A machine-learning-driven approach to mapping team tactical identities and identifying market inefficiencies across North American leagues.
Phase 1 PCA-K-Means clustering identifies tactical archetypes across American soccer leagues based on a 15-metric profile.
Identified by elite Possession Quality. These teams dominate games through patient, high-completion passing sequences in the middle and final thirds.
Identified by extreme Opponent Directness. They restrict short build-ups, forcing direct clearances by deploying aggressive high presses.
Identified by intensive Defensive Disruption. Midfield defensive units that break up transition plays and force high turnover volumes.
Identified by elevated Receiving Impact. High-mobility teams that excel at finding playmakers in space between the opposition lines.
Identified by massive Possession Volume. These systems dictate the match tempo by chaining long, secure passing sequences to choke out opposition transitions.
Identified by high Attacking Directness. High-tempo setups that bypass midfield build-up to quickly hit spaces in behind defenses.
How we quantify tactical DNA using Principal Component Analysis.
We extract a robust 15-dimensional matrix for every team from American Soccer Analysis, including Base xG, Passing Volume & Directness, and specialized g+ components.
Rather than forcing teams into preconceived boxes, we use Principal Component Analysis (PCA) to mathematically identify the strongest vectors of variance across playing styles.
Teams are grouped via K-Means clustering. An automated labeling engine then scans the centroids to assign sharp, human-readable tactical identities based on the data.
Stop overpaying for raw output. TVMS enables clubs and analysts to evaluate players based on how their specific statistical archetype fits a target team's tactical DNA. By measuring system-interaction, we project performance translation across leagues and uncover hidden market value before the competition does.
Clustering team identities in USLC and MLS.
Classifying player roles using PCA reduction.
Predicting Fair Market Value based on system fit.
I'm always building new tools and analyzing the modern game. Follow along on X (Twitter) for the latest insights and updates to TVMS.
Connect with me