Research

FSL begins with a narrow question:

Can different ways of failing be distinguished more precisely than existing football language allows?

Its first question is therefore:

Does using the language help people see, describe or communicate how a team is failing differently?

A useful distinction may remain qualitative. It may become part of football language without ever becoming a metric.

But some distinctions may invite further investigation:

  • Can they be observed reliably?
  • Can different people recognise them?
  • Can they be operationalised?
  • Can they be measured?
  • Do they add anything beyond existing football concepts or data?

This page sets out potential ways of finding out.

The sequence

Recognition → Distinction → Reliability → Measurement → Application

Recognition. People can recognise the phenomenon.

Distinction. People can distinguish one form of failure from another.

Reliability. Independent observers apply the distinctions consistently.

Measurement. The distinctions correspond to observable features that can be operationalised and, where appropriate, measured.

Application. The distinctions change analysis, communication or practice.

FSL does not begin by assuming that a pattern is measurable. It begins by asking whether the distinction is useful. It is not claiming that Drift is a measurable state.

A rule the research depends on

The scoreline must not define the pattern.

A useful Failure Pattern should be capable of occurring without the corresponding outcome, and an outcome should be capable of occurring without the corresponding pattern.

What makes two instances the same pattern?

A useful category needs more than a recognisable name. It needs enough internal coherence that two observers can identify what different instances have in common. Similar-looking passages are not necessarily the same phenomenon.

FSL therefore needs to distinguish between:

  • different expressions of the same underlying condition
  • superficially similar passages produced by different conditions
  • genuinely different Failure Patterns

This is part of the conceptual and reliability work, not something assumed in advance.

The categories themselves are hypotheses

The research may show that a proposed pattern is:

  • a distinct condition
  • a subtype of another pattern
  • a transition rather than a condition
  • a response to another pattern
  • an expression of several underlying dimensions
  • or not a useful distinction at all

Four kinds of question

  • Conceptual — Does the vocabulary make coherent distinctions? Do categories overlap unnecessarily? Does a proposed distinction add something to existing football language?
  • Perceptual — Does using FSL change what people notice or how specifically they describe it? Can observers distinguish patterns more clearly? Can those distinctions be communicated?
  • Methodological — Can different observers apply the definitions consistently? Can the categories be coded without relying on scoreline, reputation or hindsight?
  • Empirical — Do the proposed patterns relate to observable features of football? Do they add information beyond existing concepts or measures?

Related work

FSL is not starting from nothing, and this list is not exhaustive.

Momentum. Recent work has used expert coaches’ video-based descriptions of teams’ collective organisation during positive and negative momentum phases. It found negative phases associated with a stretched, deeper block and with individual actions disconnected from it. A follow-up suggests some patterns may precede transitions between phases. Momentum there is defined through score changes; FSL asks whether failure can be described independently of the scoreline, and whether it comes in distinguishable kinds.

Stability and instability. Dynamical-systems work in football has treated play as moving between stable and unstable states, and studies of perturbation have used expert judgement. These tend to use a stable/unstable distinction and are mostly about creating advantage, not about types of failure.

Coordination in tracking data. Studies of defensive synchrony suggest breakdowns in coordination coincide with opponents’ scoring opportunities. These may offer candidate measures if a distinction survives.

A fuller review of team cognition, shared mental models and coaching literature is still needed. If existing concepts describe the same phenomena equally well, that is a finding.

Phase 1: Practitioner Exposure

Can practitioners distinguish different kinds of failure?

Before testing reliability or formal category recognition, FSL explores how experienced coaches, analysts and observers encounter the vocabulary.

The purpose is not to establish whether the framework is correct. It is to investigate whether it changes what people notice, how they describe a passage, and which distinctions they find useful.

This phase may include:

  • guided conversations with coaches and analysts
  • comparative descriptions before and after exposure to FSL
  • discussion of terms that clarify, obscure, or fail to fit
  • identification of distinctions practitioners already make but existing vocabulary does not adequately express

The findings inform later perceptual and reliability work, including the refinement, removal, or addition of terms.

A first, informal look: Quick Match Read

This was an informal pilot, not a preregistered study.

Ten respondents, including coaches recruited through WhatsApp and LinkedIn, watched a passage without prior FSL context or definitions. They first described what they saw in free text, then selected from the seven proposed Behavioural Expressions plus None.

The exercise was not blinded, did not use trained or calibrated coders, and does not meet the reliability standard proposed elsewhere in this programme.

