Brief 02 / Team coordination

Coordination under pressure

Read a team as a sequence of connected responsibilities. The important question is not simply where everyone stood, but which actions made the next action possible.

Method brief / No player rankings

A cue moves through a teamContact, relay and response form a gold path, with cream support nodes connected underneath. This is a conceptual sequence, not recorded player movement.CONTACTRELAYRESPONSECOVERAGERESERVE

Team coordination is often described with broad words such as chemistry, discipline or cohesion. Those terms may capture an impression, but they do not identify a reviewable mechanism. A better starting point is an event chain: someone encounters a cue, another player responds, and other actions protect or enable that response. AI can help locate repeated chains. It cannot establish a shared plan merely because several players moved at similar times.

Start with the shared problem

Define what the team was trying to resolve. Was it entering a contested area, protecting an objective, recovering information or disengaging after a resource loss? The same spacing can be appropriate for one problem and dangerous for another. A review that begins with a generic ideal formation risks treating deliberate adaptation as failure. Write the objective and the relevant constraints before measuring the players' positions.

In Dota 2, information denial, teleport availability and hero roles may shape the response. In Rocket League, boost availability, rotations and recovery positions create a different support structure. Counter-Strike 2 trades depend on visibility, route access and timing. These games reward coordinated action in different ways; a metric should not travel between them without a new explanation of what it measures.

Follow cue, response and coverage

Mark the earliest observable cue and the next action that plausibly responds to it. Then inspect the supporting actions around that pair. Did another player hold a route, reserve a resource or create an escape? A player who never appears in the decisive engagement may have made it possible. This is why reviewing only damage or eliminations produces an incomplete account of contribution.

A

Cue

What changed, and who could observe it?

B

Response

Which action followed within a relevant window?

C

Coverage

What protected the response or preserved an alternative?

A response window needs a game-specific definition. Distance alone may not describe support if terrain, cooldowns or an obstructed route prevent interaction. Likewise, a short delay may be necessary to align resources. Explain why the selected window matters instead of treating fast responses as universally better.

Use communication evidence carefully

Replay movement is not a transcript. Without voice recordings, it is usually impossible to distinguish an unheard call from a missing call, an intentional delay or conflicting information. Label communication-related explanations as hypotheses. If authorised recordings exist, preserve timing and speaker context, and obtain appropriate permission before using them in a wider review.

Do not upload private team communication to an external model simply because transcription is convenient. Check access controls, retention terms and team policy first. Sensitive tactical information and personal discussion may coexist in the same recording. Minimise what is retained and keep player identity out of examples when identity is not necessary to understand the sequence.

Look for repeated handoff failures

AI-assisted event alignment can help find situations where a cue is followed by a response without adequate coverage. Inspect the source replay for each candidate. A detector may mistake a planned disengagement for hesitation, or miss support delivered through pressure elsewhere. Include counterexamples where the same visible pattern worked as intended. These cases often expose an omitted variable.

ObservationQuestion to investigate
Players arrive at different timesWere they meant to arrive together?
A trade does not occurWas the second player able to see and reach the interaction?
Several resources overlapWas redundancy planned, necessary or avoidable?
A player remains distantWhat responsibility was that position covering?

Make the lesson collective

Finish with a responsibility, a cue and a confirmation rule. For example, a team might agree to confirm supporting resource availability before beginning a particular entry. The review should specify who recognises the cue, who responds and what cancels the action. This creates a practice question without attributing the entire sequence to one player's character or awareness.

Revisit the agreement in later sessions under comparable conditions. If it fails, inspect whether the rule was unclear, the cue was unavailable or the environment changed. A coordination model is useful when it sharpens that discussion. It becomes misleading when its score replaces the discussion, or when uncertainty is translated into blame.

A small team reviewing notes and discussing a shared screen around a table
Team discussion provides context that positional data alone cannot recover. Supporting photography is not a record of the conceptual sequences above.