game-test-categories-reference

Overview

A video-game build is tested against a stable set of categories that the industry has converged on over three console generations. Public sources - Microsoft's Certification step-by-step guide, Microsoft's Xbox Accessibility Guidelines (XAG) v3.2, and Steam's App Review Process - and longstanding QA practice in the field (Schultz & Bryant, Game Testing All-In-One, 3rd ed., ISBN 9781305077133, cited by stable ID per PLUGIN_AUTHORING.md Step 4 fallback; GDC Vault talks on test taxonomy, gated, cited by stable ID) name the same six buckets with minor wording differences.

This skill is the pure-reference taxonomy consumed by:

• the engine-specific skills (unity-test-framework, unreal-automation-system, godot-gut-tests),
• the builders (multiplayer-state-machine-coverage, gameplay-recording-replay),
• and the platform submission reference (platform-cert-overview-reference), which maps each category onto the four platform holders.

When to use

• Scoping QA effort for a new title - what categories must be covered before submission?
• Mapping a failed platform-cert finding back to an internal test category (e.g., "XR-130 fail" → compatibility-test).
• Planning a test-pass cadence (functional every PR, compliance + performance pre-submission, accessibility per milestone).
• Triaging a bug - which category does it belong to, who owns the fix?

The six canonical categories

These six categories are referenced throughout platform-cert-overview-reference when describing each platform's submission gates.

Category detail

1. Functional

Definition. Verifies that gameplay mechanics, AI, scripting, UI, save/load, networking, and economy match the design specification.

Sub-axes:

• Mechanics correctness - physics, hit detection, damage formulas, win/lose conditions.
• AI behaviour - pathfinding, decision trees / utility AI, difficulty curves.
• UI flow - every menu transition, every controller-disconnect recovery prompt, every "Press Any Button" front-end.
• Save / load - round-trip integrity, cross-platform saves, cloud sync conflicts, corrupted-save recovery.
• Economy / progression - XP curves, currency sinks, unlock triggers, anti-grief.

Typical tooling. Engine-native test framework (unity-test-framework, unreal-automation-system, godot-gut-tests) for unit + play-mode tests; recorded-input replay (per gameplay-recording-replay) for scripted scenario coverage.

Where it shows up in cert. Microsoft's Build Verification Testing (BVT) stage is largely functional: "BVTs is a reduced test pass designed to make sure a product is fully testable and configured properly before entering full certification testing." Functional bugs that break BVT block the entire cert pass.

2. Compliance

Definition. Verifies the build conforms to the platform holder's published Requirements / Policies / TRC document. Each platform holder maintains its own:

Each requirement maps to one or more test cases with explicit Test Steps, Expected Behavior, and Pass / Fail examples - see Microsoft's Certification Tested Xbox Requirements for Xbox Console Games.

Severity model (Xbox example, applies broadly). Per the Microsoft cert guide above, findings are classified as:

Sony TRC and Nintendo Lotcheck use comparable severity tiers internally (NDA - cite by stable ID).

See also. platform-cert-overview-reference for the submission workflow per platform.

3. Compatibility

Definition. Verifies the build runs across the full matrix of supported hardware generations, console SKUs, display modes, audio configurations, storage devices, controller types, and accounts.

The Xbox bench layout - per the Certification step-by-step guide - exemplifies the matrix:

(table from Microsoft's certification-guide § Certification test bench configuration)

The same matrix logic applies to Sony (PS4 base / PS4 Pro / PS5 / PS5 Pro), Nintendo (Switch original / OLED / Switch 2), and PC (GPU vendor × driver version × OS × DirectX/Vulkan).

Cross-generation specifically. Microsoft tracks this under XR-130: Xbox Console Families and Generations.

4. Performance

Definition. Verifies the title hits its declared frame-time, load-time, memory, thermal, and (handheld) battery budgets on each SKU in the compatibility matrix.

Sub-axes:

• Frame-time - target 16.67 ms (60 Hz) or 33.33 ms (30 Hz), measured under representative load (combat encounter, open-world traversal, cinematic).
• Load time - boot, level transition, fast travel.
• Memory - peak working-set, fragmentation under long sessions (a "soak" test in compliance vocabulary).
• Thermal - sustained-performance scaling on closed-chassis devices (Switch, Steam Deck, Series S in Energy Saving mode).
• Battery - handheld + portable (Switch, Steam Deck, mobile).

