We Recovered Mismatched Save Slot Labels After Steam Branch Promotion - 2026 Case Study

This is a synthesized case study—a field pattern seen across May–October 2026 fest demos on Steam Windows builds, especially teams that promoted a playtest or internal branch to a public fest depot without updating save slot label maps. It is not a named studio turnaround story. There are no invented revenue figures, no fake Discord quotes, and no made-up refund percentages.

What follows is the failure signature, the isolation order that worked, the artifacts that proved recovery, and how the same pattern maps to your project if players say “I saved in Slot 1 and it loaded empty.”

Pair this narrative with GDevelop save-path freeze, GameMaker export sanity, Ren'Py label freeze, and BUILD_RECEIPT so you have procedure plus story.

Non-repetition note: Freeze-week posts teach prevention. This case study describes what broke after promotion when slot UI and persistence keys diverged—not another seven-day gate list.

Why this matters now (May–October 2026)

1. Weekly fest promotions — Producers merge playtest into fest_public without a slot-label diff.
2. “Lost progress” threads — Comments rhyme with gameplay / save tags in refund dashboards—but root cause is label map drift, not cloud outages.
3. Multi-slot UI debt — Menu shows Slot 1 while code writes save_slot_b after a refactor.
4. Cross-engine blind spots — GDevelop Storage keys, GameMaker ini sections, and Ren'Py persistent fields each fail differently; the symptom is identical.
5. Evidence gap — Teams ran Wednesday smoke on boot path only, not save/load matrix.

Direct answer: Recovery required save_slot_label_map_v1.json, a promotion-day slot round-trip on installed Steam build, save_slot_recovery_receipt_v1.json, and a BUILD_RECEIPT row tying build_label to the map hash before the next branch merge.

Beginner quick start — what recovered means here

Recovered in this pattern means:

1. Reproduced wrong slot load on installed Steam build—not editor only.
2. Mapped every UI label to exactly one persistence key per build_label.
3. Logged receipt JSON and one screen recording of save → quit → load.
4. Reduced new “lost save” comments for seven days on the same build_label (qualitative, not a fabricated %).

It does not mean cloud saves work globally or that Steam Cloud was enabled—many fest demos are local-only with honest FAQ copy.

Success check: A producer can load Slot 2, quit, relaunch, and Slot 2 still shows the same chapter or floor—without engineer jargon.

The player-facing failure (typical signals)

Internal team language before fix: “Saves work on my machine” — because QA used one slot and local dist folder, not installed client with three labeled slots.

Starting state (what was wrong)

Gameplay was fine. Trust in saves was not.

Root cause pattern (synthesized)

Three recurring mechanics—often combined:

None require malice—only missing map file at promotion time.

Timeline (five working days — pattern timing)

Teams with existing GDevelop freeze save_path_map.json recovered in two days; teams without any map took five.

Hour-by-hour D1 (reproduction discipline)

If Slot 1 loads empty but Slot 2 shows Slot 1’s progress—you are in this case study.

D2 — save_slot_label_map_v1.json (inventory)

Create release-evidence/saves/save_slot_label_map_v1.json:

{
  "schema": "save_slot_label_map_v1",
  "build_label": "fest-demo-2026-05-24-rc2",
  "slots": [
    {
      "ui_label": "Slot 1",
      "ui_index": 1,
      "persistence_key": "demo_save_slot_a",
      "storage_surface": "local_file",
      "expected_path_hint": "%APPDATA%/Studio/Game/saves/"
    },
    {
      "ui_label": "Slot 2",
      "ui_index": 2,
      "persistence_key": "demo_save_slot_b",
      "storage_surface": "local_file",
      "expected_path_hint": "%APPDATA%/Studio/Game/saves/"
    },
    {
      "ui_label": "Quicksave",
      "ui_index": 0,
      "persistence_key": "quicksave",
      "storage_surface": "local_file",
      "notes": "Hidden from fest UI—must not steal Slot 1 key"
    }
  ],
  "faq_claim": "three manual slots local only",
  "cloud_enabled": false
}

Pass: Every visible UI slot has exactly one persistence_key; no duplicate keys.

Engine-specific inventory notes

One map file per SKU—do not merge demo and full-game maps without a sku_id column.

D2 deep dive — grep and screenshot discipline

Programmers export evidence—not opinions:

Pass: Screenshot string Slot 1 appears on same line in map JSON as ui_label, and persistence_key appears in grep log exactly once.

