@ak--47/dungeon-master 1.5.4 → 1.6.0

package/dungeons/vertical/gaming/gaming.sql

@@ -0,0 +1,249 @@
+-- ============================================================
+-- gaming.js — v1.6 human-inspection queries (DuckDB)
+--
+-- Every query is keyed to a story id in gaming.js's `stories` export;
+-- the machine-checked verdicts come from:
+--   node scripts/verify-stories.mjs dungeons/vertical/gaming/gaming.js --data-prefix verify-gaming
+-- Generate first:
+--   node scripts/verify-runner.mjs dungeons/vertical/gaming/gaming.js verify-gaming
+-- Run this file:
+--   duckdb -c ".read dungeons/vertical/gaming/gaming.sql"
+-- ============================================================
+
+-- ── identity-resolution prelude ─────────────────────────────
+-- avgDevicePerUser: 2 — resolve device_id → canonical distinct_id via the
+-- USERS shards so uniques are identity-correct even for device-only rows
+-- (pre-auth steps of the first funnel carry device_id only).
+CREATE OR REPLACE VIEW users AS
+SELECT * FROM read_json_auto('data/verify-gaming-USERS*.json', sample_size=-1, union_by_name=true);
+
+CREATE OR REPLACE VIEW device_map AS
+-- profiles store the device pool under the legacy "anonymousIds" key
+-- (buildIdentityMap in lib/verify/identity.js reads the same field)
+SELECT unnest("anonymousIds") AS device_id, distinct_id FROM users;
+
+CREATE OR REPLACE VIEW ev AS
+-- ::VARCHAR casts — user_id sniffs as UUID, device_id as VARCHAR; DuckDB
+-- refuses to coalesce mixed types
+SELECT coalesce(m.distinct_id::VARCHAR, e.user_id::VARCHAR, e.device_id::VARCHAR) AS uid,
+       e.time::TIMESTAMP AS t,
+       e.*
+FROM read_json_auto('data/verify-gaming-EVENTS*.json', sample_size=-1, union_by_name=true) e
+LEFT JOIN device_map m ON e.device_id = m.device_id;
+
+
+-- ── H1-compass-heavy-rewards ────────────────────────────────
+-- heavy compass users (2+ "use item" w/ Ancient Compass, ~48%) earn ~1.5x
+-- reward_gold and reward_xp on quest turned in (measured 1.56x / 1.58x —
+-- bonus-quest clones are built at the boosted rate)
+WITH cnt AS (
+  SELECT u.distinct_id AS uid,
+    count(e.uid) FILTER (WHERE e.event = 'use item' AND e.item_type = 'Ancient Compass') AS n
+  FROM users u LEFT JOIN ev e ON e.uid = u.distinct_id GROUP BY 1
+)
+SELECT CASE WHEN c.n >= 2 THEN 'heavy' ELSE 'light' END AS cohort,
+  count(DISTINCT c.uid) AS users,
+  round(avg(e.reward_gold), 1) AS avg_gold,
+  round(avg(e.reward_xp), 1) AS avg_xp
+FROM ev e JOIN cnt c ON e.uid = c.uid
+WHERE e.event = 'quest turned in'
+GROUP BY 1 ORDER BY 1;
+
+
+-- ── H2-cursed-week ──────────────────────────────────────────
+-- injected Curse deaths cluster in days 40-47 of each user's own timeline.