The free-text descriptions showed some consistency around:

  • calm, comfortable possession
  • time and space
  • limited pressure
  • difficulty finding a route forward
  • circulation without penetration or progress

The forced-choice responses were:

  • Connect: 4
  • Settle: 3
  • None: 3

Two respondents explicitly said the closest available option did not quite fit. Three rejected the list altogether. Their descriptions suggested possession without threat, patience without a clear route forward, or difficulty turning circulation into progression.

Those descriptions might be consistent with Drift or Stall, but the pilot cannot establish that. The forced-choice list offered only Behavioural Expressions plus None. No Failure Pattern was available to select.

The finding is narrow. Some respondents described something they felt the available list did not adequately label. That makes it worth testing formally whether a passage can be described more precisely as a Failure Pattern, a Reference Condition, or neither.

The next step is not to repeat this exercise at larger scale. The perceptual programme below formalises the comparison: preregistering what counts as greater descriptive specificity or convergence, offering Failure Patterns as response options, introducing independent blinded assessment, and comparing FSL-trained observers with appropriate control conditions.

Phase 2: Perceptual Exploration

Can observers distinguish how a team is failing?

The most direct formal test is the simplest:

Does having the language change what people see?

A useful distinction should help an observer notice, describe or communicate something more precisely, not simply give them more football terminology.

A controlled study could compare people trained in FSL with people using an alternative structured football vocabulary, and with people using a single-dimension description of “momentum” or “team performance”.

Independent judges could then assess:

  • behavioural specificity
  • temporal precision
  • distinction between similar patterns
  • collective rather than purely individual description
  • ability to communicate what changed
  • convergence between independent observers

Observers are always asked for the evidence in the passage that led them to their reading. Otherwise the study tests vocabulary recognition rather than observation.

Recognition

Observers identify or describe unfamiliar passages. Training is tested for changes in recognition, specificity and consistency.

Discrimination

Can observers distinguish adjacent patterns that look superficially alike? Drift and Fog, Stall and Drift, Scramble and Overheat, Overheat and Collapse. This is one of the most important tests of whether six patterns are really six.

Transmission

One observer describes an unseen passage to another, who attempts to reconstruct it. This tests whether the vocabulary is communicable.

Convergence

Independent observers describe the same passage without consultation. Convergence tests shared distinctions, not just individual clarity.

Transfer

Observers apply the vocabulary in novel contexts. A useful language should not depend entirely on having seen the original examples.

Exit (exploratory)

Used only on passages where independent observers have already agreed which pattern is present; if they cannot agree on the pattern, the exit is not interpreted. Observers say, in their own words, what direction of change the passage appears to call for. Answers are later coded to a direction.

The expected mapping is preregistered:

  • Fog → Restore Clarity
  • Stall → Restore Agency
  • Overheat → Rebalance
  • Collapse → Reduce demand
  • Drift and Scramble → left open

If exits do not differ across patterns, the patterns add nothing at the intervention level. If observers agree on the pattern but choose widely different directions, the pattern may carry no practical meaning. If agreed patterns consistently produce different directions, that is evidence the vocabulary carries practical information.

Answers are coded at the level of direction first; tactical specifics are captured in free text.

Phase 3: Reliability and Comparison

Can independent observers reliably distinguish the proposed ways of failing, and how does that compare with alternatives?

Reliability matters especially here because the categories describe subtle behavioural conditions, not obvious events such as shots, turnovers or goals.

The current vocabulary is a working coding system, not a finished taxonomy.

Coding system

Failure Patterns are provisionally coded as single-primary categories.

Reference Conditions are non-exclusive; more than one may appear in the same passage.

Transitions are coded separately, recording:

  • starting pattern
  • destination
  • approximate timing
  • behavioural trigger or context
  • confidence
  • whether the change appears abrupt or gradual

Reliability requirements

  • use at least two coders where practical
  • define coding rules before analysis
  • keep coders independent
  • avoid outcome, reputation or tactical shortcuts
  • retain disagreements rather than resolving them silently
  • use adjudication only after independent coding
  • test adjacent categories where confusion is expected
  • report uncertainty rather than forcing false precision

Reliability is a prerequisite for interpreting later empirical tests. It is not evidence that a category is true.

Sample requirements

A useful sample should contain enough variation to expose strengths and weaknesses, and should deliberately seek counterexamples:

  • tournament and high-stakes defeats
  • late collapses
  • goals conceded after sustained pressure
  • visible periods of collective deterioration
  • teams that show a pattern without conceding
  • teams that concede without a preceding pattern
  • teams that recover from an apparent failure
  • teams that win while in a pattern
  • teams that lose while behaviourally stable
  • similar outcomes preceded by different patterns
  • similar patterns followed by different outcomes

Failure must be separable from outcome. Without these cases, “Failure Pattern” quietly becomes “bad result.”