Every sports fan, diehard or casual, has watched Moneyball,
the movie about the use of statistics in baseball. While sports has become
more receptive to the use of statistics to identify players, many fans still do
not like to use or misuse statistics to back up their opinions. As an avid NBA
fan, I too love to concoct fictitious trades to help make my team better. Through
the use of statistics, I am going to try to make well informed decisions
regarding player acquisitions for the San Antonio Spurs, my favourite NBA team.
To tackle this problem, I used a linear regression model.
To create the model, I first collected box score data for the
Spurs’ 2019-20 season. This data was then used to create a model that will give
a composite score, which predicts a team’s record. According to the model, a
score closer to 1 indicates a better record, while a score closer to 0
indicates a worse record. Using Basketball Reference, I identified 8 players
who the Spurs could feasibly acquire and who would improve their record. Then,
I extrapolated all of the thirteen players’ stats to 48-minute stats, this is
useful because it focuses on if the main 5 played all 48 minutes of a game,
which delivers a better estimate of the team’s quality. As it is almost
impossible to completely change a team’s roster within one season, I limited
the number of new additions to either 1 or 2 players.
The 13 players involved in this are the San Antonio Spurs main
5 (Dejounte Murray, Bryn Forbes, DeMar DeRozan, Trey Lyles and LaMarcus
Aldridge) and 8 other players who I thought could join the Spurs. The players
are Malik Monk, E’Twaun Moore, Otto
Porter Jr., Will Barton, Aaron Gordon, Dario Saric, Mo Bamba, and Montrezl
Harrell.
When the model was run, the top team
consisted of Murray, DeRozan, Aldridge, Gordon, and Barton. The aforementioned
team achieved a score of 0.802, while the Spurs current main 5 achieved a score
of 0.579. DeRozan and Murray were present in all of the top 10 results, and
Aldridge was in 7 out of the 10. Will Barton was in five of the top 10 results,
and Aaron Gordon was in 7 of the top 10. Mo Bamba was in 4 of the top 10,
though this can be attributed to his amazing blocks numbers in limited playing
time. Most of the best results substituted Bryn Forbes and Trey Lyles for Gordon
or Barton. These are the 10 best lineups, with their Win Co-Efficient.
On the other hand, the 10 least productive
lineups according to the model all featured Bryn Forbes, LaMarcus Aldridge and
Trey Lyles. These lineups mostly failed due to a lack of passing, and all the
lineups had two Lou Williams-type ‘Professional Scorers’ (Bryn Forbes and Malik
Monk/E’Twaun Moore) and three big men. Otto Porter Jr., who makes two
appearances in the 10 least productive lineups, also possesses sub-par passing
ability. In fact, none of these lineups had more than 18.5 assists. They are
listed below along with their Win Co-Efficient.
A quick study shows that
the Spurs should likely let Bryn Forbes play in his natural sixth man role,
where he can be best utilized as a gunner and 3-point specialist, and trade for
Aaron Gordon or Will Barton. This will allow the Spurs to strengthen positions
of need without gutting the rest of their roster. The data shows that the
Spurs, while unlikely to win a championship in the near future, can contend for
a spot in the Western Conference playoffs with a few savvy trades, and can keep
their record 22-year playoff streak alive.
In an upcoming article, I
will also be detailing the trades that the Spurs could make to acquire the
players I identified as reasonable targets.
Discussion about this post