+-- Measured on 48d+ lifetime users: per-day density in/out ≈ 36x; raw in/out
+-- count ≈ 2.7 (out-of-window span is ~10x longer)
+WITH firsts AS (SELECT uid, min(t) AS t0, max(t) AS tN FROM ev GROUP BY 1),
+longlife AS (SELECT uid, t0, tN FROM firsts WHERE tN >= t0 + INTERVAL 48 DAY),
+curse AS (
+  SELECT e.uid, epoch(e.t - l.t0) / 86400.0 AS dol
+  FROM ev e JOIN longlife l ON e.uid = l.uid
+  WHERE e.event = 'player death' AND e.cause_of_death = 'Curse'
+)
+SELECT
+  count(*) FILTER (WHERE dol BETWEEN 40 AND 47) AS in_window,
+  count(*) FILTER (WHERE dol < 40 OR dol > 47) AS out_window,
+  (SELECT count(*) FROM longlife) AS longlife_users,
+  round((count(*) FILTER (WHERE dol BETWEEN 40 AND 47) / 8.0)
+        / nullif(count(*) FILTER (WHERE dol < 40 OR dol > 47)
+                 / ((SELECT avg(epoch(tN - t0)) / 86400.0 FROM longlife) - 8), 0), 1) AS density_ratio
+FROM curse;
+
+
+-- ── H3-guild-rescue ─────────────────────────────────────────
+-- among users with 3+ week-1 deaths: early guild joiners (first 3 days) are
+-- exempt from the spiral → ~2.5x post-week-1 volume vs spiraled peers
+-- (knob ceiling 1/0.3 ≈ 3.3x; both cohorts select front-loaded players)
+WITH firsts AS (SELECT uid, min(t) AS t0 FROM ev GROUP BY 1),
+flags AS (
+  SELECT f.uid,
+    count(e.uid) FILTER (WHERE e.event = 'player death' AND e.t < f.t0 + INTERVAL 7 DAY) AS early_deaths,
+    count(e.uid) FILTER (WHERE e.event = 'guild joined' AND e.t < f.t0 + INTERVAL 3 DAY) AS early_guild,
+    count(e.uid) FILTER (WHERE e.t >= f.t0 + INTERVAL 7 DAY) AS post_events
+  FROM firsts f JOIN ev e ON e.uid = f.uid GROUP BY 1
+)
+SELECT CASE WHEN early_guild > 0 THEN 'guild_saved' ELSE 'spiral' END AS cohort,
+  count(*) AS users, round(avg(post_events), 1) AS avg_post
+FROM flags WHERE early_deaths >= 3 GROUP BY 1 ORDER BY 1;
+
+
+-- ── H4-death-spiral ─────────────────────────────────────────
+-- 3+ week-1 deaths and no early guild → 70% of post-week-1 events dropped.
+-- spiral post/pre event ratio ~0.11 vs healthy; spiral cohort ~11% of the
+-- guild-free population
+WITH firsts AS (SELECT uid, min(t) AS t0 FROM ev GROUP BY 1),
+flags AS (
+  SELECT f.uid,
+    count(e.uid) FILTER (WHERE e.event = 'player death' AND e.t < f.t0 + INTERVAL 7 DAY) AS early_deaths,
+    count(e.uid) FILTER (WHERE e.event = 'guild joined' AND e.t < f.t0 + INTERVAL 3 DAY) AS early_guild,
+    count(e.uid) FILTER (WHERE e.t >= f.t0 + INTERVAL 7 DAY) AS post_events,
+    count(e.uid) FILTER (WHERE e.t < f.t0 + INTERVAL 7 DAY) AS pre_events
+  FROM firsts f JOIN ev e ON e.uid = f.uid GROUP BY 1
+)
+SELECT CASE WHEN early_guild = 0 AND early_deaths >= 3 THEN 'spiral'
+            WHEN early_guild > 0 THEN 'guild' ELSE 'other' END AS cohort,
+  count(*) AS users,
+  round(avg(post_events)::DOUBLE / nullif(avg(pre_events), 0), 3) AS post_pre
+FROM flags GROUP BY 1 ORDER BY 1;
+
+
+-- ── H5-lucky-charm ──────────────────────────────────────────
+-- Lucky Charm Pack buyers (~9% of spenders) pay 2.5x on the pack itself and
+-- get 35% bonus-purchase clones → avg spend ~3.4x non-buyers. Scoped to
+-- NON-whales so H10's 1.8x multiplier doesn't contaminate the ratio.