Studios using validate-packet added a optional check: save_slot_label_map_v1.json must exist when demo_has_save_slots=true in packet manifest.

D3 — fix mapping and migration

Minimum fix set:

1. Align ui_index to persistence_key in code or rename keys to match shipped UI.
2. On load, if old key exists under legacy name, migrate once with logged message.
3. Show save_incompatible label instead of silent empty slot when migration impossible.

Example migration block (conceptual—adapt to engine):

IF legacy_key "save1" exists AND "demo_save_slot_a" empty
  COPY legacy → demo_save_slot_a
  DELETE legacy
  LOG "migrated save1 → demo_save_slot_a for build_label rc2"

Scope rule: No new features during D3—mapping fixes only.

Align FAQ save lines with faq_claim row in map.

D4 — slot round-trip matrix

Log gate_slot_matrix.log:

Run matrix on:

• Fresh install (no prior saves)
• Install after one promotion (simulates fest player updating mid-week)

Pair with metadata diff on same promotion day—store copy and slot behavior ship together.

D5 — save_slot_recovery_receipt_v1.json

{
  "schema": "save_slot_recovery_receipt_v1",
  "build_label": "fest-demo-2026-05-24-rc2",
  "map_path": "release-evidence/saves/save_slot_label_map_v1.json",
  "map_sha256": "sha256:example…",
  "matrix_log": "release-evidence/saves/gate_slot_matrix.log",
  "slots_tested": 3,
  "matrix_pass": true,
  "migration_applied": true,
  "recovery_pass": true,
  "promotion_allowed": true,
  "signed_at": "2026-05-28T17:00:00Z",
  "reviewer": "GamineAI Team",
  "notes": "Installed Steam build; cloud off; pairs GDevelop freeze week 2026-W21"
}

Attach to BUILD_RECEIPT upload row:

What promotion day should have caught

Total ~27 minutes—cheaper than five-day recovery.

Add row to Wednesday smoke optional S9: “Slot 1 round-trip pass.”

Refund tag crosswalk (qualitative)

In refund dashboard spreadsheets, teams reported clusters of:

After receipt shipped, new rows on same build_label with save language dropped in several teams’ qualitative notes—not a guaranteed metric for your studio.

Playtest form fields (D2 add-on)

Use 18 playtest tools for intake—require these four fields on save-related tickets.

Branch promotion failure modes

Rule: Promotion checklist includes map hash diff, not only binary smoke.

Comparison to replay refresh case study

Both live under release-evidence/—different subfolders.

Comparison to Ren'Py persistent drift

Ren'Py freeze week prevents label renames mid-fest. This case study covers UI slot numbers vs persistence keys after branch promotion—VN teams hit both; run freeze week before promotion, slot map on promotion day.

Developer checklist (PR template)

• [ ] save_slot_label_map_v1.json matches UI screenshot
• [ ] No duplicate persistence_key values
• [ ] Slot matrix pass on installed Steam build
• [ ] save_slot_recovery_receipt_v1.json updated
• [ ] FAQ save claims match faq_claim
• [ ] BUILD_RECEIPT row includes map hash
• [ ] Migration log attached if keys renamed

Producer one-pager

Save slot honesty means Slot 1 on screen writes to the key players expect after every promotion. We prove it with save_slot_recovery_receipt_v1.json before fest branch merge. Playtesters: send build_label and which slot you used.

Anti-patterns in the same threads

Ninety-minute “are we in this case study?” audit

If first step fails—you are in this case study.

Incident retrospective (45-minute workshop)

1. Why was slot matrix missing from promotion checklist?
2. Who owns map updates when UI strings change?
3. Do playtest forms capture ui_slot?
4. Is playtest branch isolated from public fest depot?
5. Does Wednesday smoke include save round-trip?

Forward-link: playtest invite isolation playbook when backlogged—separate scopes need separate maps.

When this case study does not apply

• Single-slot arcade games with no load menu
• Roguelites with one run autosave only (see replay case study instead)
• Games with no Steam branch promotions this quarter
• Teams that only ship itch HTML5 without labeled slots

Community response template (Steam discussion)

Thanks for the report. Build fest-demo-2026-05-24-rc2 fixed a slot-label mismatch from the previous branch—please update. If you are on rc2 and Slot 1 still loads empty, reply with build_label from the title screen and which slot you used so we can match your save file.

No invented stats—build label and slot only.

Cost of five-day delay (opportunity framing)

Five days beats two weeks of fest-week comment firefighting.

