M1 starter pack: local playtest record¶
Reviewed 8 September 2026. M1 is implemented and available in the local game. Human pacing, physical-controller ergonomics and a longer native-display soak remain playtest work; this record does not claim the six-frontier war is already built.
Play the route¶
- Enter Greyfen with a new character, or continue an existing save. Press J for the journal.
- Take A Blade Before Breakfast from Rusk. Use sword and mace at the training yard.
- Elen offers Tracks by the Road. Inspect the broken wagon on the west fork, defeat three wolves and return.
- Find Nella for The Hollow Howl. Inspect the spoor, then defeat the scarred pack leader. Sidestep its committed lunges and interrupt the howl.
- Complete Rusk's existing Goblins at the Fenline. Elen then offers Greyfen's Missing Stores: recover both marked bundles inside the camp.
- Return the stores, then take Quiet the Watch from Rusk. Cut the alarm rope before freeing Lysa. Return for the reward.
- Continue Varyn and Rusk's existing shrine/Emberwake sequence. Archery, magic and the tougher Ironwood branch remain available.
Actual local scene, positioned using the isolated QA server's capture controls. The cage door and short exit respond to each character's acknowledged rescue.
Completed checks¶
| Check | Evidence |
|---|---|
| Content migration | 12 quests, 590 zone objects, six wolf spawns and nine residents. Repeating installation does not duplicate content; eight original quest schemas/unlocks remain intact |
| Starter expedition | A fresh character completed all four M1 quests plus the original steel lesson, Fenline, shrine and Emberwake quests through ordinary server commands. Two deaths and recovery; no admin, invulnerability, teleport or injected progress |
| Shared / personal progress | Two players independently recover the same bundles and rescue the captive. Recent nearby damage contributors receive kill credit. Ordered objectives and interaction distance/cover are enforced |
| Reconnect / rewards | Partial objectives survive a new persistence instance. Returning players finish and turn in once; replayed turn-in messages cannot duplicate the reward; final state and coins survive another reload |
| Wolf combat | Fixed-direction lunge can be sidestepped; stagger interrupts it; cover blocks hits; howl alerts at most two existing local wolves; alarm reserves are bounded |
| Blender / animation | 16 model variants pass attachment, atlas, LOD and geometry-budget checks. Wolf pose tests cover 30/60/200 Hz. Browser encounters show idle, running, windup and attack states |
| Journal | Search, prerequisites, ready/archive states, three-pin limit, reorder, per-character reload, input blocking, Tab wrap, standard-controller API navigation and small-screen scrolling pass |
| Regression checks | Type checking and all 31 movement, character, environment, batching, starter-content and wolf-animation tests pass |
The expedition advances server simulation faster than wall-clock time and uses an automated aiming/navigation driver. Its 375.1 simulated seconds is not a human completion-time estimate. The proposed 45–60-minute first session needs human playtesting. Archery/magic lessons and the two Ironwood quests were not completed in this expedition; their definitions were preserved.
Frame delivery with the new content¶
Production client, isolated QA backend, automated local Chrome at 1280 × 633, DPR 1, 100% render scale, at its available 60 Hz cadence:
| Capture | Frames | p95 | Worst | Frames over 33 ms |
|---|---|---|---|---|
| Straight road movement | 240 | 17.9 ms | 18.4 ms | 0 |
| First burst/turning sample before expanded warm-up | 600 | 17.7 ms | 49.3 ms | 1 |
| Same sequence repeated before the fix | 600 | 17.3 ms | 17.8 ms | 0 |
| Fresh page after expanded effects warm-up | 600 | 17.1 ms | 19.7 ms | 0 |
The straight traversal has no frozen or oversized steps across its 210 steady-state samples. The effects check includes fast rotation, movement, real spell input and staged repeated bolts/rifts. After the initial first-use hitch, loading now uploads the common effect meshes and warms the fog shader variant used by the world. The final fresh-page sample reports no failed shaders or browser runtime errors.
These short captures support the local improvement. They do not establish 120/200 Hz presentation on the physical M4 Pro display, a five-minute soak or 20-player capacity. The hardware acceptance target remains explicit.
Records and commands¶
Browser, UI and frame-time record · Server expedition events · M1 content brief
npm run typecheck
npm run test:starter
npm run test:smoothness
npm run test:characters
npm run test:environment
npm run validate:characters
node --import tsx tools/starter-expedition.ts
npm run build
Tests use temporary copied zone/quest data. The normal local server was restarted with the updated content after saving its existing player records. No public deployment or live territory event was performed.