+WITH lucky AS (SELECT DISTINCT uid FROM ev WHERE event = 'real money purchase' AND product = 'Lucky Charm Pack')
+SELECT
+  round(avg(price_usd) FILTER (WHERE uid IN (SELECT uid FROM lucky)), 2) AS lucky_avg,
+  round(avg(price_usd) FILTER (WHERE uid NOT IN (SELECT uid FROM lucky)), 2) AS nonlucky_avg,
+  count(DISTINCT uid) FILTER (WHERE uid IN (SELECT uid FROM lucky)) AS lucky_users
+FROM ev WHERE event = 'real money purchase' AND price_usd IS NOT NULL
+  AND substr(uid, 1, 1) NOT IN ('0','3','6','9','c','f');
+
+
+-- ── H6-strategic-explorers ──────────────────────────────────
+-- strategic explorers (6+ inspect AND 6+ search for clues, ~42% of users):
+-- 85% dungeon completion vs ~54% baseline, 2x treasure (measured 1.7x blend)
+WITH cnt AS (
+  SELECT u.distinct_id AS uid,
+    count(e.uid) FILTER (WHERE e.event = 'inspect') AS ins,
+    count(e.uid) FILTER (WHERE e.event = 'search for clues') AS sea
+  FROM users u LEFT JOIN ev e ON e.uid = u.distinct_id GROUP BY 1
+),
+coh AS (SELECT uid, (ins >= 6 AND sea >= 6) AS strategic FROM cnt)
+SELECT c.strategic, count(DISTINCT c.uid) AS users,
+  round(avg(e.treasure_value) FILTER (WHERE e.event = 'find treasure'), 1) AS avg_treasure,
+  round((count(*) FILTER (WHERE e.event = 'exit dungeon' AND e.completion_status = 'completed'))::DOUBLE
+        / nullif(count(*) FILTER (WHERE e.event = 'exit dungeon'), 0), 3) AS completion
+FROM coh c JOIN ev e ON e.uid = c.uid GROUP BY 1 ORDER BY 1;
+
+
+-- ── H7-shadowmourne ─────────────────────────────────────────
+-- Shadowmourne Legendary drops only after dataset day 45 (2% roll on find
+-- treasure); zero pre-release sightings; owners win ~96% of combats vs ~55%
+WITH owners AS (SELECT DISTINCT uid FROM ev WHERE event = 'find treasure' AND treasure_type = 'Shadowmourne Legendary'),
+combat AS (
+  SELECT uid, count(*) FILTER (WHERE outcome = 'Victory')::DOUBLE / count(*) AS win_rate
+  FROM ev WHERE event = 'combat completed' GROUP BY 1
+)
+SELECT
+  (SELECT count(*) FROM ev WHERE event = 'find treasure' AND treasure_type = 'Shadowmourne Legendary'
+     AND t < (SELECT min(t) FROM ev) + INTERVAL 45 DAY) AS pre_release_drops,
+  (SELECT count(*) FROM owners) AS owners,
+  (SELECT round(avg(win_rate), 3) FROM combat WHERE uid IN (SELECT uid FROM owners)) AS owner_win,
+  (SELECT round(avg(win_rate), 3) FROM combat WHERE uid NOT IN (SELECT uid FROM owners)) AS rest_win;
+
+
+-- ── H8-subscriber-tiers ─────────────────────────────────────
+-- Elite: 1.8x quest gold, completion flip → ~0.90; Premium: 1.4x gold,
+-- completion → ~0.82; Free baseline ~0.68 (completion = f + (1-f)·flip)
+SELECT subscription_tier, count(DISTINCT uid) AS users,
+  round(avg(reward_gold) FILTER (WHERE event = 'quest turned in'), 1) AS avg_gold,
+  round((count(*) FILTER (WHERE event = 'exit dungeon' AND completion_status = 'completed'))::DOUBLE
+        / nullif(count(*) FILTER (WHERE event = 'exit dungeon'), 0), 3) AS completion
+FROM ev GROUP BY 1 ORDER BY 1;
+
+
+-- ── H9-level-gold-scaling ───────────────────────────────────
+-- reward_gold scales with profile level: mult = 1 + 0.15·(level-1).
+-- formula check: hi/lo gold ratio should equal
+-- (1 + 0.15·(mean_hi - 1)) / (1 + 0.15·(mean_lo - 1)) — measured 1.95 vs 1.94
+WITH bucketed AS (
+  SELECT CASE WHEN u.level <= 5 THEN 'lo_1_5' WHEN u.level >= 16 THEN 'hi_16_20' ELSE 'mid' END AS bucket,
+    u.level, e.reward_gold
+  FROM users u JOIN ev e ON e.uid = u.distinct_id AND e.event = 'quest turned in'
+)
+SELECT bucket, count(*) AS quests, round(avg(level), 2) AS mean_level,
+  round(avg(reward_gold), 1) AS avg_gold,
+  round((1 + 0.15 * (avg(level) - 1)), 3) AS predicted_mult
+FROM bucketed GROUP BY 1 ORDER BY 1;
+
+
+-- ── H10-whale-purchases ─────────────────────────────────────
+-- whales = distinct_id first hex char in ('0','3','6','9','c','f') → 37.5%
+-- of ids nominally, ~63% of spend rows; 1.8x price_usd. Scoped to NON-lucky
+-- buyers so H5's multiplier doesn't contaminate the ratio.
