Project report card · 2026-10-08
Kion’s playtest grades across presentation, gameplay and completeness.
Overall is an editorial assessment, not an average or a measured fidelity percentage. Grades improve through playtesting; passing automated checks alone does not raise them. Bars show the letter-grade scale, not percent complete.
Kion’s subjective assessment of the launch-experience updates on 2026-10-10 (revision 30700f02): the shared Rappy icon, minimal fan-remake notice and quieter cached startup bring the experience to A. This records the owner’s review of the updates, not a completed exported-device playtest.
Reviewed 2026-10-10.
Maintain the current presentation and finish exported-device verification. A startup movie is an optional path toward S; Kion is not prioritizing that level of polish.
Kion rates asset sync A following the launch and verified-cache improvements in PR #744 (merged revision d170622d). This records the owner’s assessment; exported-device verification remains a follow-up.
Reviewed 2026-10-10.
Playtest the verified-cache fast path, first download and retry recovery on real devices; confirm repeat launches do not flash the patch screen.
Faithful to the game and good overall; a few small improvements remain.
Preserve the current quality while addressing specific playtest observations.
Liked overall, but pixelated compared with the original Three.js mock.
Compare the same framing at target resolutions and fix the source of pixelation.
Presentation exceeds the original in places, but some angles break down. Principal’s office needs a final pass.
Inspect multiple camera angles and finish the principal’s office. Verify the Coliseum Master’s shared left-menu/right-NPC framing and playtest the drifting market cloud shadows and animated background crowd; grade unchanged.
Better than the original overall, but pixelated and still improvable.
Verify crisp rendering, readability and layout at supported resolutions, including effective levels in boss health bars; grade unchanged.
Close to the original’s feature set; needs substantial playtesting and object-placement polish. Wetlands floors/routing can trap the player.
Fix Wetlands stuck routes, then test traps, objects and traversal across room variations. Valley native-mirror free roam is visually approved. Playtest the expanded field-stage rollout, including area lighting, glow masks and mipmapped materials; broader traversal remains pending.
Kion raises Traps from D to B after the PR #761 laptop playtest on 2026-10-13. Distinct trap meshes, corrected materials and layouts, tactical enemy statuses, cycling needle/poison plates, rescuable bear traps, visible ball arming and the Coliseum trap page provide a playable tested baseline.
Reviewed 2026-10-13.
Blast and burn feedback are approved in #783; Kion approved #780 turret enemy targeting and spin/explosion behavior on 2026-11-05. Next, refine balance across field encounters. Original hostile radii, Light confusion, exact environmental timing/damage parity and the broader status/weapon-element audit remain open; intentional owner-directed remake behavior is documented in trap-validation.
Playable from start to finish; much of the early effort went here, but a return pass is needed.
Replay the story, recording progression, routing, encounter and presentation defects.
Kion raises Bosses from D to C on 2026-10-13: the existing boss encounters have received a first pass and work mechanically, even though they are not 1:1 with the original. This owner assessment follows the playable Shrine, Reyburn/Octo, Paru robot and Arca Humilias work in PRs #753, #754, #756, #757 and #758. Reyburn and Octo were accepted in Mac playtests for #754; the later Reyburn behavior and canyon backdrop were accepted on 2026-10-12 for #758 (behavior revision 08afa91f). This is a playable-mechanics baseline, not full original-game fidelity or complete campaign validation.
Reviewed 2026-10-13.
Refine boss feel, balance, hit geometry, effects and original behavior from the playable baseline. Remaining work includes Octo swallow/escape, robot targeting/movement and floor/status variants, Trinity piece visuals and attack delivery, and Falz form progression. Review campaign boss assignments and the Trinity-to-Falz handoff separately from the dedicated arena encounters. Complete Tower Mother placement/validation and the new Snowfield/Makara encounter designs; these are not covered by the existing-boss first pass.
Basic mechanics exist; lock-on and better enemy behavior are still needed.
Implement lock-on (#775) and continue enemy-family attack-rhythm review. Kion approved the #773 collision, obstacle-detour, Ghowl reach and post-hit protection changes in local Mac playtesting on 2026-11-05; recheck the final small Helion radius increase to 1.4 and crowded narrow passages. Playtest the shorter, model-sized box and wall destruction bursts.
Kion raises Systems from D to C after local Mac playtesting of PR #752 (6c2230c2 plus the playtest fixes committed in this PR) on 2026-10-11. Combat progression, inventory grinding with visible per-instance ATP and grind levels, dedicated Photon Blast activation and the segmented HUD gauge now support the core loop. The Coliseum exit blocker was reproduced and fixed with regression coverage; broader balance and fidelity work remain.
Reviewed 2026-10-11.
Playtest Equip → Sets (save/recall, bank retrieval, missing-item refusal, controller naming) and the EXP bar/level-up flash. Verify multi-area enemy levels and both death paths. Continue quest and free-roam playtesting after the Coliseum exit fix. Validate economy, class/EXP and enemy-level tuning, weapon specials/status fidelity, technique weapons and party Photon Blast behavior. Coliseum now exposes player/enemy levels, scaled HP previews and debug invincibility; verify hands-on tuning and controller use. Existing traversal defects remain separate follow-up work. Grade and assessment date unchanged.
A means convincing, faithful where intended, and polished with minor issues. B means good with visible gaps. C means the main loop works but important behavior or feel is incomplete. D means placeholder or rough implementation. Plus/minus refines the assessment; an unreviewed area stays ungraded.
A grade increase SHOULD cite a dated user playtest, the revision, the scenario and the observed improvement. Automated passes MUST NOT independently promote a grade. Better-than-original presentation is a deliberate design choice and SHOULD be distinguished from original-game fidelity.
Every PR SHOULD identify affected scorecard areas and link their issue filters, or state “No scorecard impact” with a short reason. Apply the relevant area labels to concrete GitHub issues; maintain membership there, not in a source-code list.
For affected areas, review the assessment and next review target in spec/src/data/fidelity-grades.json. Update stale wording when the change resolves a target or reveals a new gap. If implementation is complete but a playtest is pending, say so and retain the grade.
Change a grade only with Kion’s assessment backed by a dated playtest, tested revision, scenario and observations. Record that evidence in the PR and the affected area’s assessment; update the assessment date only when an assessment actually occurs. A merge, passing test or closed issue alone does not change a grade. Partial reviews MUST identify the reviewed areas rather than imply every score was reassessed.
Playtest base-disk availability, class restrictions, charged upgrades and fresh Force save/reload. Original timing, overall Force balance, enemy levels and economy remain open. Kion raised Systems to C after the PR #752 Mac playtest on 2026-10-11; these remaining technique and balance checks are follow-up work.
The existing suites provide substantial mechanics coverage. They do not establish every original attack, visual alignment or combat feel. Existing user approval of the gameplay foundation remains valid; family-level unknowns remain open.
Each area links directly to open issues with its score:* label. An issue may improve several areas and carry several score labels. The scorecard owns the assessment and review targets; issues describe concrete defects, investigations or improvements with a clear completion condition.
Use these filters to choose work without an umbrella issue or milestone per grade. Empty filters mean no open issues have been associated yet, not that the area is complete. Kion’s playtest observations can refine each review target independently of the issue list.