clock menu more-arrow no yes

Filed under:

538 finally favors the Warriors to win the title

It’s been a long journey to win over Nate Silver and his computer model, but the Warriors are almost there.

2022 NBA Playoffs - Golden State Warriors v Dallas Mavericks
Andrew Wiggins, presumably reading about how Nate Silver thinks he sucks.
Photo by Noah Graham/NBAE via Getty Images

Game 5 was a huge night for the Golden State Warriors. They got one game away from their fourth championship in eight years, they pushed the Boston Celtics to the brink of elimination, and they finally became the title favorites for Nate Silver’s FiveThirtyEight.com.

It’s been quite a journey for the Warriors, who are now a 63% favorite to win one of the next two games and capture the title. Although, if you considered the teams to be of equal talent, the Warriors would have a 75% chance, so the model still mostly doesn’t believe in them. They gave the Dubs a 48% to win Game One (accurate), a 43% chance to win Game Two (off but close), a 25% chance to win Game Three (right on), a 23% chance to win Game Four (nope!), and just a 47% chance to win Game Five (a smaller nope). Basically, the Warriors are up 3-2 on the Celtics and on the robots. They’re not just fighting for an NBA title, they’re fighting against the singularity.

If you’re wondering, their robots give Golden State just a 26% chance to win Game Six, only slightly lower than the 28.6% chance Nate Silver gave Donald Trump of winning the presidency on the morning of Election Day in 2016.

Before the series began, Neil Paine wrote a piece called, “We Might Be Overrating The Celtics, But You’re Probably Underrating Them”, which was justifying the site’s robot overlords’ projection that Boston had an 80% chance to beat the Warriors. The article acknowledged that the Warriors won two more games against Boston against a slightly better schedule, but credited the Celtics’ eight-game underperformance versus their Pythagorean record to them being “the unluckiest team in the league.”

But that’s still more optimistic than Team Silver’s pre-season evaluation of the Warriors, which had the team 11th in the West with a 37-45 record, with a 22% chance of making the playoffs at all, and a 0.5% chance of making, much less winning the Finals.

We next have to address the Warriors, who are projected to finish worse than they did last year and miss the play-in tournament entirely. And that’s not because the model isn’t accounting for Klay Thompson playing games. It’s actually the opposite. 5 RAPTOR penalizes players who suffer Achilles tears, as Thompson did prior to last season. Throw in the fact that he was already coming off a torn ACL prior to that, and the model now views him as a net-negative player. The model views the Warriors, like the Lakers, as being chock-full of negatives, with only Stephen Curry, Draymond Green and Gary Payton II (The Mitten!) grading out as net positives, among the team’s projected rotation players.

The model also controversially put the Lakers 9th in the West, though it actually overrated them by nine games, by projecting 42 wins. They were 12 games too conservative on Memphis (44 projected wins), 12 games too optimistic on New Orleans (the computers weren’t given Zion Williamson’s off-season weight numbers), and had the 27-win Blazers winning 48 games, though to be fair, no robot could have anticipated the level of tanking Portland did. The human capacity to lose on purpose still baffles our mechanical friends.

Even after the season, the robots gave Golden State just a 22% chance to make the Finals and a 9% chance to win it all. After beating the Nuggets in Round One, those odds didn’t change at all. After going up 3-2 on Memphis, their title odds actually dropped to 8%, and going into the Conference Finals, the Warriors were slight underdogs to Dallas (54-46) and still had just a 10% chance of winning the title. Of course, Silver also declared La Taqueria had the best burrito in the country, when everyone knows it’s not even the best burrito in San Francisco - El Farolito is the best, of course.

Look, 538 is backing up their (stupid, inaccurate, Klay Thompson-slandering) computer model, in a loyal but misguided attempt to justify why they even exist as a sports site. They employ some excellent journalists, but it’s all centered around the prediction models (RAPTOR, which outlasted its successor metric PREDATOR). It’s like Nate Silver defending police brutality because some cops let him eat a burrito in his jail cell - misguided, and proven to be irrelevant.

But for now, the Warriors have won over even the computer living under Nate Silver’s Park Slope home, gears whirring away, frantically trying to add inputs for the effects of cold ocean water and podcasts per game to see why this team has defied the expectations of doctors, bettors, and washed up radio hosts. And if they can close out the Celtics, there’s a 63% chance they can celebrate with El Farolito - I recommend the el pastor super burrito, and pay the extra 75 cents for chips. My computer model says it’s a bargain.