+WITH lucky AS (SELECT DISTINCT uid FROM ev WHERE event = 'real money purchase' AND product = 'Lucky Charm Pack')
+SELECT
+  round(avg(price_usd) FILTER (WHERE substr(uid, 1, 1) IN ('0','3','6','9','c','f')), 2) AS whale_avg,
+  round(avg(price_usd) FILTER (WHERE substr(uid, 1, 1) NOT IN ('0','3','6','9','c','f')), 2) AS rest_avg,
+  count(DISTINCT uid) FILTER (WHERE substr(uid, 1, 1) IN ('0','3','6','9','c','f')) AS whale_users,
+  count(DISTINCT uid) FILTER (WHERE substr(uid, 1, 1) NOT IN ('0','3','6','9','c','f')) AS rest_users
+FROM ev WHERE event = 'real money purchase' AND price_usd IS NOT NULL
+  AND uid NOT IN (SELECT uid FROM lucky);
+
+
+-- ── H11-alignment-archetype ─────────────────────────────────
+-- deterministic profile mapping: Good alignments → hero, Evil → villain,
+-- rest → neutral. Counts must match EXACTLY; shares ≈ 26/25/49
+SELECT
+  count(*) FILTER (WHERE alignment IN ('Lawful Good', 'Neutral Good')) AS good_aligns,
+  count(*) FILTER (WHERE archetype = 'hero') AS heroes,
+  count(*) FILTER (WHERE alignment IN ('Chaotic Evil', 'Neutral Evil')) AS evil_aligns,
+  count(*) FILTER (WHERE archetype = 'villain') AS villains,
+  count(*) FILTER (WHERE archetype = 'neutral') AS neutrals,
+  count(*) AS total
+FROM users;
+
+
+-- ── H12-combat-ttc-by-tier ──────────────────────────────────
+-- funnel-post compresses combat initiated → combat completed → use item by
+-- tier (Elite 0.30x, Premium 0.70x, Free 1.40x). The story uses the funnel
+-- emulator (median TTC, 6h window); this cross-event MIN→MIN approximation
+-- shows the same ordering for eyeball purposes.
+WITH ci AS (SELECT uid, min(t) AS t0 FROM ev WHERE event = 'combat initiated' GROUP BY 1),
+ui AS (
+  SELECT e.uid, min(e.t) AS t1
+  FROM ev e JOIN ci ON e.uid = ci.uid AND e.t >= ci.t0
+  WHERE e.event = 'use item' GROUP BY 1
+)
+SELECT u.subscription_tier, count(*) AS converters,
+  round(median(epoch(t1 - t0)) / 60, 1) AS median_ttc_min
+FROM ci JOIN ui USING (uid) JOIN users u ON u.distinct_id = ci.uid
+GROUP BY 1 ORDER BY median_ttc_min;
+
+
+-- ── H13-prep-magic-number ───────────────────────────────────
+-- prep actions (inspect + search for clues) between first quest accepted and
+-- first fight boss: sweet band 3-6 → 1.3x treasure (measured 1.24x scoped to
+-- non-strategic users); over-prepared 7+ → boss win flipped to ~0.73x sweet
+WITH anchors AS (
+  SELECT uid,
+    min(t) FILTER (WHERE event = 'quest accepted') AS qa,
+    min(t) FILTER (WHERE event = 'fight boss') AS fb
+  FROM ev GROUP BY 1
+),
+prep AS (
+  SELECT a.uid, count(e.uid) AS n
+  FROM anchors a LEFT JOIN ev e ON e.uid = a.uid
+    AND e.event IN ('inspect', 'search for clues') AND e.t > a.qa AND e.t < a.fb
+  WHERE a.qa IS NOT NULL AND a.fb IS NOT NULL AND a.fb > a.qa
+  GROUP BY 1
+),
+bands AS (SELECT uid, CASE WHEN n BETWEEN 3 AND 6 THEN 'sweet' WHEN n < 3 THEN 'low' ELSE 'over' END AS band FROM prep)
+SELECT b.band, count(DISTINCT b.uid) AS users,
+  round(avg(e.treasure_value) FILTER (WHERE e.event = 'find treasure'), 1) AS avg_treasure,
+  round((count(*) FILTER (WHERE e.event = 'fight boss' AND e.victory = true))::DOUBLE
+        / nullif(count(*) FILTER (WHERE e.event = 'fight boss'), 0), 3) AS boss_win
+FROM bands b JOIN ev e ON e.uid = b.uid
+GROUP BY 1 ORDER BY 1;

package/dungeons/vertical/gaming/gaming.verify.mjs

@@ -0,0 +1,70 @@
+#!/usr/bin/env node
+/**
+ * gaming.verify.mjs — thin wrapper around the story runner.
