What Do the Numbers Say? Probability of the Cavs Beating the Dubs

Ahead of the NBA conference finals this Sunday, many have already come to the conclusion that the Cavaliers and Warriors will meet again in the NBA finals, for the record fourth straight time. For some, this is an exciting rematch. For others, it is another wasted opportunity for good basketball as they expect the Cavs to receive another beatdown. The dubs won 6-4 in 2015, the Cavs came back 4-3 in 2016, and dubs won again 4-1 in 2017.

While the match up has not always been close, these two teams have played a total of 18 finals games against each other, which provides a decent sample for a logit probability model. Using stats from these 18 games,  I created logit models to estimate the probability of a Cavs victory given Cavs’ performance. So, what do the numbers tell us? Is the Cavs on pace for another beating? Or is there a chance that the “new squad” can takedown the dubs?

“We got a fucking squad now” – J.R. Smith

The Data

I gathered data from the 18 finals games the Cavs and the Dubs have played since 2014. The result column has 1 for win and 0 for lose. The other columns contain key game stats: Field Goal % (FG%), 3-point % (3%), total rebounds (TRB), assists (AST), turnovers (TOV), and points (PTS). The three data sets below organize by the game stats by Lebron James (LJ), the rest of the team (R), and the entire team.


I created three logit models that estimate the probability of a Cavs win using different game stats as variables. The way to interpret this table is that a unit change in the game stat leads to a certain percentage point change in the probability of a Cavs win. For instance, the Logit Model 1 says that an additional point scored by Lebron James increases the probability of a Cavs win by 9.62%. (Certain coefficients do not make sense such as an increase in field goal percentage leads to a decrease in the probability of a Cavs win. This is likely caused by an outlier game in the 18 games sample) Overall, these three models show that points, assists, and 3-point percentages matter a lot!

Now, here is the fun part. These models tell us the importance of these game stats to the probability of a Cavs win, but we can also plug-in stats to find the probability of a Cavs win. The line below is Lebron James’ average game stat against the Dubs in the Finals. Even with his near triple double performance, Logit Model 1 says that the probability of a Cavs win is only 32.48%!

The plot below shows that if we hold Lebron James’ game stat at his finals average, but vary his game points, then the probability of a Cavs win increases up to 65%.

We already looked at what happens if we plug-in Lebron James’ average finals game stat. The table below shows the probability of a Cavs win based on different scenarios. For instance, if Lebron James performs his best and the rest of the team hits their finals high of 56.76% for their 3-point %, then the probability of a Cavs win is almost guaranteed at 97.43%!

The most important analysis here is to see what happens if plug-in game stats from the series against the Pacers and the Raptors. The results show what our eyeballs already told us. The Cavs better bring their “Lebronto” selves to the Finals if they want to beat the Dubs. If Lebron performs his Toronto average and the rest of the team hits their average 3-point %, then the probability of a Cavs win is 58.44%. If they perform their best, then the probability is as high as 83.09%.

Knowing this is helpful, but what about the probability of the Cavs beating the Dubs in the best of 7 series? The table below is just one of the paths, but things do not look good for the Cavs.

Combining these probabilities with some general statistics class probability work, I came up with the probability of the Cavs and the Dubs winning the chip in 4, 5, 6, and 7 games.


In layman terms, this says that even if Lebron has his average finals performance (which is amazing already) for G1 and G2, the Cavs bring their average and best Lebronto-selves for G3 and G4 at home, and Lebron has his best finals performance for G5, G6, and G7, the total probability of the Cavs winning in 4, 5, 6, and 7 games is only 17.02%, compared to the Dub’s 77.11%.

Yeah, the numbers say this will be another tough June for the Cavs.

MVP and “Best Player on Best Team”

Investing the notion that the NBA MVP traditionally goes to the best player on the best team.

Every year when talking about who should win NBA MVP, the following argument is always brought up:

MVP traditionally goes to the best player on the best team, therefore Player X should win because he is the best player on the best team.

I’ve always been bothered by this argument because of two reasons. One, it’s a lazy argument. People don’t have to watch a second of basketball to argue this. Two, if MVP always went to the best player on the best team, there would be nothing to debate besides whether Kevin Durant is better than Steph Curry.

Now, to investigate if this argument is actually true. Does MVP actually usually go to best player on the best team??

Continue reading “MVP and “Best Player on Best Team””

[Fastbreak Data] Using Machine Learning to Find the 8 Types of Players in the NBA

Alex Cheng is Founder of Fastbreak Data and a guest contributor for Ball and One. Follow his work at fastbreakdata.com.


Using machine learning, my goal is to uncover the positions that are intrinsic to today’s NBA players and classify players with a position that best encapsulates their skill sets.

Inspired by the topological data analysis of Muthu Alagappan in “From 5 to 13: Redefining the Positions in Basketball” and the “Periodic Table of NBA Elements” by Stephen Shea in Basketball Analytics: Spatial Tracking, this study proposes an alternative method in which to classify players in today’s NBA….

Click the image below to read the full article. Or you can click here to read it on my blog, Fastbreak Data.

Click on the image above to view the full article at fastbreakdata.com

Why the Golden State Warriors are Absolutely Fine

Let me just say that all this fanfare about the Warriors being in trouble is way overblown.

A team on pace to win 64 games hits a rough patch, and all of a sudden it seems like the entire sports media world piles on with completely baised and random stats to prove how much the team is in trouble.

Only a half game ahead of the Spurs! Stephen Curry 0-5 with under 10 seconds remaining! Warriors playing worse without KD (who saw that coming?)! Nevermind that the Warriors have played 8 games in 13 days, all in different cities, and 17 of the last 24 games on the road. At the end of the day, after the trade deadline, and even after KD’s sprain, the Warriors are still favored to take the championship this year – and for good reason. Let’s dive into the numbers.

Continue reading “Why the Golden State Warriors are Absolutely Fine”

What on earth is Phil Jackson doing with the Knicks?

Breaking down Phil Jackson’s bizzare tenure as GM

The NBA is a zero-sum game. For every win a team earns, another takes a loss. For a champion to be crowned, 29 other teams must lose. Every GM has a different way of approaching the game; some are in win-now mode, some stash picks and look to the future, and some are patiently waiting for their blossoming young stars to lead the way. There is no right answer; history has shown us that there is no single guaranteed path to winning a championship.

However, at some point in a GM’s career, whatever strategy he has chosen has to start producing Wins. At some point, your team is good enough, or it’s just pretending (Clippers). At some point, all your high draft picks have to turn into players (76ers). At some point, your #1 picks have to develop into stars (Timberwolves). No matter how clever or innovative a GM is, at the end of the day, the goal isn’t to stash draft picks or assets, the goal is to Win.

Continue reading “What on earth is Phil Jackson doing with the Knicks?”

Musings on the first third of the 16-17 season

Musings on every team for the first third of the 2016-17 NBA season.

Merry Christmas and happy 2/3 of the NBA season left! I’ve been able to watch a little bit of every team and to celebrate the holidays, I’ll be going through each one and giving my take on they stand. Team records are before games on 12/19.

Continue reading “Musings on the first third of the 16-17 season”