Pass 1 recruited 14 testers across casual, mid-core, and core gaming profiles. Pass 2 used a 7-tester subset focused on verification of high-priority fixes, weighted toward the casual profile that experienced the highest abandonment in Pass 1.
Resolved rows are dimmed. Status reflects Pass 2 verification. Frequency shown is Pass 1 (n=14) unless marked P2.
| # | Location | Description | Severity (P1) | Frequency (P1) | Status |
|---|---|---|---|---|---|
| F-01 | Launch & Permissions | Camera, microphone, and contacts requested on first launch before any gameplay context. Caused 5 of 14 testers to uninstall immediately. | Critical | Resolved | |
| F-02 | IAP Prompt | Monetisation overlay appeared after first dungeon death, before the player reached any reward. 9 of 14 testers reacted negatively; 3 uninstalled at this point. | Critical | Resolved | |
| F-03 | IAP Prompt | No dismiss option visible on IAP overlay for 4 seconds. Testers assumed the app had locked them into a purchase flow. | Critical | Resolved | |
| F-04 | Combat Tutorial | Two-finger diagonal swipe mechanic taught by static icon only — no animation, haptic feedback, or retry prompt. All completing testers attempted this incorrectly multiple times. | High | Open | |
| F-05 | Combat Tutorial | Tutorial countdown timer runs during instructional overlays, creating pressure to skip reading. 6 testers did not read combat instructions before proceeding. | High | Open | |
| F-06 | Map Navigation | World map dungeon icons unlabeled. Testers could not distinguish story from challenge dungeons; 7 entered challenge content with starter equipment. | High | Partial | |
| F-07 | Map Navigation | "Return to base" button not visible on small displays (≤5.4"). Hardware back gesture caused accidental app close on iOS. | High | Resolved | |
| F-08 | IAP Prompt | Currency naming ("Vault Gems" vs "Vault Coins") confused testers — unclear which was premium vs earned. 5 testers were uncertain whether they were spending real money. | High | Resolved | |
| F-09 | Combat Tutorial | Enemy attack audio warnings too quiet at device default volume. Visual fallback absent. All testers missed audio cues in ambient conditions. | Medium | Open | |
| F-10 | First Dungeon | Death screen shows no path back to dungeon selection — only "Try Again" or "Home." 4 testers assumed they'd lost progress and quit the session. | Medium | Resolved | |
| F-11 | Character Select | Character stat tooltips require undiscoverable long-press. All testers chose characters by appearance; none read stats before committing. | Medium | Open | |
| F-12 | First Dungeon | Reward animation plays before loot is actually credited to inventory. 3 testers navigated to inventory mid-animation and saw empty slots, assumed a bug. | Medium | Resolved | |
| F-13 | Map Navigation | No visual indicator of player level or progression status on map screen. Testers returning to map after dungeon completion couldn't assess what to do next. | Medium | Open | |
| F-14 | Combat Tutorial | UI elements overflow their containers on MIUI / ColorOS. Ability icons partially obscured on Xiaomi and OPPO devices. Not reproducible on stock Android or iOS. | Low | Open | |
| F-15 | First Dungeon | MIUI background app management terminated session during dungeon load. No session restore on relaunch — tester returned to home screen with no progress saved. | Low | Resolved | |
| F-16 | Character Select | Confirmation dialog for character selection uses "OK" as the confirm label — not the character's name. Low urgency; caused brief hesitation in 4 testers. | Low | Resolved |
With 14 testers in Pass 1, several behavioural patterns emerged across player profiles that would not be visible at pilot scale. These patterns inform the prioritisation in Section 08.
All 5 casual-profile testers who uninstalled cited the permissions prompt as the trigger. Core and mid-core testers showed no abandonment at this step — they were accustomed to the pattern. This profile-specific finding would not be statistically visible with fewer than 8–10 testers and demonstrates why panel size matters for risk surface coverage.
Across all player profiles, the two-finger swipe failure pattern was consistent — though core gamers recovered faster. The shared failure suggests this is a UX design gap, not a familiarity gap. The static icon instruction format fails regardless of gaming experience level.
Mid-core testers were the most likely to spend time on the map trying to understand the icon system before selecting — and the most likely to feel frustrated when they couldn't decode it. Casual testers simply tapped the nearest icon. Core testers explored methodically. The mid-core pattern is the highest-risk for early churn from this group.
The IAP overlay triggered negative verbal reactions from 9 of 14 testers across all profiles. The 4-second dismiss delay compounded this — testers assumed dark pattern intent. After the fix in Pass 2 (delay removed, prompt timing adjusted), zero testers cited the IAP flow as a negative.
The power slash two-finger diagonal swipe should be demonstrated through a looping animation at native resolution, showing finger placement and direction. Add a subtle haptic confirmation on correct execution, and a "try again" coaching prompt after 2 failed attempts. This is the highest-remaining-priority fix — it affected 9 of 14 Pass 1 testers and remains the primary driver of tutorial abandonment in the casual segment. Estimated dev effort: 1–3 days (animation asset + haptic trigger).
The tutorial timer should pause automatically whenever any instructional modal or overlay is shown and resume only when the tester dismisses it or begins interacting. This is a behavioural design fix, not a visual one — the timer can remain visible but inactive. Consider also offering a "take your time" mode as a first-launch option for players who indicate they are new to action RPGs.
The client's patch added colour coding to dungeon icons, which was insufficient — 3 of 7 Pass 2 testers still entered incorrect content. Icons must include a short text label visible on the map view (not only on tap), plus a level range indicator (e.g. "Lv 1–5"). A confirmation prompt for out-of-range dungeons ("This dungeon is recommended for Lv 10+. You are Lv 2. Continue?") should be implemented as a short-term safeguard while the icon system is fully redesigned.
Enemy attack warnings require a visual component — a brief screen edge flash or ability icon pulse — that functions without audio. Increase default mixing for warning sounds by approximately 6dB relative to background music. This is an accessibility improvement as much as a usability fix and should be treated as baseline quality for a mobile release.
Add a visible info icon (ⓘ) to each character card that opens the stat tooltip. The long-press interaction can remain but should not be the only path to this information. Consider displaying one or two key stats (e.g. Attack, Defence) directly on the card without requiring any interaction at all, to make the character choice feel more informed from the outset.
The ability icon overflow on Xiaomi and OPPO devices is caused by non-standard safe area insets in MIUI 14 and ColorOS 13. These manufacturer skins represent a significant share of the Android install base in key markets (South/Southeast Asia, parts of Europe). Recommend adding MIUI and ColorOS to the standard device test matrix and auditing all UI containers against their safe area specifications.
The following table summarises verification results for each issue the client addressed between passes. Pass 2 testers were not informed which issues had been fixed.