The Toronto Raptors have never seen a team quite like the Golden State Warriors in the playoffs. It’s not that they haven’t had strong competition in the past: the Philadelphia 76ers are loaded with talent, and the Milwaukee Bucks had the best regular season record in the NBA this year, as well as the probable league MVP in Giannis Antetokounmpo.
The Warriors, however, are a different beast altogether. They’ve won three of the past four NBA championships, have a full starting lineup’s worth of All-Star-calibre players, and two of the best players in the NBA. Yet the Raptors — a team which has never made an NBA Finals — are being given a chance to win them, and were even favoured at the beginning of the series.
How can this Canadian expansion project, with a history of underwhelming playoff performances, be favoured to defeat the greatest dynasty that the league has ever seen? Is it time to question the validity of these predictive models? Or is it possible that the Raptors are actually a good team?
Roster Construction
Raptors President Masai Ujiri and General Manager Bobby Webster were diligent in addressing their shortcomings this offseason. Last July, Ujiri made the tough decision to trade fan favourite DeMar Derozan, along with Jakob Poeltl, to the San Antonio Spurs in exchange for Kawhi Leonard and Danny Green.
Both Green and Leonard have played in two NBA Finals, and won an NBA championship together, with Leonard winning Finals MVP. Ujiri and Webster not only upgraded the team’s talent, but its experience and championship pedigree as well. The Raptors front office also reportedly refused to include forward Pascal Siakam in the trade for Leonard and Green.
This seemed like an odd decision, given that Siakam had only averaged 20.7 minutes and 7.3 points per game in the previous season, but he has since blossomed into a starter and the Raptors’ secondary scoring option after Leonard. At the trade deadline, the Raptors acquired veteran centre Marc Gasol, who brought more playoff experience, as well as elite defence and passing.
By the numbers
The American data outlet FiveThirtyEight gave the Raptors a 54 per cent chance to win the NBA Finals at the start of the series. That number dropped down to 49 per cent after game two but has since increased to 87 per cent. FiveThirtyEight uses a projection model called CARMELO, which takes into account numerous factors.
CARMELO is a more advanced version of Elo ratings, which factor in which team won each game, the margin of victory, and where each game was played. Elo is a somewhat useful statistic, but has numerous flaws in evaluating future performance. For example, if a player is injured, or is resting — as was the case with Leonard throughout the regular season — Elo would not be able to account for that.
Adding Leonard, Gasol, and Green to the roster has quite clearly made the Raptors a better team, but this did not show up in their Elo score: the Raptors highest ever Elo rating came in March 2018, before acquiring any of these players. CARMELO incorporates individual player projections to account for offseason transactions, injuries, and rest. FiveThirtyEight also later added a playoff experience adjustment to account for the advantage that more experienced teams have.
A U of T model
U of T statistics professor Jeffrey Rosenthal has been working on his own model for the NBA playoffs, although he admits that it is much less advanced than other models like FiveThirtyEight’s. He created this model in response to media inquiries asking him to calculate the probabilities of Kawhi Leonard’s buzzer beating shot in game seven against Philadelphia.
“That was an interesting one because I couldn’t find actual statistics about how often the shot… bounces four times. It’s extremely unlikely,” Rosenthal recalls.
Rosenthal’s predictive model is heavily based on past performance. “I just looked at the regular season records and outcomes of the competing teams, and compared their performances at home and away, and extrapolated from that into the playoffs to give an estimate for each game of the probability that one team would win or the other, taking account of home court advantage and that kind of thing. And then do that to get an estimate of the probability for each game,” he explained.
He found that there was a huge difference in how certain teams performed at home versus on the road, noting that Toronto and Milwaukee played much better at home, whereas Golden State was about even in both settings. The Raptors had a better regular season record, giving them home court advantage in four of the seven games this series. Rosenthal gave the Raptors a 51 per cent chance at the start of the series, dropping down to 48 per cent after game two, and now at a high of 89 per cent going into game five.
Many are also predicting that the tides will turn even more in Golden State’s favour when star forward Kevin Durant returns from an injury, but Rosenthal isn’t so sure. “You can say, ‘He’s a great player, and coming back, it’s going to make all the difference.’ Or you could say, ‘They’ll have a new guy back in the lineup, he’s missed a few games, he’s out of rhythm and he’s still hurting,’ or whatever. So, it’s hard to say.”
With only one more win needed, Raptors fans are counting on the team to overcome the odds and bring home its first championship title. “It’s the cliché, but there’s a reason you have to play the game, right? You can [only] get so much by trying to predict,” Rosenthal notes.
