This week's mailbag features your questions on Game 7 win probability, 3-point shot quality, and more.
You can tweet your questions using the hashtag #peltonmailbag or email them to peltonmailbag@gmail.com.
"Game 7 at home is an 80% proposition."
— Ruminative Orangutan (@Ruminorang) April 21, 2018
Agree to disagree. @kpelton #peltonmailbag
In a typical series, a home team is expected to win Game 7 about 80 percent of the time because historically that has been the case. NBA teams are 102-26 all time in Game 7s, almost precisely an 80 percent winning percentage.
So should the Cleveland Cavaliers be overwhelming favorites Sunday against the Indiana Pacers despite a dismal first-round series thus far? Let's dig a little deeper.
The historical home-court advantage in Game 7 is a product of two things: the edge provided by playing at home as well as the fact that the higher seed with home court was likely better during the regular season. Indeed, despite the fact that the two teams have by definition split the first six games of the series, since 1984 the typical Game 7 host had a point differential 2.3 points per game better during the regular season than the visitor.
Subtracting that from the average home margin of victory in Game 7s (8.6 points per game since 1984, when the playoffs first expanded to their current 16-team format), we get that home court is worth about 6.3 points in Game 7 -- more than double the typical home-court advantage during the regular season in the modern NBA, which has averaged about three points per game since 2000.
As my former colleague Tom Haberstroh explored in 2015, home-court advantage has been in decline in the NBA. This season's margin (plus-2.1 points per game) was the lowest on record, replacing the previous low for a full season -- a plus-2.4 advantage in 2014-15.
Naturally, that trend has carried over to the playoffs. Though home court fluctuates wildly from postseason to postseason because of the smaller number of games, regression analysis suggests a decline of about one point per game in home-court advantage from the mid-1980s to today -- not quite as pronounced as in the regular season, but notable nonetheless.
Since there are so few Game 7s, any change in home-court advantage would be more difficult to detect (because of the small sample size). And in fact, over the past five postseasons, home teams have actually done even better, with a plus-9.1 margin of victory. So it's probably worth sticking with the longer-term estimate of home-court advantage in Game 7.
Now let's get back to the Cavaliers-Pacers series, which is unusual for a couple of reasons. First off, Indiana (plus-1.4) actually had the better point differential of the two teams during the regular season despite Cleveland (plus-0.9) winning two additional games to claim home court. Based strictly on regular-season results and the historical home-court advantage in Game 7, we'd give the Cavaliers about a 70 percent chance to win.
Second, the Pacers come in off a 34-point win in Game 6 that gave them a plus-44 margin for the series. No other seven-game series since 1984 has seen the lower-seeded team so dominant through the first six games. In series since 1984 where the lower-seeded team outscored the higher-seeded team through the first six games, the home team has gone just 22-10 (.688) since 1984. Four of the six teams with the best cumulative margins through six games went on to pull the Game 7 upset on the road.
So in this case, I'll agree with you that 80 percent surely overstates the Cavaliers' chances of winning Game 7. The ESPN Forecast panel gave Cleveland a 61 percent chance of winning, and that's about where I'd land.
Intriguingly, the Cavaliers enter as 5.5-point favorites and with an implied 70 percent chance of winning based on the moneyline, consistent with the regular-season analysis above. But I'd assume at least some of those figures are based on the respect given to LeBron James in a Game 7 rather than anything we've seen from Cleveland in this series.
"After LeBron James went 15-of-15 on free throws in Game 5, what's the lowest season free throw percentage to make all of 15 or more attempts in a game?"
-- Stephen Boesch
It's tricky enough just getting 15 free throw attempts, let alone making them all. So Basketball-Reference.com shows just 121 players doing both since 1963-64, with James joining that group earlier this week. Here are the players who did so during a season (or the playoffs after a season) with the lowest free throw percentages:
While the full group of players who have gone perfect on at least 15 free throw attempts -- including players not listed above -- is naturally solidly above average at making free throws (their weighted average is 83.9 percent shooting during the season in question), it's possible to have such an accurate, high-volume night without being a good free throw shooter overall.
"How rare is it that there is only one first-round sweep? A follow-up question: how about where the single sweep is of the higher-seeded team?"
-- Wes Ivers
Since the first round went to a best-of-seven format, there has been an average of 1.6 sweeps per year. So having only one is not particularly unusual. It last happened in 2014 and has now happened seven times in the past 16 years. Having no first-round sweeps is notably unusual: the only time it happened was 2003, the first year after the change.
As for the second part of your question, a sweep by a lower-seeded team is quite rare. It has happened only twice in the first round in the best-of-seven era: this year's Pelicans and the 2015 Washington Wizards, who swept the higher-seeded Toronto Raptors.
"There seems to be a huge amount of blowouts happening in the league, and especially in the playoffs. If I had to take a guess, I would say it's because the strong uptick in metric-based offenses running on efficient 3-point shooting and layups. Is there a measurable difference in the margin of victories in recent years? Could it be linked to the volatility of increased 3-point shooting?"
-- Ian Stratton
This was true last season, as I noted in a mailbag at the time, but not so much so far this season. There have been only six games decided by 20 or more points, a relatively lower rate (14.0 percent of all games) than the typical postseason since 2003 (16.1 percent). The average margin of victory has been 11.8 points per game, a little lower than the average for the first round since 2003 (12.0).
If we look more generally, margins of victory and the percentage of games ending in blowouts have increased somewhat since 2003. I'm not sure how much of that to attribute to 3-pointers -- given, as I explored when discussing the Houston Rockets this year, their impact on a team's consistency is overstated -- as compared to growing stratification of the league with stars teaming up.
also probably the case with the Wolves defense. Maybe a question for #peltonmailbag, is there a tradeoff between better shot quality from 3 and more 3 point attempts?
— Krishna Narsu (@knarsu3) April 23, 2018
It seems natural that there would be due to diminishing returns, but the data is mixed. I took a look at shot quality as measured by Second Spectrum's quantified shot quality (qSQ) measure, which accounts for the location and type of shot and nearby defenders, for teams over the past five seasons as compared to their 3-point attempts per 100 possessions -- both relative to league average for the season. Here's how that looks graphically.
Naturally, the Houston Rockets of recent vintage are off the charts here, attempting way more 3s per 100 possessions than the rest of the league. (The 2015-16 Rockets were a relative outlier in this regard.) After maintaining good shot quality on their 3s at high volume in 2014-15, Houston has had some of the league's lowest-quality 3s by this measure the past two seasons. (Although, as the chart shows, this year's Golden State Warriors actually got slightly worse quality on their 3s at lower volume.)
More generally, the relationship between 3-point attempts and their quality exists but isn't particularly strong. I think that's probably because in addition to diminishing returns, there's another effect at play here. Teams with weaker offenses both aren't generating many 3s and aren't getting high-quality ones when they do shoot them.