Claude for Football Tactics Intelligence

Distinguish Claude’s tactical reasoning responsibilities from the deterministic responsibilities of the 2D pitch interface.

Define the target users, match contexts, and decisions the football intelligence Artifact will support.

Convert the eight desired capabilities into a prioritised prototype scope and an experimental extension scope.

Write user stories for creating, analysing, comparing, animating, explaining, and exporting tactical situations.

Write observable acceptance criteria for each core prototype capability.

Identify the practical limitations of a self-contained Claude Artifact and design fallbacks for unsupported behaviour.

Specify a normalised coordinate system for locating objects consistently on a football pitch.

Define a scene model containing teams, players, the ball, zones, labels, and directional annotations.

Represent formations, player roles, and in-possession or out-of-possession team shapes in a consistent format.

Represent tactical actions such as passes, carries, runs, pressure, cover, and rotations in a consistent format.

Represent multi-step tactical patterns as ordered frames with timing and transition information.

Create a controlled vocabulary for phases of play, spatial relationships, and tactical principles.

Define rules for handling ambiguous, incomplete, or contradictory tactical instructions.

Write a master build prompt that communicates the product goal, interaction model, visual system, and technical constraints to Claude.

Break a large Artifact request into staged build prompts with one testable change per stage.

Use plain-language descriptions of interface components, application state, and user events when directing Claude.

Create checkpoint prompts that preserve working features while Claude adds new ones.

Report visual defects and runtime errors to Claude with reproducible evidence and expected behaviour.

Use rollback and comparison prompts when a Claude-generated revision breaks an existing feature.

Prompt Claude to generate a proportionally accurate and responsive 2D football pitch.

Prompt Claude to add draggable players and a draggable ball with valid movement boundaries.

Prompt Claude to add editable teams, player identities, colours, roles, and formation presets.

Prompt Claude to add selection, duplication, deletion, reset, and undo interactions for pitch objects.

Specify visual layers for zones, arrows, passing lines, movement paths, pressure, and annotations.

Apply a clear visual hierarchy that keeps dense tactical diagrams readable.

Test pitch interactions across different canvas sizes and correct failures through targeted prompts.

Write a tactical instruction template that separates context, team shape, player actions, constraints, and desired outcome.

Direct Claude to translate a tactical instruction into the structured pitch-scene format.

Visualise Claude’s interpretation using positions, zones, arrows, labels, and movement paths.

Design clarification behaviour for instructions that do not provide enough tactical or spatial detail.

Constrain generated diagrams so player locations and actions remain physically and tactically plausible.

Test text-to-diagram behaviour with build-up, pressing, transition, and final-third scenarios.

Create a structured handoff that allows Claude to interpret the current manually arranged pitch state.

Prompt Claude to separate visible observations from tactical inferences about a pitch situation.

Prompt Claude to identify numerical, positional, qualitative, and dynamic advantages from the arranged scene.

Prompt Claude to diagnose space, compactness, pressure, cover, balance, and likely progression routes.

Require Claude to link every tactical claim to visible evidence on the canvas.

Display analysis findings as coordinated pitch annotations and concise natural-language explanations.

Define keyframes for a tactical sequence using player positions, ball position, actions, and duration.

Prompt Claude to add play, pause, restart, speed, and step-through controls for tactical animations.

Use movement trails and timed annotations to make rotations, passes, runs, and pressing actions legible.

Create a side-by-side comparison of two formations or tactical approaches under the same match conditions.

Compare tactical approaches using consistent criteria such as access, width, depth, rest defence, and pressing security.

Represent changes between attacking, defensive, and transitional shapes within one sequence.

Create an analysis framework that moves from evidence to diagnosis, options, recommendation, and trade-offs.

Specify team principles and match constraints that Claude must respect when recommending tactical solutions.

Prompt Claude to generate and rank multiple tactical responses to the same match problem.

Require each recommendation to identify intended benefits, dependencies, risks, and opponent countermeasures.

Prompt Claude to express uncertainty when the canvas lacks information required for a confident conclusion.

Create counterfactual comparisons that show how one tactical change alters space, roles, and likely outcomes.

Test recommendations against deliberately misleading or tactically impossible user requests.

Prompt Claude to add local saving, naming, reopening, and deletion of tactical projects.

Prompt Claude to add import and export of structured tactical-scene data.

Prompt Claude to export a readable pitch diagram for use in presentations and reports.

Design a screenshot-to-scene workflow that extracts teams, player locations, the ball, and visible annotations.

Create a review step for correcting uncertain or inaccurate information extracted from a screenshot.

Design an experimental video workflow that reconstructs a tactical sequence from selected key frames rather than full-match tracking.

Evaluate media reconstruction by comparing extracted positions and actions with a manually produced reference.

Create a test suite covering tactical correctness, interaction reliability, visual clarity, explanation quality, and data persistence.