Now that most of the NBA teams have passed the halfway point of the season, we recalculated our end-of-season win total predictions and playoff probabilities. In truth, we don’t think that the second half of the season matters that much, in terms of playoff probability.
Since the top eight teams from each conference make the playoffs, the likelihood that even the top four teams in wins per conference at mid-season would fall out of the top eight at the end of the regular season is astronomically small. However, the second half does matter in terms of the teams fighting for the last couple of spots, as well as, of course, playoff seeding.
How accurate will our predictions be? Well, at this point last season, every team with a greater than 90% of making the playoffs did, our top 7 teams (and 8 of our top 9) in playoff probability in the Western Conference made it into the postseason, and our top 6 teams (and again, 8 of our top 9) in playoff probability in the Eastern Conference made it into the playoffs. Our biggest miss was the Detroit Pistons, who collapsed epic-ally in the second half. We expect our predictions to be just as accurate this year.
If you need a refresher on our methodology, feel free to check out our 2018-2019 preseason playoff predictions, or our look back at the 2017-2018 NBA season. Since we did make one small change, here is a quick refresher.
At the beginning of the season, we gather win total predictions from a variety of sources, including casinos, ESPN, and other sports websites. We then average, or take the median, all of these predictions together to generate an expected win total for each team. By a phenomenon called Wisdom of the Crowds, this win total is usually more predictive than any one single estimate.
Using this average, and a variance determined by historical data, we are able to calculate a Beta distribution to model each team’s expected winning percentage for the upcoming season. We call this distribution a Prior Distribution, since it is calculated prior to collecting any data from the season.
For example, here is the Golden State Warriors’ (GSW) prior distribution for the 2018-2019 season. The beta distribution was determined to be Beta(19.82, 6.18). As you can see, we expect them to have a win probability between 0.600 and 0.850 for this season. As of 1/12/2019, they have a win probability of 0.667.
As we did for the 2017-2018 season, after each game we increment the first parameter of the prior distribution by the number of wins a team has, and the second parameter by the number of losses.
All data below uses NBA standings as of 1/12/2019.
For example, the GSW have a record of 28-14. Their Beta distribution on this date would be Beta(19.82 + 28, 6.18 + 14) = Beta(47.82, 20.18). We call this a Posterior Distribution, since it is calculated after we have seen data from the regular season.
In addition to posterior distributions calculated by win-loss records, we also calculated a second posterior distribution using Pythagorean win-loss records. Just as in baseball, Pythagorean win-loss records are determined by the number of points scored and allowed by each team so far on the season.
The NBA Pythagorean formula is: Runs Scored^16.5 / (Runs Scored^16.5 + Runs Allowed^16.5).
While GSW has a win percentage of 0.667 for a win-loss record of 28-14, they have a Pythagorean win percentage of 0.697, for an expected win-loss record of 29-13. This indicates that GSW has been slightly unlucky in the first half of the season.
Since a team is more likely to regress to their Pythagorean win percentage as the season progresses, we used a team’s expected win-loss record to calculate a, hopefully more accurate, posterior distribution. In the case of GSW, that would be Beta(19.82 + 29, 6.18 + 13) = Beta(48.82, 19.18).
Below are the two posterior distributions calculated above for GSW. While they don't look too different, since they only had a swing of 1 win, for teams like Boston who have a Pythagorean record of 5 wins higher than their actual record, it can really swing a team's playoff probability.
Using the posterior distributions for each team, we calculated playoff probabilities and expected end of season win totals by performing Monte Carlo simulation. For example, for the GSW, we can sample from Beta(48.82, 19.18) for a sampled winning percentage, say .650. With 40 games left in the season, that would put GSW at .650*40 = 26 more wins, for an end of season expected win total of 26+28 = 54. After 100,000 samples, we get an average end of season win total, and can calculate playoff odds by seeing how often a team finished in the top 8 in total regular season wins.
Here are the expected win totals and playoff probabilities for each conference, first using a team’s current win-loss record.
These next graphics use a team’s Pythagorean record to predict end of season records and playoff probabilities.
Most of the Eastern conference is already set, with the top 5 teams set to have a minimum of 10 games over the 6th seed. Only the last two seeds really have any competition. We have Toronto, Philadelphia, Indianapolis, Boston, Milwaukee, and Miami as locks to make the playoffs, with Charlotte and Brooklyn bringing up the last two spots with 85% and 71% playoff odds respectively. If one of those two does fall out, either Detroit or Washington will take the last spot.
The Western Conference consists of a much tighter race. We predict 7 teams to finish within 5 wins of each other, yet only 4 of them will make the playoffs. Our locks are thus only Golden State, Denver, Oklahoma City, and Houston. The next 7 teams, in terms of playoff probability, are San Antonio, Portland, LA Lakers, Utah, LA Clippers, New Orleans, and Minnesota.
Here are our predicted standings:
At the beginning of the season, we used our preseason predictions to try and win $1,000,000 through NBA’s pick-em challenge. If our mid-season predictions hold, here are the number of picks we would get right:
While in the East, we are only projected to miss on the Wizards, the West is a different story. The West is projected to be a lot closer, in terms of spread, than the East, which decreases the parity. We had a few close calls, but it look like we’ll miss on Suns, Trail Blazers, Spurs, Kings, Jazz, Mavericks, and Rockets if our projections hold sadly.
We have to hit at least 25/30 to be entered for any prize. Wish us luck, we’ll need it.
Is your favorite team going to make the playoff? Do you disagree with any of our seedings? Let us know in the comments below! As always, our code, data, and visuals can be found on our GitHub. Be sure to check back regularly as we update our NBA playoff odds and expected wins daily!
The SaberSmart Team
P.S. If you enjoyed this article, and need something off Amazon anyway, why not support this site by clicking through the banner at the bottom of the page? As a member of the Amazon Affiliates program, we may receive a commission on any purchases. All revenue goes towards the continued hosting of this site.