Active players / day (DAU proxy — by last save)
New players / day (by createdAt)
🎯 Real DAU — distinct players / day (event feed: login)
New vs returning / day (event feed)
Gold distribution
Gems distribution
🪙 Gold gained by source
💎 Gems gained by source
🪙 Gold spent by category
💎 Gems spent by category
Net flow — minted vs spent
📈 Daily minted vs spent — 🪙 Gold
📈 Daily minted vs spent — 💎 Gems
🪙 Gold spent — source → category → item
💎 Gems spent — source → category → item
⏱️ Offline return — away-time distribution
🐾 Pet boxes opened by source
📑 Feature usage — tab views
🐾 Pet gacha — pulls & new unlocks by source
Total level distribution
Quests completed / player
Average level per skill
Quest chain funnel (players reached → completed)
Weekly board — quests completed / player
Weekly board — milestones reached
🐾 Ownership by pet (players who unlocked each)
🎯 Pet usage — players currently deploying each
⭐ Star-level distribution (across all owned pets)
Pets owned / player
📈 Pet boxes opened / day (event feed)
🔓 Skill unlock rate (% of players)
🗺️ Episode progress (reached → fully completed)
Furthest episode reached / player
📜 Clue Scrolls held / solved by tier
Scrolls held / player
⚡ Clue Booster tier (players)
Held = unsolved scrolls sitting in inventory right now (a snapshot, not a lifetime drop count). Solved = lifetime expeditions completed per rarity from clueLog. A big held pile with few solves means players collect clues but don't run Expeditions.
Clears by dungeon
Dungeon clears / player
🔀 Divergence stashes — by winner & reason
🔒 Lease supersessions — device forced read-only
♻️ Silent rebases — conflicts resolved with no player-facing UI
Low counts are healthy. A rising divergence_stash rate means real players are hitting cross-device data-loss reconciliations (loser branch preserved to the conflict stash) — sizes the need for restore UI + server merge. lease_superseded counts every time the single-writer lease forces a second device read-only.
Since v1.81.0 (2026-08-20) the cloud doc is authoritative at session start and conflicts resolve by sync_rebase instead of a reload banner. Read it this way: replayed > 0 proves the event outbox rescued un-synced actions across a conflict; divergence stashes with reason Rebase Hydrate are expected (rebase still preserves the local branch), while stashes carrying the OLD reasons (“Higher Progress …”, “Higher Version”) after that date mean stale bundles are still in the wild. This card is the gate for Sync V2 Phase 7 part 2 — deleting the save-transaction merges and the total-XP guard.