(01) THE UTILITY OF THE NBA DRAFT COMBINE IN PREDICTING PLAYER SUCCESS

Abstract Details:

Background: NBA Drafts have prioritized size and athleticism; the NBA Draft Combine (NBADC) is an assessment of these traits. Advanced performance metrics (APM) such as Player Efficiency Rating (PER) and Win Share/48 (WS/48) are often used to measure an athlete’s performance in basketball.

Purpose: The purpose of this investigation was to assess relationships between NBADC metrics and success in the NBA.

Methods: Complete NBADC data sets were downloaded from NBA.com from 2000-2022 yielding a sample of 1287 players. Players were divided by position and level of accomplishment in the NBA: L0=Never played in the NBA (n=399); L1=Non-Qualifier < 58 games (n=369); L2=Rotation Player (n=456); and L3=All-Star (n=63). A One-way ANOVA was conducted to see if there were significant differences between the 4 groups. A Tukey post-hoc analysis was then performed to show where the differences lay between the groups. The Spearman’s rank correlation assessed relationships and a stepwise linear regression determined predictive relationships between NBADC metrics and APM. Statistical significance was set a priori at p≤0.05.

Results: PG. ANOVA, Wingspan (WS)(p=0.023), Wingspan-Height (W-H)(0.033), ¾ Court Sprint (¾)(0.012), Vertical Jump (VJ)(< 0.001). Tukey analysis, L1 players are 6 cm ↔ in WS than L0 players & 2.31 cm ↔ in W-H. L2 are .06 sec faster in the ¾ than L0. L1, L2, L3 all outperformed the L0 in the Raw Max VJ (RMVJ) with L3 jumping 10.49 cm ↑ than the L0. SG. Height (0.009), Weight (0.006), Lane Agility Drill (LAD)(< .001) VJ (< .001). L2 are 1.34cm ↑ than L1. L3 are 5.24 kg heavier than L1. L2 are 0.28 sec faster than L1 in the LAD. L3 can jump 7.48 cm ↑ than L0 in the RMVJ. SF. LAD (0.002), VJ (0.007). L2 are .24 sec faster than L0 in the LAD. L1 can jump 4.29 cm ↑ than L0 in Max VJ and 3.36 cm ↑ in the RMVJ. PF. Standing Reach (SR)(< .001), LAD (< .025), VJ (< .001). L3 are 3.9 cm ↑ in SR. L2 are .21 sec faster than L0 in the LAD. L3 can jump 7.38 cm ↑ than L0 in the RMVJ and 7.58 cm in the Raw No Step VJ (RNSVJ). C. WS (< .001), LAD (0.041), VJ (0.006). L3 are 7.76 cm ↔ in WS than L0 and .33 sec faster in the LAD. L1 can jump 5.57 cm ↑ than L0 in the RNSVJ and 6.25 cm ↑ in the RMVJ. A statistically significant relationship (SSR) exists between PER and W-H (t=2.122, p=.038), which explained 6.9% of the variability (V) in PER. Each 1 unit ↑ in PER is associated (assoc) with an ↑ of .136 units in W-H. A SSR exists between WS/48 and WS (t=2.177, p=.033) accounting for 7.2% of the V in WS/48. A 1 unit ↑ in WS/48 is assoc with an ↑ of .001 units in WS. A SSR exists between MPG and Max VJ (t=2.363, p=.021) explaining 8.4% of V in MPG. Each 1 unit ↑ in MPG is assoc with an ↑ of .090 units in Max VJ.

Conclusions: Raw VJ metrics illustrated associations across all positions. Anthropometrics of WS and W-H and physical performance in the vertical jump is predictive of performance in All-Star populations. Significant differences in performance and anthropometrics were observed between different levels of NBA players.PRACTICAL APPLICATION: While there are a myriad of factors which are likely to contribute to basketball ability in the NBA our data suggests that lower body muscular power is associated. As such, coaches and other individuals wishing to project future contributions may wish to prioritize vertical jump in comparison to metrics of physical performance in the NBADC. Some justification exists to prioritize WS and W-H when evaluating future NBA players.

Acknowledgements: None