This week's mailbag features your questions on injuries to star players, pull-up transition 3s, Tyus Jones and more.
You can tweet your questions using the hashtag #peltonmailbag or email them to peltonmailbag@gmail.com.
"Which non-Hall of Famer had the best peak as a player?"
This question arose from a conversation during All-Star Weekend and I figured it was worth taking a deep dive in the mailbag. As a first cut, let's take a look at the leaders in my wins above replacement player (WARP) statistic since 1977-78 among eligible players who are not in the Hall of Fame, including a few who might soon get there.
As you can see, WARP highly valued '90s point guards, particularly those with the surname Hardaway. So too does Basketball-Reference.com's value over replacement player (VORP) metric, which I've included for an alternative perspective. Conventional wisdom at the time was not necessarily so high on these players. Blaylock wasn't an All-Star in 1996-97, while Francis -- who led the league in VORP -- didn't make any of the three All-NBA teams.
Taking those subjective assessments into account, a stronger case can be made for Penny Hardaway, who was an All-NBA First Team pick and third in MVP voting during 1995-96, when the Orlando Magic had the league's third-best record despite Shaquille O'Neal missing 28 games because of injury. Hardaway's peak was relatively short because of injuries, but before those injuries he was a phenomenal young player.
That said, a few other candidates are coming down the line. Assuming Derrick Rose isn't voted to the Hall of Fame, he'll be the first NBA MVP ever to miss out. However, Rose's advanced stats from his MVP campaign (16.6 WARP, 6.0 VORP) don't match up to Hardaway's.
The statistical contender who would only be picked for the Hall of Fame on the strength of his international career is Andrei Kirilenko, whose 2003-04 campaign rated off the charts by all-in-one stats. Kirilenko rated as worth 19.9 WARP -- more than any current non-Hall of Famer -- and his 7.5 VORP ranked second in the league behind Kevin Garnett.
I haven't yet mentioned ESPN's real plus-minus (RPM) because it has not been calculated back to the 1990s, but Kirilenko was third in the league in RPM, too. He could never again repeat that success, and wasn't picked for an All-NBA team, but there's a case Kirilenko had the best season of any non-Hall of Famer.
"If you weight the number of injuries and the total games lost by WARP, do you agree with your former colleague Tom Haberstroh's conclusion that these metrics have risen significantly compared to seasons past?" -- Tarek Mohamed
Yes, they generally suggest the same thing. Because of the way I've set up my injury database, I can't compare games lost through the same date, but so far there have been an average of 4.6 players out due to injury or rest in each game -- an increase of more than 30 percent from last year's final average of 3.6 per game.
Naturally, there tends to be more games lost to injury late in the season, both because season-ending injuries can only increase in number and because teams out of the playoff hunt are less likely to ask their players to play through minor injuries. (Or, a cynic would note, prefer them not to play so as to improve draft positioning.) So that 30 percent figure should continue to rise. Notably, because I'm including games missed to rest, that increase cannot be attributed to teams saying their players are out with minor injuries rather than resting.
I also track the average playing time and player win percentage (the per-minute version of WARP) for injured players, and that does reveal one interesting trend. The average minutes per game for injured players (weighted by games missed) is down substantially, presumably because I'm including players on two-way contracts in my injury calculations.
Player win percentage is not off as much, which might reflect the number of All-Stars we've seen miss time because of injury. It's also worth noting after looking at this chart that part of the reason this year's increase is so substantial is because the rate of injuries per game was lowest in 2016-17 in four years. As a result, I'm hesitant to ascribe the uptick to any underlying trend. More data is needed to determine what's going on.
"A lot of players love to pull up for three in transition with their momentum still moving forward (see Caldwell-Pope, Kentavious). It seems like that's a more difficult shot than they think it is. Do you have data to show whether these shots are made more or less often than expected, given the player and the shot location?" -- David Harwitt
Second Spectrum data powered by NBA Advanced Stats offers an interesting perspective on this question. Players are in fact shooting a high percentage on pull-up 3s in transition: 43.0 percent collectively, which is an effective field-goal percentage (eFG) of 64.6 percent.
Clearly, that's much better than we'd otherwise expect. But I think the better point of comparison is the other transition attempts players are passing on in order to pull up for 3, and those are even more efficient. Players are shooting 67.5 percent on 2-point attempts in transition and an incredible 54.2 percent on catch-and-shoot 3s in transition, good for an eFG of 81.3 percent.
Of course, the average shot of each type isn't always available on each transition opportunity, so the pull-up 3 may be worth it if there isn't a teammate running open either to the basket or the 3-point line. (The 2-point figure also includes completely uncontested layups or dunks off steals, which are rarely an option.) Still, players may be shooting a few too many pull-up 3s.
That definitely does go for Caldwell-Pope, whose 23 pull-up 3s in transition rank him fifth in the league behind Kyle Lowry, Devin Booker, Kevin Durant and James Harden. Caldwell-Pope's eFG on those shots? A relatively dismal 45.7 percent.
@kpelton RPM currently shows Tyus Jones as the 9th-most valuable per-minute player in the NBA. Can we get a deep dive into other stats to explain if this is for real? #peltonmailbag
- Jeff Allen (@TYTEJEFF) February 1, 2018
Since this question was posed, Jones has dropped a few spots to 12th in the single-season version of ESPN's real plus-minus that is on the website.
If the question is whether Jones is one of the 20 best players in the NBA, the answer is obviously not. Jones' box-score stats show him as a solid backup but generally not as a league-average player; he rates a little worse than average in my wins above replacement player (WARP) metric and a little better than average in Basketball-Reference.com's box plus-minus. Even taking multiple years into account in the version of RPM that has more predictive power shows Jones as less effective, if still better than his individual stats: he ranks 33rd in multiyear RPM with a plus-3.2 rating.
So why has Jones rated so well by RPM this season? The Minnesota Timberwolves have allowed 6.8 fewer points per 100 possessions with Jones on the court, which is particularly impressive because he often shares the backcourt with Jamal Crawford, who currently ranks 480th in defensive RPM.
As you might imagine, opponent 3-point shooting is a factor. Teams are shooting 34.6 percent on 3s against the Timberwolves when Jones plays, per NBA Advanced Stats, and 38.0 percent with him on the bench.
Minnesota has also rebounded better and fouled much less with Jones on the court this season, which his page on CleaningTheGlass.com shows was not the case in 2016-17. Overall, I'm inclined to think that Jones' defensive value is at least somewhat fluky, and without it he looks more like a good backup than a superstar.
