Forms Academy · FORMAX Research
A Research Hub · May 2026

The Constructed Chain

A developmental theory of tactical possibility in football, presented through papers and interactive Markov experiments.

The existing Markov chain literature in football is descriptive. It models the chain that emerges from match data and assigns value to its component actions. This program of work moves the question upstream. The chain a team can execute is a downstream product of the technical substrate of its players. Random play is the chain you collapse to when the substrate is absent. Structured football is the chain that becomes available when the substrate has been built. FORMAX is the substrate, organized as a five-domain by five-layer matrix, and the experiments below make the relationship between substrate and chain concrete.

Ron Hogsett · Founder, Forms Academy

Interactive Experiments

Experiment 00 · Foundational · R = D / T engine

The Constructed Chain Simulator

A Markov chain whose transition matrix is generated, not chosen. Adjust the five FORMAX domain sliders and watch the matrix regenerate through R = D / T applied to every forward transition. Each transition shows its task difficulty, the team's technical capacity, and the resulting perceived risk in real time.

Claim: The chain you can execute is determined by where your players sit on the FORMAX matrix, through the equation R = D / T. Move a slider; the chain follows.
Experiment 01 · Cross-Level Comparison

Process Entropy by Playing Level

Five chains modeled across recreational, club youth, semi-professional, elite professional, and world-class levels. Mean row entropy and expected goals computed for each. The data behind the gap between camps.

Claim: Process entropy decreases monotonically with playing level. The "football is random" thesis is largely a selection effect from high-entropy populations.
Experiment 02 · Substrate vs. Ambition

The Cost of Imported Philosophy

Two sliders: substrate quality and tactical ambition. When ambition exceeds substrate, transitions fail and turnovers compound. When the two are aligned, expected goals maximize. The heatmap reveals the developmental constraint.

Claim: You cannot prescribe a chain your players cannot execute. The cost of imported philosophy is structural, not stylistic.
Experiment 03 · Distribution Analysis

The Variance Gap

Ten thousand simulations of structured and direct chains. Outcome histograms placed side by side. The structured chain wins on the mean and on the variance simultaneously, which is the statistical case that the public discourse never makes.

Claim: Low-entropy chains produce both higher mean outcomes and lower variance. Structure is not a high-risk choice; it is the opposite.

The Papers

Paper I

The Constructed Chain

A Developmental Theory of Tactical Possibility in Football

Argues from the analytics side inward. Surveys the existing Markov literature (Rudd 2011, Singh 2018, Cervone et al. 2014/2016, and onward) and locates the contribution. The chain a team can run is constructed by the substrate; the substrate is built by the methodology.

Process entropy Markov chains Substrate theory
Paper II

FORMAX

A Five-by-Five Substrate for Football Development

Argues from the developmental side outward. Describes the matrix in detail (five domains by five layers), the assessment principles that operationalize it, and the curricular implications for academy programs. The companion to Paper I.

Player development Technical substrate Curriculum design

About this research program

FORMAX Research is the long-term intellectual program of Forms Academy and AlphaForms. Its purpose is to make the substrate-side argument for football development explicit, formal, and testable. The simulator and the experiments are working tools, intended for use by coaches, technical directors, federation curriculum designers, and analytics departments who want to engage the substrate question alongside the chain question.

The agenda from here is threefold: deeper assessment instruments for each FORMAX domain, longitudinal yield studies of FORMAX academy programs against comparators, and chain-construction studies that connect specific tactical models to their FORMAX prerequisites. Collaborators welcome.

Standing on shoulders

Sarah Rudd · 2011
A Framework for Tactical Analysis Using Markov Chains
The foundational application of absorbing Markov chains to soccer. NESSIS. The intellectual progenitor.
Karun Singh · 2018
Introducing Expected Threat (xT)
The piece that gave xT its name and broke through to the broader analytics community. Now at Arsenal.
Cervone, D'Amour, Bornn, Goldsberry · 2014/2016
A Multiresolution Stochastic Process Model for Predicting Basketball Possession Outcomes
Expected Possession Value (EPV). The technical apex of possession valuation, originally for basketball and extended to soccer by Fernandez, Bornn, and Cervone (2019).
Hirotsu, Wright · 2002, 2003
A Markov Process Model of an Association Football Match
Full-match Markov framework used to optimize substitution timing and tactical decisions.
Decroos, Bransen, Van Haaren, Davis · 2019
Actions Speak Louder Than Goals (VAEP)
A machine-learning alternative to xT that handles defensive actions and full match context. KDD 2019.
Van Roy, Robberechts, Yang et al. · 2021
Leaving Goals on the Pitch (MDP for Soccer)
Markov Decision Process framework asking not just what value was created, but whether a better action was available.