Basketball Training Load Data Means Different Things for Different Metrics, Study Finds

The research, published in Frontiers in Sports and Active Living, analysed external load data from 84 professional male players across four Spanish first division teams and found that the most appropriate level of interpretation — whole team, positional group, or individual player — differs from one variable to the next.

The study drew on local positioning system and inertial measurement technologies collected over eight measurement windows spanning three competitive seasons. The dataset comprised 7,211 individual player drill observations from 608 unique session-drill units in training sessions. Rather than assuming a single best way to benchmark load, the researchers analysed fourteen variables separately, using linear mixed models with random intercepts for player and session-drill unit.

To decide which interpretation each variable supported, the team compared models with no positional classification against models with two, three, or five positional groups, using the Bayesian information criterion. Team measurement window was retained as a contextual adjustment when the data supported it. The result was a variable-specific framework: some metrics demanded team-level interpretation, others required positional splits, and one called for a finer three-way division.

Whole team interpretation was selected for eight variables: distance, explosive distance, accelerations, decelerations, impacts, jumps, high metabolic load distance, and energy expenditure. For these measures, the study found no evidence that splitting players by position added meaningful information beyond the drill context and the team as a whole.

Positional interpretation using two groups — perimeter and post players — was selected for five variables: peak speed, peak acceleration, peak deceleration, high-intensity take-offs, and high-intensity landings.

Player Load was the only variable for which a three-group positional interpretation — guards, forwards, and centers — was supported. Team measurement context was retained for six of the fourteen variables, as a contextual adjustment when supported by the data.

Beyond the group-level findings, the study also examined whether between-player variability supported complementary individual interpretation after accounting for drill context. It did, across all variables. In other words, even after the drill setting and the appropriate group level were taken into account, individual players still differed in ways that individual interpretation could capture.

The practical implication, according to the authors, is a layered approach to load monitoring. Drill context should provide the common basis for interpretation, positional interpretation should be applied selectively — only for the variables where the data support it — and individual interpretation should complement the supported group level.

This study suggests that decision is best made variable by variable rather than by adopting a single convention for all metrics.

The research was conducted exclusively in professional male basketball in the Spanish first division, and the authors note that the supported interpretation levels inform the selection of interpretation levels for external load monitoring in that setting.

Subject of Research: Interpretation levels for external training load monitoring in professional male basketball

Article Title: A variable-specific framework for interpreting external training load in professional male basketball

Article References: López-Sierra, P., Scanlan, A. T., Ferioli, D., & Ibáñez, S. J. (2026). A variable-specific framework for interpreting external training load in professional male basketball. Frontiers in Sports and Active Living, 8, Article 1952209. https://doi.org/10.3389/fspor.2026.1952209

Image Credits: AI Generated

DOI: 10.3389/fspor.2026.1952209

Keywords: basketball, training load, load monitoring, sports science, local positioning system, inertial measurement, positional analysis, drill demands, player load, professional basketball, mixed models, external load