This edition of the NBA mailbag features questions on the effects of preseason travel to Asia, a difficult early schedule and spot free throw shooting.
"Is traveling to Asia during the preseason to blame for slow starts by the Kings, Nets, Pacers and Rockets?"
-- Chris R.
Starting with this question from one of our editors because it's something I've been curious about during the first two weeks of the season. So far, only the injury-riddled Golden State Warriors have underperformed their preseason projections using ESPN's real plus-minus more than the 1-5 Kings and the Rockets, who are 4-3 but have been outscored by 3.8 points per game. Meanwhile, the 3-4 Nets and 3-3 Pacers have also started slower than expected.
To try to answer this question based on historical data, I looked at a pair of factors:
How well teams that have traveled to Asia have done compared with their preseason over/under win totals.
How their performance early in the schedule -- evaluated by point differential adjusted for opponent and location -- compared with the course of the full season.
The latter looks at whether there's a jet-lag/fatigue effect after returning, the former at whether that hurt the team's overall record.
Historically, the answer is no. Over the past decade, 20 teams have played preseason games in Asia, mostly in China. As a group, these teams have played almost exactly to their preseason over/under totals, going under by an average of 0.3 wins. And their performance over the first five, 10 and 20 games -- as well as the first half of the season -- has exceeded their season-long results.
That said, we still want to monitor whether shortening the NBA preseason has made travel more difficult. Starting in 2017-18, the league shaved about a week off the period from the start of training camp through opening night, and over the past two years, the four teams that played preseason games in China have gone under by an average of 4.3 wins. Meanwhile, three of the four have gotten off to relatively slow starts through their first 10 games, with the 2017-18 Warriors -- healthy at the start of the season before dealing with injuries later -- being the lone exception.
There's also the question of whether travel to India might be more taxing than playing in East Asia, and our sample size there is solely two teams thus far. But it's also tempting to succumb to confirmation bias, where slow starts by teams that played preseason in Asia are held up as evidence of the effect and strong starts by other teams (the Lakers and Raptors, both among the 11 teams most outperforming their RPM projections) are ignored.
A broader question but is there any evidence of a frontloaded/backloaded/evenly distributed strength of schedule being beneficial or detrimental? (If you're reading this Seth, apologies for double timing you) #peltonmailbag— EuanDewar (@EuanDewar) October 30, 2019
Staying on the subject of early-season schedule/performance, I'd say the conventional wisdom is that it's preferable to have an easy schedule early on to build confidence, whereas a tough schedule can discourage teams. And there doesn't really appear to be any evidence for this over the past decade.
Again, I've rated schedule strength relative to the full season over the first five, 10 and 20 games as well as the first half of the schedule. There's no relationship between any of those measures and how well teams perform compared with their preseason over/under total. The schedule really does seem to even out no matter how it's distributed.
There is a modest relationship between having a tougher schedule in the second half of the season and having a better record. But there's a simple explanation for this: The national TV schedule is backloaded with good teams playing one another after the NFL season ends, so there's actually a stronger relationship between the preseason total -- a measure of presumed ability entering the season -- and a more difficult second-half schedule.
"Your model has Dallas as a playoff team, but it seems a lot of other pundits who did not have the Mavericks in the postseason are higher on their playoff chances through four games. Is this small-sample-size theater?"
Yes and no. If the question is strictly whether we learn that much from teams based on their first four games, the answer is no. Performance over the first four games of the season explains about 35% of the variation in winning percentage over the full season, which is to say still not very much. A team's preseason win projection, for example, explains about 60% of the variation in their final record. It takes almost 15 games for performance to date to become as predictive.
Based on this track record, if you thought Dallas was a .500 team entering the season, the way the Mavericks have started should add only a couple of wins to their final estimate -- probably not enough to make them a likely playoff team.
At the same time, I'd guess people who have rethought the Mavericks' outlook are probably also taking into account the fact that statistical projections were higher on Dallas entering the season. If you've decided that the first handful of games back up those more optimistic preseason expectations, then it's perfectly reasonable to recalibrate after only a few games.
"To continue the free throw conversation, I've always thought it odd that no matter where a player gets fouled, he then has to or gets to take uncontested shots from 15 feet. What about taking a soccer-like approach and giving players one uncontested shot from the spot of the foul (and still shoot from the FT line for non-shooting fouls). Seems like this would cut down on the intentional fouls near the basket and not "punish" bigs like Andre Drummond for getting fouled shooting near the rim. I'm sure there's a reason not to do it but wanted your thoughts."
-- Matt Jasinski
A number of people had similar responses to last week's mailbag on the topic of shooting one free throw for all points and whether shooting fouls on 3-pointers should yield two free throws instead of three.
Certainly, an uncontested shot from the spot of the foul makes sense for jumpers. I'd guess the conversion rate on 3s would be fairly similar to what we see in the 3-point contest, where the best shooters in the league typically make a little better than half of their attempts. That would put us in the sweet spot where a foul is clearly better than a contested attempt but not so valuable as to make those calls huge game-changers.
Finding that sweet spot around the basket would be more challenging. Soccer solves this with penalty kicks rather than free kicks inside the box, but I'm not quite sure how you'd handle where the player shoots from and where other players line up for a possible rebound in basketball.
If you did it strictly like a soccer free kick where other players can't be within a certain distance, that would seem to guarantee two points for any foul near the basket. That might not be the worst thing in terms of taking away semi-intentional fouls on those players, but it would encourage offensive players to do everything possible to draw a foul.