Title: How the Raptors shifted individual projections for the 2018-19 season
Date: July 5, 2018
Original Source: The Athletic
Synopsis: In my latest for The Athletic Toronto, I utilized a popular projection system to determine how much 2017-18’s over-performance changed the future outlook for the Raptors’ young players.
In the lulls of the free agent moratorium, I got playing with some data. Basically, what I wanted to know was this: Did the Toronto Raptors outperforming expectations last year improve their projections moving forward?
It seems simple enough. At the individual level, most of the Raptors’ young players delivered on their promise, cruising past what was a reasonable (or 50th percentile, if we had PECOTA-like projections) expectation. Even with some regression to our Bayesian priors about their career outlooks, their prior performances, and the league replacement level to account for smallish one-year samples, it seems obvious that over-performing as a youth would beget a rosier future projection, at least at the individual level.
Some of this is cloudy amid a coaching change and with some potential roster flux up in the air. It’s also difficult to use performance within a role to project how a player may do in a bigger role (like, say, if Fred VanVleet were suddenly a 35-minute starter). Even season samples are noisy, and basketball doesn’t yet have the luxury of a strong predictive catch-all metric the way baseball, a far more finite, discrete, and measurable sport, has. And even if we could accurately project individual performance, how those performances work together within the context of a team would never be as simple as adding up each player’s contributions (though some individual-performance based models do a fairly good job of projecting team performance). These are fine limitations, as this isn’t meant to be perfect. I just wanted to know if we can safely bump up expectations for the Gang of Youths on the roster heading into 2018-19.