks star Luka Doncic recently expressed about the play-in tournament. It's especially bad by comparison for the seventh seed, which could get bumped down to eighth. But just getting into the play-in tournament doesn't actually convey a great shot at making the playoffs if you're ninth or 10th.
"When evaluating the value of players, how do you account for players who put up bigger numbers for a while when key players are out? For example, people keep talking about All-NBA and MVP cases for Kevin Durant, Kyrie Irving and James Harden. Aren't each of their numbers skewed by the other players missing significant numbers of games, which results in the other players on the team having to pick up the slack? I really don't think you can take all of their stats and say they would each average those same points, rebounds, assists, etc. if they all played 90% of their games together."
-- Shaun, West Seattle
My first answer to this is that's part of why statistical analysts don't look at points, rebounds and assists per game very often as measures of player value. Instead, we've developed all-in-one metrics that should be less sensitive to the role a player is in.
To test my own wins above replacement player (WARP) metric, one study I've done is what happens to players' stats when they change teams based on the quality of their new team relative to their old one. Naturally, you find a strong relationship between the change in team ability and how players' minutes and points per game change. (Assists per game are stronger yet.) But there's basically no relationship at all between the difference in team quality and the difference in player win percentage, the per-minute component of WARP:
All of which is to say the concept of "good stats, bad team" -- understandably driven home to NBA observers by decades of seeing players' box-score stats change with their team quality -- shouldn't exist with well-calibrated player value metrics.
To me, the award that's really influenced here is Rookie of the Year, where Anthony Edwards has been putting up massive box-score stats without as much efficiency as players like Durant, Harden and Irving maintain when their usage rates go way up. Since the All-Star break, Edwards is averaging 23.8 PPG, which is more than any rookie since David Robinson (24.3 PPG in 1989-90) has managed over a full season. Edwards has certainly been much better over that stretch, lifting his season WARP total above replacement level, but not nearly enough so to catch Tyrese Haliburton (4.8 WARP) or LaMelo Ball (4.5 before his wrist injury).
For the true offensive superstars, there definitely is a benefit to playing with weaker teammates because the ability to carry heavier loads on offense without suffering as much in terms of efficiency is what makes those players so rare and special. Harden in particular would not likely rate nearly as valuable if Durant had been healthy all season.
Still, as my ESPN colleague Zach Lowe has highlighted, that ability to help the team survive injuries to other stars is a huge part of the value the Brooklyn Nets sought by trading for Harden. So I don't think it's anything that should be held against him come awards time.
"I was looking at the RealGM future draft picks page and, noting all the long-term protection on many draft picks, thinking about how this reduces the ability to trade future picks. Is putting too much protection on a traded draft pick sometimes counterproductive? Could it sometimes be better to put less (or no?) protection on a pick to have some certainty when it would convey and to allow you to have the ability to trade other picks before the protected/traded pick finally conveys?"
-- Matt
The first thing to note here is that there are ways around those protections. On a technical level, the Detroit Pistons cannot trade a first-round pick because of the protected pick they sent to the Houston Rockets on draft night last year, which could convey any time between now and 2027.
In practice, however, teams will often trade a first-round pick in the soonest year possible -- two years after the previously traded pick conveys. That's what the Denver Nuggets, for example, did with the pick sent to the Orlando Magic in the Aaron Gordon trade -- the clock starts on that pick (which has its own protection) two years after the Nuggets send a first-rounder to the Oklahoma City Thunder from their draft-night trade for RJ Hampton, which could happen from 2023 through 2025.
Still, I think we're going to see fewer picks that are protected as long as the Detroit pick to Houston going forward because of the challenges that creates. The Pistons can't guarantee the team they're trading with a first-round pick in trade conversations until after the 2023 draft, when they get access to 2029 draft picks.
I don't think we're going to see teams respond by making picks fully unprotected, which has the potential to backfire along the lines of what we saw when Detroit got the No. 2 pick in the 2003 draft from the Memphis Grizzlies thanks to a 1998 deal sending Otis Thorpe to the Grizzlies with declining protection over time.
Instead, we're seeing teams offer shorter windows where the pick converts to second-rounders after a certain point. That's what Denver did with the pick headed to Oklahoma City, which ultimately allowed the Nuggets to make the Gordon trade.