Where it shows up in cert. Microsoft requires a minimum sustained framerate per Xbox Requirements (XR) testing. Sony TRC includes performance gates (NDA - cite by stable ID). Nintendo Lotcheck includes thermal + battery rules for Switch (NDA - cite by stable ID).

5. Localization

Definition. Verifies that translated text, voice-over, dates, times, currencies, and culture-specific content render and behave correctly per locale.

Sub-axes:

• String fitting - does the German translation overflow the UI control? (German averages ~30 % longer than English.)
• Script support - CJK rendering (subpixel hinting, vertical text where used), Arabic / Hebrew bidirectional layout.
• Voice-over coverage - does every English VO line have a matching localized take? Lip-sync drift.
• Culture-sensitive content - flags, gestures, gambling references, religious symbols. (Each platform holder maintains per-region content rules under its compliance category - see Microsoft's Xbox Network Policies.)
• Date / time / currency formatting - locale-aware presentation.

Where it shows up in cert. Microsoft's bench layout (table above) explicitly varies Console Language across five consoles to exercise localized assets in the cert pass.

6. Accessibility

Definition. Verifies the title meets the accessibility guidelines published by the platform holder (Microsoft XAG v3.2 per learn.microsoft.com/.../accessibility/guidelines) and the regulatory framework that applies in the target market (US CVAA for in-game communications; EU EAA from June 2025).

XAG-style category structure. Each Xbox Accessibility Guideline, per the page above, has:

• Goal - desired player impact.
• Scoping questions - does this XAG apply to your title?
• Key areas to target - game elements the XAG covers.
• Implementation guidelines - prescriptive minimum bar.
• Example content - images / video of good implementation.
• Potential player impact - disability types affected.
• Resources and tools.

Common accessibility test sub-axes (cross-referenced to XAG identifiers where they exist; see learn.microsoft.com/.../accessibility/guidelines):

• Input - remappable controls, hold-to-press toggle, single-stick control schemes, Xbox Adaptive Controller support.
• Visual - high-contrast modes, text size, color-blind modes, screen reader / narrator output for menus (XAG 101 text size, cited in the guidelines overview above).
• Audio - subtitles (font size, background opacity, speaker attribution), visual cues for audio events.
• Cognitive - difficulty options, save-anywhere, pause-anywhere, reduced motion / camera-shake toggles.

Service. Microsoft offers an explicit Microsoft Gaming Accessibility Testing Service (MGATS) that tests against the XAG.

Cross-axis: multiplayer / online

Multiplayer is not a seventh category - it cuts across functional + compliance + compatibility + performance. The multiplayer-specific test surface is:

• State-machine correctness under packet loss / latency / disconnect - see multiplayer-state-machine-coverage.
• Matchmaking + session lifecycle - joins, drop-ins, host migration.
• Anti-cheat compatibility with platform integrity systems (Xbox LIVE; Sony's authentication; Nintendo Network).
• Cert-blocking risk. Per Microsoft's certification-guide, "Multiplayer does not work as expected" is one of the most common reasons a title is placed on Hold during a Xbox cert pass.

Cross-axis: content rating

Content rating (ESRB / PEGI / CERO / USK / IARC) is not a testing category - it is a regulatory classification authored from a content questionnaire, then validated by the rating board. However:

• Steam requires the Content Survey before review.
• Platform holders gate distribution on the local rating board's certificate (cite by stable ID for the NDA-only details).
• Localization (category 5) may interact with content rating - removing or recoloring blood for the German release, removing gambling minigames for the Belgian release.

Bug severity → category mapping cheatsheet

When triaging a bug, map it to one category to clarify owner + urgency:

Anti-patterns

Limitations

• NDA-only platform details. Sony TRC and Nintendo Lotcheck exact requirement numbers are gated. This skill cites them by stable ID per PLUGIN_AUTHORING.md Step 4 fallback; partners with NDA access should consult their developer portal for the authoritative current revision.
• Microsoft XR numbering drifts. XR identifiers (XR-024, XR-130, etc.) are revised across GDK releases - verify against the current Xbox Requirements list before quoting.
• Steam Direct is lighter than console cert. Per partner.steamgames.com/doc/store/review_process, Steam's review is store-page + build-startup focused (3 - 5 business days), not a console-style XR pass.
• Content rating boards are out of scope of this skill - they are regulatory classification, not QA categories. Consult ESRB / PEGI / CERO / USK / IARC documentation directly.
• No regional rating list here. Per-region content rules (German violence colouring, Belgian / NL loot-box rules, China / Korea content rules) are platform + region specific - consult the platform holder's regional supplement.