Player Engine — Goals Skeleton
The data side now has phases, gates, pre-registered experiments, and
tripwires (data-engine roadmap). The player side
has stronger vision documents (The Journey, the
Reward Economy, the spatial-wiki narrative) but no
equivalent goal structure — The Journey's second test ("does it make the
community stronger?") currently has no metric, no gate, no experiment.
This skeleton proposes that structure with identical discipline. Every
judgment call is marked [ALE] — it was drafted from the engineering seat
and the player side is not the engineering seat's to decide. Suggested use:
Ale edits ruthlessly, the offsite ratifies, and the
Operating Picture GAP rows close.
The claim (proposed wording — [ALE] to own it)
People play through Playroll because it makes their play better — more purposeful, more social, more rewarded, and eventually part of a living map of the worlds they love — not because it pays. Pay is the respect, not the reason. If that claim is false, the data engine is renting a workforce, and rented workforces churn.
Phases and gates (mirroring the data side, sharing its clock)
P0 — Keep the promises (now → +1 quarter, parallel to data Phase 0)
The trust foundations that make everything after honest:
| # | Item | Notes | Done when |
|---|---|---|---|
| P0.1 | Promise audit | The published promises ("you keep your keys", "no chat, no voice, no personal data", transparent agreement, quality-based pay) checked against built reality — the data side found one violation already (incoming voice in game audio); the player side owns the communication of the fix | Every public promise is either true or corrected, with Pioneers told directly [ALE: wording + channel] |
| P0.2 | Payout reliability baseline | Manual Altitude rail today; disputes and delays are trust events, not ops noise | Payout-dispute rate and time-to-paid tracked per tranche |
| P0.3 | Consent as a player experience | The new consent scopes (AI-training, sale, narration) arrive via re-consent overlay — the player side decides how it feels (plain language, no dark patterns, easy withdrawal) | Re-consent completion ≥ [ALE: target] with withdrawal genuinely one click |
| P0.4 | Recruiting spec handshake | The data side's thin input regimes (KBM-precise, narrators, full parties) handed to Kame + Mattia as an actual recruiting brief for the first time | Recruiting brief exists; funnel tagged against it |
Gate: no player promise is false, and paying a Pioneer is boring (predictable, on time, undisputed).
P1 — Prove the pull (parallel to the data side's POC + experiment)
The player-side founding experiment — the centerpiece this side is
missing. Candidate designs [ALE picks one, or replaces]:
- Quest-cohort vs free-play-cohort: same weeks, randomized offer; measure D7/D30 retention, session frequency, and voluntary free-play share after quests. Claim under test: quests add pull rather than substituting pay for joy. (This is the player half of the data side's naturalness measurement — same collection window, two lenses.)
- Bounty-framing A/B: identical quest, framed as "mission + reward" vs "task + payment"; measure completion quality and sentiment. Claim: framing changes whether it feels like play or work.
- The wiki teaser: show a cohort their own sessions as a 3D artifact (the offsite's reconstructed-session tab); measure effect on recording frequency. Claim: seeing the map is itself a reward.
Pre-registered like the data side's: metrics frozen before the cohort starts, negative results reported, no post-hoc threshold moves.
Gate: one pre-registered player-experiment result worth building on — the second test finally has a number.
P2 — First durable cohort (parallel to data Phase 2)
The player-side equivalent of "first revenue" is first retention: a
cohort that stays without novelty effects. Candidate gate [ALE]:
week-8 retention of the Phase-1-style cohort ≥ [target], with quest
participation voluntary-majority (players choose quests over the minimum).
Plus: the Ambassador program shaped ([ALE] — the program shape is flagged
as needing alignment since May), and the Season/EXP loop live enough to
carry identity between quest campaigns.
P3 — The community that maps worlds (parallel to data Phase 3)
The spatial wiki graduates from R&D teaser to the thing players show each
other. Community cartographers exist as an identity (the vision doc's
"cartographers, guides, and archivists of worlds"). Data-side campaigns
become invisible infrastructure behind experiences players would choose
anyway. Gate [ALE]: some share of new players arrive via community/wiki
pull rather than paid acquisition.
Metrics menu (drawn from what already exists — [ALE] sets targets)
From Player-Tier KPIs and the dashboards: retention cohorts, session frequency, validated hours per contributor. New, proposed: quest NPS (or a lighter "was this fun?" pulse), voluntary re-uptake rate, free-play share post-quest, payout-dispute rate, time-to-paid, consent-withdrawal rate, support sentiment, Discord participation breadth (not just volume). Principle inherited from the data side: a small set of honest numbers beats a dashboard of vanity ones.
Player-side tripwires (drafted for [ALE]'s edit)
| # | Signal | Response |
|---|---|---|
| PT1 | Quest NPS or re-uptake drops while data-side naturalness still looks fine | The metrics disagree — trust the player one; pause quest volume, redesign with Phil's UX findings |
| PT2 | "Gig work" framing appears organically in community language | Reframe rewards presentation before scaling anything (bounty-framing experiment gives the lever) |
| PT3 | Payout dispute or delay touches >[x]% of a tranche | Freeze campaign launches until the rail is boring again |
| PT4 | A trust event (privacy surprise, consent confusion) reaches the community | Direct, fast, plain-language response owned by community leads — the data side's provenance receipts exist to make this answerable |
| PT5 | Quest participation becomes compulsory-feeling (minimum-chasing dominates) | Rebalance economy toward EXP/identity rewards; quests must stay chosen |
The paired commitments (what each side owes the other)
- Data side owes players: bounties priced so honest play beats farming; quests triaged for fun, not just coverage; provenance receipts that make every trust question answerable; revenue that funds the reward economy.
- Player side owes the data engine: a fleet that stays; recruiting against the catalogue's real needs; the naturalness that makes the data worth a premium; the community pull that makes all of it cheaper than paid acquisition.
Open decisions for Ale (the actual asks)
- Own or rewrite the claim.
- Pick (or replace) the player-side founding experiment — ideally sharing the POC/experiment collection windows so one campaign feeds both sides' evidence.
- Set the P2 retention gate and the metrics-menu targets.
- Edit the tripwires into ones the community team would actually pull.
- Decide the Ambassador program shape (flagged open since May).
- Name the owner-pairing for each metric (Phil's UX findings, Kame+Mattia's funnel, Denise+Luca's community signals all feed this — who reports what, at what cadence).
Drafted 2026-07-08 from the engineering seat as a mirror of the data-engine roadmap's structure. Everything above is Ale's to keep, bend, or bin — the only position this document defends is that the second test deserves the same rigor as the first.