Methodology · v1
Analysis methodology
This document specifies the scoring framework the Steam Capsule Analyzer applies to capsule artwork: the rubric, the viewing conditions, the scoring scale, and how the overall score is computed.
Scope
The rubric evaluates the visual capsule asset on six dimensions: genre clarity, title readability, contrast and color, uniqueness and polish, brand consistency, and composition.
Each dimension is scored as a 1-10 integer against published anchors, evaluated at three rendered sizes on Steam's #1b2838 background. The overall 0-100 score is the average of the six dimension scores, multiplied by ten.
The framework is versioned. Material revisions advance the version stamp on this page and are recorded with a dated description of the change.
Viewing conditions
Every dimension is evaluated at three rendered sizes, on the background color Steam actually uses, under attention conditions that match real browsing. A capsule that performs at full size but degrades at thumbnail size fails in production, where the large majority of impressions happen at the smaller renders.
| Test | What it means |
|---|---|
| Full size | 920 × 430, the header capsule on the store page. |
| Small | Around 231 × 87, the home-page rail render. The size where ornate fonts start to fail. |
| Tiny | 120 × 45, the auto-generated thumbnail in search and discovery. The render where most impressions happen. |
| Background | Steam dark, #1b2838. Every capsule is judged on this color, beside other capsules on this color. |
| Quick scroll | Slight blur, low attention, under one second to parse. The realistic browsing condition. |
| Squint test | If the design collapses when squinting, dimensions that depend on fine detail are scored down. |
| Grayscale test | Applied for contrast and silhouette. Strong value separation survives desaturation; muddy mid-tones do not. |
The 120 × 45 thumbnail is auto-generated from the small capsule (462 × 174), which is why designing for the smallest render first is the standing recommendation.
Scoring anchors (1-10)
Every dimension shares the same anchor scale. Each dimension starts at 6, the competent baseline, and moves up or down on observable evidence in the asset. Scores of 9 and 10 are reserved for execution that holds up at small and tiny size, not at full size alone.
| Score | Label | What it means |
|---|---|---|
| 9-10 | Excellent | Stands out clearly at small and tiny size. Strong hierarchy, clean craft, memorable. |
| 7-8 | Good to solid | The normal range for a well-executed asset. Reads well small, with minor issues. |
| 6 | Competent baseline | Functional, with no notable strengths or weaknesses. Every dimension starts here. |
| 4-5 | Below average | Clarity, readability, or execution issues at small size that hurt discoverability. |
| 1-3 | Significant problems | Confusing, illegible, or actively misleading. |
The six dimensions
The full criteria for each dimension. The same criteria apply to every capsule the analyzer scores.
1. Genre Clarity
JudgesWhat kind of game this is at tiny (120 × 45) size. The first second of a player's attention.
Rewards
- Theme-specific cues a player can name without context
- Familiar genre iconography: card fans, dungeon tiles, base grids, an isolated horror subject
- Pose, setting, and composition that imply the actual gameplay
- Readable fantasy that signals the subgenre, not just the genre
Penalizes
- Mixed messaging where the visuals could imply two unrelated genres
- A generic theme with no gameplay type implied
- Misleading signals: cozy aesthetics on a horror game, and the reverse
| Score | What it looks like |
|---|---|
| 10 | At tiny size, the genre or subgenre is nameable from visuals alone. |
| 5 | It reads as a game; the genre stays ambiguous. |
| 1 | Communicates the wrong genre, or actively misleads. |
2. Title Readability
JudgesLetterforms, spacing, contrast, and legibility at full size and at tiny size, and whether the logo collapses when Steam scales it down.
Rewards
- Placement on a controlled background zone rather than noisy texture
- Letterform weight and outline that hold at 120 × 45
- High contrast between the title and the artwork behind it at the rendered size
Penalizes
- Decorative or ornate fonts that lose legibility small
- Taglines or subtitle text that become unreadable at small size
- Insufficient contrast between letterforms and background at tiny size
| Score | What it looks like |
|---|---|
| 9-10 | Reads cleanly at every rendered size, including 120 × 45. |
| 6 | Reads at full and small size; collapses at tiny. |
| 4 or below | The title is unreadable at tiny size. This is the usual score unless the title is intentionally minimal and still recognizable. |
3. Contrast & Color
JudgesHow the capsule holds up against Steam's #1b2838 background in a quick scroll, evaluated in color and in grayscale.
Rewards
- Strong value separation between subject and background
- A clean silhouette that survives desaturation
- Saturation control that keeps the image from going muddy
- Lighting separation between the subject and the background plane
Penalizes
- Limited value range and muddy mid-tones
- A subject that blends into its own background
- Unclear edges or weak silhouette readability in grayscale
- Dark-on-dark compositions that disappear against #1b2838
| Score | What it looks like |
|---|---|
| 9-10 | Strong silhouette and value contrast at every size, including in grayscale. |
| 6 | Reads cleanly at full size; blends at tiny size. |
| 1-3 | The subject is lost in grayscale or against #1b2838. |
4. Uniqueness & Polish
JudgesHow premium and distinct the capsule feels relative to common capsules in the same genre. The dimension that determines whether it is memorable.
