NBA In-Season Tournament by the numbers, plus suggested future improvements
With the completion of the last of the 60 group play games on Tuesday night, we can draw a line through the group stage of the first installment of the NBA ’s In-Season Tournament.
While it has not been perfect — we have a few format suggestions below — the launch can’t be described as anything but a major success with the games drawing better ratings and more discussion than is standard for this part of the season.
Moving past the pageantry and increased interest relative to past Novembers, I was curious as to whether there were any discernible differences in play between tourney and non-tourney games through Tuesday, Nov. 28. Certainly, some teams performed better in tourney games than the rest of the schedule, and unsurprisingly the top eight teams by Net Rating are the ones advancing:
Of those eight, only the Boston Celtics failed to improve on their non-tournament play, but maintaining that level of play — the league’s best in non-IST situations — was enough to see them through.
The Los Angeles Lakers — with the largest difference in performance between the two types of fixture — had an average Net Rating 26.4/100 higher in tourney games. One of the main differences in league aggregate performance is the 37.3 percent 3-point accuracy on Tuesdays and Fridays compared to 35.7 percent in normal regular-season play. The Lakers are responsible for around a third of that gap, going from 29.7 percent in non-tourney to 47.5 percent in their four tournament contests. Every other team in the league was within seven points of their non-tourney accuracy, making the Lakers 17.8 point gap stand out all the more.
Interestingly, the other Los Angeles team had the largest decline relative to non-tourney play. The Clippers had an average Net Rating 12.4/100 worse in group play than the rest of the season.
Overall, the league is slightly more efficient in group play — an average offensive rating (ORtg) of 114.3 compared to 113.0 — with the main driver being increased 3-point accuracy. I’m reluctant to ascribe this to much more than variance. While it would be tempting to posit that since tourney games were never the second night of a back-to-back, those fresher legs might be the culprit. This season, teams have been slightly more accurate (36.5 to 36.0 percent) during those second-night games.
While not nearly to the same extent as in the playoffs, especially not in the later rounds, teams did shorten their rotations somewhat. Just under 1.4 players per team per game have seen 35 or more minutes in each contest during non-tourney play, while a little more than 1.6 did so in tourney group play. On the other end of the spectrum, teams gave 10 minutes or more to 9.1 players in non-tourney games compared to just under 8.9 in tourney play.
As mentioned earlier, similar patterns of bottom-end rotation players getting squeezed out with more playing time concentrated in the top few players is something typical of the postseason, but the effect, depending on the playoff round, is three to five times larger in the postseason than was seen in the tourney.
Additionally, some of this concentration of minutes could be the result of slightly more tourney games resulting in some amount of “clutch” with a margin of five points or less at any point in the last five minutes. Thirty-four of the 60 group games (56.7 percent) compared to 99 of 197 (50.3 percent) of other games saw games reach that state.
While it took teams a few rounds to figure out the importance of point differential as a tiebreaker, the incentive to leave stars in to chase a big margin was present.
This brings me to a few critiques and/or suggestions. None of these are meant to knock the concept or execution of the tourney. Even though I thought there were more misses than hits in the court designs themselves, the idea of having specialty courts for these games is a good one and should be kept. This is a perfect illustration that the league knew the first iteration of the tourney wouldn’t be perfect, but improvements would only be revealed through experience, and I think there are a couple of lessons to be had.
For one, I think the idea of a point differential as a tiebreaker is sound, I would consider replacing it or at least moving it lower in the tiebreak order for a few reasons.
First, though the schedule-makers did a great job of “protecting” tourney games by not having them occur as the second leg of a back-to-back, many were first legs, which creates some weird incentives and decisions balancing the desire to maximize scoring margins with the need to have something in the tank for the following night’s game. The schedule is complex enough that ensuring tourney games never have to serve as the front half of a BTB might not be achievable.
Second, while there was an amount of drama and comedy in Golden State, realizing they needed to beat Sacramento by 12 and emphasizing hitting that target rather than winning the game and Boston intentionally fouling Andre Drummond , allowing them to run up a big margin on Chicago, it’s not hard to imagine a scenario where two teams reach a comfortable margin, which would mean both teams advanced.
Soccer tournaments with group play used to run into this problem, the most famous example of which was the West Germany-Austria match at the 1982 World Cup, which saw the sides sleepwalk through a 1-0 West German victory with both teams knowing that a one- or two-goal margin in favor of the Germans would see both advance at the expense of Algeria.
Since then, the final round of games in each group for tournaments such as the World Cup or European Championships has been played simultaneously. In fact, the final day of league play is treated the same way in many countries. While the flow of games sometimes still allows the two teams in a game to realize mutually advantageous outcomes, it’s harder to keep track of the scenarios in real time. If point differential remains a key tiebreaker, this is a must.
But I would prefer deemphasizing point differential and instead making something like quarters won the standard. Older fans might remember that the old Continental Basketball Association had a similar system whereby instead of organizing the standings by wins and losses, there were 7 points at stake every game. The team ahead at the final buzzer received 3 points, with an additional point awarded for each quarter won. While I wouldn’t go that far with the tourney, I do think that winning quarters would serve a similar purpose as point differential while removing some of the weird side effects.
Over the next few seasons, I’m sure we’ll find a few more of these edge cases, as the teams themselves become more familiar with the possible scenarios and game out ways to maximize their chances of progression. I look forward to seeing what those innovative tactics might be.
But for now, I’m happy this is part of the NBA landscape and am very much looking forward to the quarterfinals, where I expect to see a genuine playoff atmosphere.
(Photo of LeBron James : Tyler Ross / NBAE via )
