NBA – Probability of Drafting Star / Above-Average Player by Draft Pick
The 2015 MIT Sloan Sports Analytics Conference recently wrapped up and we wanted to share work for our research paper -- that made it past the abstract round was considered as a finalist for the conference. Research was performed by Carlton Chin, Jesse Heussner, and Max Weisberg. This article is part one of a series of articles.
The value of professional sports draft picks is fairly well documented. No matter the metric chosen, we see the familiar exponential function, of slightly varying shape. Intuitively, we know that the performance of the best players and all-stars should be roughly the same across different eras. However, research shows that the impact of young stars has declined over time. Has the value of draft picks evolved over time?
Professional sports have become increasingly competitive – and this has changed the dynamics of professional sports drafts. That is, data shows that the value of draft picks, especially relative to the general level of play within professional leagues – is declining. The results have implications on various levels. For instance, one dynamic is the slower speed (and steeper learning curve) at which young stars develop in today’s more competitive sports environment. In addition, some analysts prefer to study value as a function of salary. Various metrics are investigated to mine for value in the increasingly competitive world of professional sports.
1. Introduction and the Value of Draft Picks
Valuations of draft picks may be performed based on a variety of parameters, objective functions, and time frames. Some analysts might compute values based on career wins (above replacement or above average), while others focus on peak value. Computations may be based on a relatively short time period or may be based on very long-term historical data. For example, the average number one pick is worth x, number two pick worth some lesser amount, and so on. No matter the method of calculation, we see the familiar exponential function, of slightly varying shape, depending on the choice of parameters.
Below, we use NBA draft data going back over several decades and derive the probability of drafting stars or above-average players. Analytical methods may be used to compute probabilities and present realistic expectations. However, the data shows that even a relatively simple task – as in valuing a draft pick – has many idiosyncrasies and can lead to studying multiple parameters including value, perceived value, salaries, competition levels, and more.
2. Advanced Analytics
Professional sports are in the midst of an analytics revolution. While baseball used to be the sport where advanced stats served the most application, analytics have become an integral part of the decision making process for most NBA franchises and have gradually entered the conversation in the NFL as well. Most MLB decision makers rely heavily on sabermetrics, and many NBA franchises are now investing in their own analytics departments to try to gain a comparative advantage over competing teams.
But public discourse is evolving at a similar rate. To many baseball fans, “classic” stats such as hits, homeruns, and batting average have already been replaced with more illuminating statistics such as On Base + Slugging (OPS) and Wins Above Replacement Player (WARP). The same trend is evident in basketball, where it is no longer as acceptable to simply look at points, rebounds and assists to gauge a player’s worth. Allen Iverson probably would not have won the 2001 MVP if we were armed with the information we currently enjoy, as statistics such as TS%, RPM, and WARP would have likely foiled much of Iverson’s case.
The Houston Rockets and Dallas Mavericks have teams of analytics people supplying information to coaches and players; the Sacramento Kings crowdsourced their draft with the hope of gaining a comparative advantage over other teams; the Houston Astros have created a “nerd cave” of analytics people to spur a rebuild . All the while, fans and media members are simultaneously going along with the ride. Fans and practitioners in the analytics community all have the same goal: to most accurately measure a player’s individual value.
Carlton Chin is a fund manager, MIT-trained quantitative analyst, and co-author of “Who Will Win the Big Game?" He has been quoted by the Wall St. Journal, New York Times, and ESPN. Jesse Heussner and Max Weisberg co-authored this paper and worked with Carlton for the Sacramento Kings on Draft 3.0.