# Phase 7.2 Next-Step Proposal: Hi Fi Snake build026.zip / build027.zip

## Proposal Summary

The primary focus of the next build cycle must shift from basic input detection to resolving critical performance bottlenecks and spatial navigation failures that are driving high early churn. While joystick responsiveness was successfully improved in build027, players are still exiting in under 90 seconds due to severe low FPS on common mobile GPUs, missing doors by a single tile (spatial navigation failures), and dying instantly to the Room 2 lava hazard. We must optimize rendering resolution on mobile, widen door alignment tolerances, and soften early hazard scaling to make the game playable and fair.

## Urgent

* **Optimize Mobile GPU Performance (Cap Device Pixel Ratio)**: Cap `window.devicePixelRatio` (e.g., to a maximum of 1.5 or 2.0) on mobile browsers, particularly for Mali-G52, Mali-G51, Mali-G57, and PowerVR GPUs, to resolve severe frame rate drops. (Evidence Class: evidence_backed; Source Finding: Low FPS / Performance issues on mobile GPUs)
* **Widen Door Collision and Alignment Tolerance**: Implement a forgiving alignment window or magnetic snapping when players approach a door. Players who are off by a single grid tile should be guided through the door rather than colliding with the adjacent wall and dying. (Evidence Class: evidence_backed; Source Finding: Spatial navigation failures)
* **Soften Room 2 Lava Hazard Pacing**: Delay the initial spawn or reduce the starting speed of the approaching lava hazard in Room 2. Players currently freeze or die instantly before they have time to react or turn. (Evidence Class: evidence_backed; Source Finding: Punishing Room 2 lava hazard)

## Next

* **Expand Virtual Joystick Active Bounds**: Allow the virtual joystick to continue tracking input even when the player's finger drags slightly outside the visual joystick boundary. (Evidence Class: likely; Source Finding: Joystick dragging outside active bounds)
* **Refine Grid Turning and Input Buffering**: Implement a short input buffer for perpendicular turns so players do not have to time their turns perfectly on the grid, reducing the perceived physical effort of turning. (Evidence Class: likely; Source Finding: Turning avoidance / high effort)
* **Add Store and Customization Error Feedback**: Provide clear visual or audio feedback (e.g., "Not enough currency") when players attempt to purchase locked skins in the store. (Evidence Class: likely; Source Finding: Store and customization flow friction)

## Later

* **Advanced Upgrade Synergies & Economy Rebalancing**: Defer adjustments to late-game upgrades and economy until first-session retention and performance are stabilized. (Reason: Players are churning too early to experience these systems.)
* **Daily Return Hook & Cosmetic Expansion**: Defer additional cosmetic items and daily challenges. (Reason: Focus must remain on making the first 2 minutes of gameplay smooth and playable.)

## Keep

* **Active-Input Player Movement**: Retain the active-input control model (stopping immediately on release) as it provides precise, intentional movement, but refine the turning mechanics to make it feel less tedious.
* **Player Speed Progression Formula**: Keep the starting speed at 4 and scaling of 0.4 per room, as this provides a solid foundation once spatial navigation and performance are resolved.

## Try

* **Visual Alignment Cues**: Try adding subtle, dynamic visual arrows on the floor that light up when the player is perfectly aligned with the door path, reinforcing the correct route.

## Avoid / Defer

* **Avoid Adding Complex Level Layouts**: Do not introduce tighter corridors or more complex hazards yet. (Evidence Class: evidence_backed; Source Finding: Spatial navigation failures and turning difficulty make tight spaces highly frustrating.)
* **Defer Shop Expansion**: Avoid adding more cosmetic items until basic transaction feedback is implemented. (Evidence Class: likely; Source Finding: Store and customization flow friction.)

## Reference Trinity Impact

| Reference Role | Impact | Explanation |
| --- | --- | --- |
| Primary | strengthened | Fixing turning effort and door alignment directly supports the classic Snake-like challenge of navigating self-generated obstacles without feeling cheated by the controls. |
| Secondary | unchanged | The room-by-room progression and upgrade cadence inspired by Archero remain intact, though early room hazards are softened for onboarding. |
| Tertiary | strengthened | Resolving the mobile GPU performance issues ensures that the neon arcade spectacle inspired by Pac-Man 256 is visible and smooth for all players. |

## Evidence Classification

| Proposed Change | Evidence Class | Supporting Evidence |
| --- | --- | --- |
| Cap device pixel ratio on mobile | evidence_backed | Multiple playtest sessions explicitly reported low FPS and lag on Mali and PowerVR GPUs, confirmed by developer notes. |
| Widen door alignment tolerance | evidence_backed | Players frequently collected pellets but hit the wall right next to the door instead of going through. |
| Soften Room 2 lava hazard | evidence_backed | Multiple players died immediately upon entering Room 2 due to lava speed or freezing under pressure. |
| Expand virtual joystick bounds | likely | Mobile players frequently dragged their fingers outside the active joystick area and lost control. |
| Add store error feedback | likely | A player spent significant time trying to buy skins in the store without having enough currency, indicating a lack of clear feedback. |

## Root Causes

* **High Rendering Overhead**: High device pixel ratios on low-end mobile devices cause WebGL performance to degrade severely, leading to low FPS and input lag.
* **Unforgiving Grid Alignment**: The grid-based movement requires pixel-perfect alignment to pass through doors, causing players to collide with walls when they are off by a single tile.
* **Punishing Early Hazard Scaling**: The Room 2 lava hazard approaches too quickly, punishing players before they can adapt to the active-input movement scheme under pressure.

## Production Risk

* **Device Fragmentation**: Ensuring robust performance optimization across various mobile browsers and low-end GPUs without compromising visual quality.
* **Tuning Balance**: Adjusting door alignment tolerances must not allow players to clip through solid walls or bypass intended obstacles.

## Next Test Hypotheses

* **Hypothesis 1**: Capping the device pixel ratio on mobile devices will eliminate low FPS reports and increase average session length on low-end Android devices by 30%.
* **Hypothesis 2**: Widening the door alignment tolerance will eliminate spatial navigation failures (hitting the wall next to the door) and increase Room 1 completion rates.

## Suggested Version Note

"Performance and navigation update: Optimized mobile GPU rendering, widened door alignment tolerances to prevent accidental wall collisions, and softened early Room 2 hazard scaling."

## What To Watch Next

* Observe if mobile players on Mali/PowerVR GPUs experience smooth frame rates.
* Track whether players successfully align with and pass through doors without hitting adjacent walls.
* Monitor if players survive the transition to Room 2 and successfully engage with the lava hazard.