NBA playoffs: Analytics predicts who will go to the NBA Finals
We already know which teams are going to the NBA Finals, because adjusted net rating told us so.
Professional sports are complicated. If they weren’t, they wouldn’t be that interesting. The amount of variables that go into a single possession on the hardwood go beyond the ten players on the court, and include things that can’t be quantified like momentum.
Unlike the NFL, the NBA doesn’t have to worry about weather adding another variable, but sightlines in different arenas can cause issues for visiting teams as does the way each game is officiated differently than the last.
This year’s NBA playoffs have been defined by injuries, and even two of the four teams left standing going into the conference finals have some to deal with. The Boston Celtics are the prohibitive favorites to win the Larry O’Brien Trophy despite missing Kristaps Porzingis to start the Eastern Conference Finals, while the Dallas Mavericks found a way to knock off the top-seeded Oklahoma City Thunder despite a noticeably hobbled Luke Doncic.
On the other hand, the Minnesota Timberwolves have been mostly healthy except for Mike Conley missing a game with a calf injury before returning for the final two games of their series against the defending champion Denver Nuggets, while the long-shot Indiana Pacers lost Bennedict Mathurin to a shoulder injury a couple of months ago but are healthy otherwise.
Pacers star Tyrese Haliburton has not looked like quite the same player he was in the first half of the season after returning to play through a strained left hamstring earlier in the season, but controlled Game 7 against the New York Knicks with 26 points, six assists and four rebounds.
Given that start, injuries are likely to be the great unknown heading into the next round. No amount of data can predict what player might go down or how quickly they will recover in between games.
Despite that, what we hope to predict here is which two teams will advance to the NBA Finals by using analytics. In the NFL playoffs, we have used DVOA to predict each round over the last two seasons. Unlike the NFL, the NBA is full of different advanced data that helps put the game into perspective. Data points like Player Efficiency Rating, LEBRON, Win Shares per 48 minutes, Win Probability Added, Real Plus-Minus, RAPTOR and others try to help put the game into context.
For our predictions, we will be using advanced net rating (aNet) which takes into account a wide variety of data points, including adjusting for schedule, to boil down a team to simple numbers to help us predict the outcome of the conference finals.
Like any analytical model, aNet is not considered predictive of what will happen. Instead, it provides a data point of what has happened to help rate teams.
But even while aNet may not be considered predictive, it does provide us with quality data points as a guide for the upcoming conference finals.
Eastern Conference Finals
Overall aNet: Boston 1st vs Indiana 13th
Offensive aNet: Boston 1st vs Indiana 2nd
Defensive aNet: Boston 2nd vs Indiana 24th
A quick look at the offensive ratings helps explain why the over-under in the ECF is 15.5 points higher than that of their counterparts in the West. The huge gap on the defensive side of the floor explains the overall difference between the top-rated Boston and an Indiana squad that is closer to the middle of the pack.
Edge: Celtics - With Boston the overwhelming favorite in the series, it is no surprise to see such a big gap between the teams overall. It will be interesting if the Pacers look to push the pace and create a shootout given their limitations on defense.
Western Conference Finals
Overall aNet: Minnesota 3rd vs Dallas 14th
Offensive aNet: Minnesota 17th vs Dallas 7th
Defensive aNet: Minnesota 1st vs Dallas 18th
This is a far more interesting set of data, with each team being significantly better than the other on one side of the game. Digging into the data beyond the ranking data shows why the overall aNet rating is so large. The Mavericks are just 2.5 points better than the Timberwolves on offense, but are 6.4 points worse on the defensive side of the ball. With Kyrie Irving and a hobbled Doncic leading the backcourt, it will be tough for Dallas to be drastically better on the defensive side of the ball.
Edge: Timberwolves - Irving has the deep playoff experience from his time with the Cleveland Cavaliers, but Conley’s veteran presence helps steady Minnesota’s young and energetic bunch. The Mavericks' best chance is to ride the scoring prowess of their dynamic backcourt in order to overcome the huge defensive advantage of the Timberwolves.
As noted in the beginning, professional sports are complicated. That is what makes them so fun. The nuance of the NBA, along with its high-flying nature, provides a level of excitement and feats of athleticism rarely seen in the common person. But boiling that game down to a simple number is an attempt to help understand the game better and, perhaps, win a little money as well.