Rewards
- A clear visual hook: a character silhouette, motif, distinctive prop, or unusual palette
- Clean craft: crisp edges, intentional typography, coherent effects
- Visual storytelling that communicates a selling point or core mechanic
- An art style distinct enough to identify the game later
Penalizes
- Cheap-asset vibe, template look, or a stock-pack aesthetic
- Random effects and default-feeling gradients
- A generic theme with no standout idea
| Score | What it looks like |
|---|---|
| 9-10 | Premium craft with a distinctive hook a player remembers. |
| 5-6 | Competent but generic. Nothing wrong, nothing memorable. |
| 1-3 | Reads as cheap or templated; indistinguishable from neighbors. |
5. Brand Consistency
JudgesInternal cohesion only: whether the capsule reads as one coherent piece, with a consistent rendering style, palette, and identity, or as a collage of mismatched assets. The engine is instructed not to assume prior familiarity with the game.
Rewards
- A consistent rendering register across every element
- A coherent palette and lighting language
- Recognizable identity cues: a character, motif, or signature treatment
- Genre conventions used well without feeling derivative
Penalizes
- Pasted-in elements that do not share lighting or rendering with the rest
- A photoreal element over a flat-shaded environment, or similar style clashes
- Generic presentation with no memorable identity cue
| Score | What it looks like |
|---|---|
| 9-10 | A single coherent image with a recognizable identity hook. |
| 6 | Mostly consistent; one element feels slightly off. |
| 1-3 | Visually mismatched assets; reads as a collage. |
6. Composition
JudgesHierarchy, focal point, balance, clutter, safe margins, and crop resilience across the capsule aspect ratios.
Rewards
- One clear primary subject at small and tiny sizes
- Supporting elements that guide the eye rather than compete with it
- Background, midground, foreground depth that creates a clear read
- Effective use of space, with no dead-center voids unless intentional
Penalizes
- Competing focal points that all collapse at tiny size
- Edge-hugging titles or subjects that get cut by Steam's crops
- Awkward empty gaps and scattered attention
- Equal emphasis on every element, with no hierarchy
| Score | What it looks like |
|---|---|
| 9-10 | A single focal point reads at every size; the composition survives every crop. |
| 6 | Works at full size; weakens at smaller crops. |
| 1-3 | Multiple competing focal points, or critical elements cropped at some size. |
How the 0-100 score is computed
The six dimension scores are averaged and multiplied by ten. There is no hidden weighting: six 9s produce a 90, six 5s produce a 50, and five 9s with one 3 produce an 80. One failing dimension costs a full tier without erasing five strong ones.
The bias toward small-size performance lives in the anchors rather than in the aggregation. A title that is unreadable at tiny size caps title readability at 4, and a subject that vanishes in grayscale scores 1-3 on contrast, so a capsule that fails at thumbnail size loses points on every dimension where that failure shows.
Score bands
The label attached to each overall score range, used on result cards and the leaderboard.
| Overall score | Label |
|---|---|
| 90-100 | Exceptional |
| 85-89 | Standout |
| 80-84 | Very Strong |
| 70-79 | Strong |
| 65-69 | Solid |
| 50-64 | Average |
| Below 50 | Weak |
Non-fabrication rules
Per-dimension explanations are constrained by the rules below. They prevent the engine from describing elements that are not present and require every judgement to be grounded in observable evidence.
- ·Only elements that are clearly visible may be described. Unreadable text, logos, characters, and details are never guessed at.
- ·If the title is not readable at a given size, the analysis states that it is not readable at that size.
- ·Brand consistency is scored from internal cohesion cues only, never from assumed recognition of the game.
- ·Every explanation must reference specific visible elements (top-left logo, center character silhouette, warm orange gradient), not generic adjectives.
- ·Every explanation must include at least one explicit statement about what works or fails at small or tiny size.
- ·If the capsule image is missing or not viewable, every dimension scores 1 and the first fix asks for a readable capsule image.
Sources we calibrate against
The rubric is calibrated against external authorities, not invented in isolation. Where Steam publishes a rule, the rule is the rule.
Official capsule dimensions and asset requirements.
What is banned on a base capsule: review scores, award logos, marketing copy, cross-promotion.
Steamworks: Visibility on Steam
What Steam's discovery algorithm rewards.
How tags drive 'More Like This', browse pages, and the Discovery Queue.
Methodology v1 · reviewed 2026-07-02