+ *
+ * All verification logic lives in the `stories` export of ./gaming.js;
+ * this script just streams the shards and delegates. It is equivalent to:
+ *
+ *   node scripts/verify-stories.mjs dungeons/vertical/gaming/gaming.js --data-prefix verify-gaming
+ *
+ * Generate first:
+ *   node scripts/verify-runner.mjs dungeons/vertical/gaming/gaming.js verify-gaming
+ * Run:
+ *   node dungeons/vertical/gaming/gaming.verify.mjs [prefix]   # default verify-gaming
+ */
+import fs from 'node:fs';
+import path from 'node:path';
+import readline from 'node:readline';
+import { execFile } from 'node:child_process';
+import { promisify } from 'node:util';
+import { buildIdentityMap, evaluateStories, VERDICT_RANK } from '@ak--47/dungeon-master/verify';
+import config, { stories } from './gaming.js';
+
+const PREFIX = process.argv[2] || 'verify-gaming';
+const prefixPath = path.join('data', PREFIX);
+
+async function loadShards(suffix) {
+	// streaming load: full-fidelity EVENTS shards exceed the readFileSync cap
+	const dir = path.dirname(prefixPath), base = path.basename(prefixPath);
+	const out = [];
+	if (!fs.existsSync(dir)) return out;
+	for (const f of fs.readdirSync(dir).filter(f => f.startsWith(`${base}-${suffix}`) && f.endsWith('.json')).sort()) {
+		const rl = readline.createInterface({ input: fs.createReadStream(path.join(dir, f)), crlfDelay: Infinity });
+		for await (const line of rl) {
+			if (line.trim()) out.push(JSON.parse(line));
+		}
+	}
+	return out;
+}
+
+const events = await loadShards('EVENTS');
+const profiles = await loadShards('USERS');
+if (!events.length) {
+	console.error(`no shards at ${prefixPath}-EVENTS*.json — run: node scripts/verify-runner.mjs dungeons/vertical/gaming/gaming.js ${PREFIX}`);
+	process.exit(1);
+}
+console.log(`gaming — events=${events.length} users=${profiles.length} (${prefixPath})`);
+
+const execFileP = promisify(execFile);
+const runSql = async (sql) => {
+	const { stdout } = await execFileP('duckdb', ['-json', '-c', sql.replaceAll('{{PREFIX}}', prefixPath)], { maxBuffer: 64 * 1024 * 1024 });
+	return stdout.trim() ? JSON.parse(stdout) : [];
+};
+
+// funnels passed raw (unvalidated) — every timeToConvert story carries an
+// explicit conversionWindowMs, so no funnel-default threading is needed
+const results = await evaluateStories(stories, events, {
+	profiles,
+	funnels: config.funnels,
+	identityMap: buildIdentityMap(profiles),
+	runSql,
+});
+
+let worst = 'NAILED';
+for (const r of results) {
+	console.log(`${r.verdict.padEnd(7)} ${r.id}`);
+	for (const a of r.assertions) console.log(`        ${a.verdict.padEnd(7)} ${a.detail}`);
+	if (VERDICT_RANK[r.verdict] < VERDICT_RANK[worst]) worst = r.verdict;
+}
+console.log(`\nworst verdict: ${worst}`);
+process.exit(VERDICT_RANK[worst] >= VERDICT_RANK.STRONG ? 0 : 1);