clock menu more-arrow no yes mobile

Filed under:

What the Computers Knew About the Raptors Before We Did

Many statistical models liked Toronto’s chances in the 2019 NBA Finals more than the public. Turns out, math was right. Here are the four things we should’ve weighed more heavily before this close series between the Raptors and Warriors began.

Getty Images/Ringer illustration

Nobody believed in the Raptors. In The Ringer’s NBA Finals preview, every writer picked the Warriors to win, and nobody thought the series would last seven games. The same was true at Sports Illustrated. It was almost true at ESPN, where 19 of 21 writers picked Golden State. And the gamblers in Las Vegas favored Golden State by a considerable amount, picking the Warriors at a nearly 3-to-1 clip.

So nobody believed in the Raptors. Or, no group of people did, anyway. The computers were comparatively enamored of Toronto’s chances. FiveThirtyEight favored Toronto. ESPN’s Basketball Power Index calculated the split at Golden State, 52 percent; Toronto, 48. Basketball-Reference’s formula spit out a result in the same range. A University of Toronto statistics professor’s model gave Toronto a slight edge. Yet even then, few believed the numbers. In a published staff chat, FiveThirtyEight writer Chris Herring wrote, “Our model is wrong. All due respect to our model.”

But after Toronto more than held its own through two games of the Finals, split two games at home, and nearly took a 2-0 series lead, those computers look more prescient than they did a week ago—or at least more prescient than all the people who doubted them. (Like myself, who picked the Warriors in five. Mea culpa.) Golden State still may win the series—and is in fact now favored in all those models—but Toronto still looks like a far more formidable opponent, with a legitimate chance to win and a strong chance to keep the series even. What did the computers see that the public missed? A few factors come to mind.

The Teams Were Equally Good in the Regular Season

The simplest way to predict a playoff series is to pick the better team, and the simplest way to identify the better team is to look at the two competitors’ regular-season results. Eighty-two games is a reasonably large sample, so the superior team during that span is usually better overall, and thus more likely to emerge victorious from a playoff clash.

That heuristic is a bit less useful in this series because the two teams ended up with nearly identical regular-season results. Toronto finished with a 58-24 record, one win ahead of Golden State, while the Warriors posted a better point differential by just 0.4 net points per game. The point here, however, isn’t to pick at those small differences to try to identify the better team, but rather to acknowledge that those numbers are darn close. We should have expected this series to be close, too.

Running a Log5 calculation—a method invented by Bill James to estimate the probability of each team winning a game—with the two teams’ regular-season records shows that the Raptors would have a 51 percent chance of winning a game against the Warriors on a neutral court. Using their point differential during the course of the season (a more powerful predictive measure than overall record), that probability would be 49 percent. So either way, the result denotes a toss-up.

Examining regular-season results is admittedly tricky for teams like the Warriors, who can coast for long stretches in the winter while retaining an extra gear for when the games truly count. But Toronto evinced signs of reaching a higher level in the playoffs, too. Kawhi Leonard, for instance, missed 22 regular-season games as Toronto carefully managed his workload as he worked his way back from an injured quad. The Raptors were admittedly excellent in those games—17-5 with a plus-11.8 point differential, versus a plus-4.0 differential in games he played—but would obviously fare better in the postseason with their star playing more minutes every game. Kyle Lowry, an All-Star this season, missed 17 regular-season games of his own.

The Raptors also added Marc Gasol at the trade deadline to boost their best lineup’s ceiling, while Golden State’s only in-season acquisition was Andrew Bogut, fresh off winning Australian league MVP honors for the Sydney Kings. Historically, championship teams haven’t needed midseason trades to bolster their roster, but Gasol proved an upgrade over Jonas Valanciunas, even before the playoffs began (plus-15.3 net rating with him on the court).

The Raptors Have Home-Court Advantage

It might be easy to mentally discount this factor because Golden State is such an accomplished road team. After their Game 2 victory, the Warriors have now won at least one game away from home in all 23 playoff series since 2013, and they were the league’s best road team in the regular season.