Cross-training for non-programmers

Producers can run D1 matrix without opening engine:

1. Install Steam build from checklist.
2. Play to mid-demo.
3. Save each slot; quit; reload each.
4. Screenshot pass/fail table.
5. File form.

Engineering fixes maps; whole team detects.

Link graph (save discipline family)

New hires read 5 → 1 for motivation before procedure.

Cloud saves caveat

If cloud saves parity applies, extend map with cloud_enabled: true and Steam Cloud API row per slot. This synthesized pattern often involved local-only fest demos where players assumed cloud sync—fix FAQ before enabling cloud.

Defold and collection saves

Defold collection audit addresses save root discovery. Slot-label mismatch can still occur if UI uses slot numbers but collections write to a single game.state file—add collection_slot column to map JSON.

Audio and trust (tangential)

Players who mute after save frustration still leave bad reviews—optional menu loop check via LUFS listicle does not fix slots but reduces compound negative first impressions after a save scare.

Friday Block 5 line

Add to Friday Block 5:

slot_matrix=pass/fail; save_slot_receipt_path=release-evidence/saves/save_slot_recovery_receipt_v1.json

Fifteen seconds prevents promoting a build that only passed boot smoke.

What we would do differently (retrospective)

Ship save_slot_label_map_v1.json before first public fest promotion—not after the first “lost progress” thread. Run slot matrix on installed Steam the same hour as BUILD_RECEIPT upload. Treat map hash diff as a merge blocker equal to broken exe.

That ordering collapses this five-day pattern into one promotion afternoon—and leaves the team arguing about story and art during fest month instead of defending slot numbering in discussion threads.

Key takeaways

• Synthesized pattern—no invented studio metrics; portable artifact list.
• UI slot labels must map 1:1 to persistence keys per build_label.
• save_slot_label_map_v1.json is the inventory heart.
• save_slot_recovery_receipt_v1.json proves recovery before next promotion.
• Reproduce on installed Steam, not editor or dist folder only.
• Five-day timeline is indicative; prevention takes one promotion checklist.
• Pairs with GDevelop, Ren'Py, GameMaker, and Construct save discipline posts.
• Refund dashboard tags often mislabel slot bugs as cloud issues.
• Add S9 slot round-trip to Wednesday smoke optional gates.
• Migration messages beat silent empty slots.
• Playtest forms need ui_slot and build_label.
• Branch promotion without map diff is the usual trigger.
• Producers can run the ninety-minute audit without engine access.
• FAQ faq_claim row must match real slot count and cloud state.

FAQ

Is this the same as corrupt save files?
No—files may exist under a different key than the UI slot the player chose.

Do we need Steam Cloud?
Not for local-only demos—honest FAQ matters more than enabling cloud.

Does this replace GDevelop freeze week?
No—freeze prevents path bugs; this case study covers post-promotion label drift.

Can one slot work and two fail?
Yes—matrix testing all visible slots is mandatory.

Should quicksave share Slot 1’s key?
No—give quicksave its own persistence_key in the map unless FAQ says they are the same slot.

Post-recovery monitoring (seven-day qualitative window)

No target percentages—track direction only.

Partner diligence attachment

Q3 diligence reviewers ask what happens to player progress when branches merge. Attach save_slot_label_map_v1.json, save_slot_recovery_receipt_v1.json, and optional 60s three-slot matrix capture—answers slot honesty without inventing retention stats.

Store-demo mismatch crosslink

When saves fail, players also report store-demo mismatch symptoms—treat slot map as binary truth beside metadata diff, not a separate crisis narrative.

Final honesty block

GamineAI Team publishes synthesized patterns to reduce repeated support questions across indie teams. Your comment volume, refund rate, and engine will differ. The artifact list is the portable part—timelines are indicative, not guarantees.

Reader assignment (optional)

Assign one person to run the ninety-minute audit and attach qa-slot-label.log to your next BUILD_RECEIPT. You either confirm you are not in this case study—or you start D1 the same day. Share the log in your next producer standup so art and narrative leads know saves are unblocked before trailer week.

Conclusion

Fest promotion without a slot-label receipt recreates “lost progress” threads even when saves exist on disk.

Map the slots. Test the matrix. File the receipt. Then merge the branch. Players forgive hard games faster than saves that lie about which file they touched.

Next reads: GDevelop save-path freeze, Wednesday demo smoke, BUILD_RECEIPT pipeline.