Best Road Teams, 2018-19 Regular Season

Team Record Net Rating
Team Record Net Rating
Warriors 27-14 6.4
Bucks 27-14 5.7
Raptors 26-15 4.5
Rockets 22-19 2.7
Jazz 21-20 2.5
Thunder 22-19 2.1
Celtics 21-20 2.1
Trail Blazers 21-20 0.1

But because Toronto beat Golden State twice in the regular season, thus besting the Warriors by a single game overall, the Raptors will host a potential Game 7—and that fact does matter. Mathematically, two teams as evenly matched as Toronto and Golden State would project to reach a seventh game about one-third of the time. And since the introduction of the shot clock in 1954-55, home teams have won 79 percent of Game 7s regardless of the round; for the Finals specifically, that rate is a healthy 75 percent (12-4). Through this lens, Golden State’s loss to Cleveland at home in 2016 was an anomaly; before that, every Finals Game 7 since the 1970s had gone the home team’s way.

If the series goes seven, Toronto would benefit. A Log5 calculation that factors in home/road point differential splits shows that Toronto would have a 55 percent chance to win a home game against Golden State, versus just a 43 percent chance to win a road game. That difference could be enough to swing the series.

Toronto’s Run to the Finals Was As Impressive As Golden State’s

After Kevin Durant left Game 5 of Golden State’s second-round series against Houston with an injured calf, the Warriors blitzed through the rest of their Western Conference slate. They won that game against Houston and finished off the Rockets on the road in Game 6; they steamrolled Portland in a sweep while completing several double-digit comebacks along the way. Meanwhile, Toronto needed a miraculous Leonard buzzer-beater to beat Philadelphia in seven games and then fell behind 2-0 against Milwaukee in the conference finals before switching the momentum.

But that description exaggerates the disparity between the two finalists’ paths. The Warriors struggled in the first round against an overmatched Clippers squad, remember, and while Houston proved as competitive as expected, Portland profiled as a relatively weak conference finals opponent. In the other conference, Milwaukee looked like a dominant force before spiraling against Toronto: The Bucks won 60 games in the regular season (with the point differential of a 64-win team), routed the Pistons in historically lopsided fashion in the first round, and easily handled a talented Celtics roster in the second. Beating Portland in a sweep may seem impressive, but statistically, it isn’t much more of a feat than beating Milwaukee at all. Toronto deserves ample credit—and computer models gave them that due—for navigating its tricky trail to the Finals.

To reach the championship round, Golden State went 12-4 against the Clippers, Rockets, and Trail Blazers; Toronto went 12-6 against the Magic, 76ers, and Bucks. The public might have leaned Golden State’s way because of the dominance that team showed in the back half of its games against the West, but Toronto’s struggles against high-quality competition should have been viewed as a fantastic achievement rather than a demerit.

Kevin Durant Is Hurt

Oh yeah, the two-time-defending Finals MVP was questionable to play before the series began, and has in fact missed the first two games. That makes a difference, because for all the pre-series chatter about Golden State’s more entertaining playing style without Durant, the team is certainly worse without him.

Some models don’t account for injuries and midseason transactions and instead just look at past team success, but for those that do, Durant’s uncertain health proved the final factor that balanced the two teams’ odds. FiveThirtyEight’s Nate Silver wrote that Golden State would have been roughly a 2-to-1 favorite by FiveThirtyEight’s calculation if both teams were fully healthy—a ratio more in line with the public perception.

FiveThirtyEight’s model uses dynamic health estimates, which means it accounted for the fact that Durant would be more likely to play games later in the series. That reason helps explain why that site thinks the Warriors are now 62 percent favorites to win the Finals, although Klay Thompson’s injury may further change the calculus. But 62 percent is still closer to a toss-up than a sure thing; it means the Raptors have a 38 percent chance, or the same chance that Leonard has of making any given 3-pointer.

The Raptors definitely still have an opportunity, even after letting Game 2 slip away. If anything, the first two games demonstrated that they have a better opportunity than many people thought. If only we had listened to the computers, or sought to see what they saw.