feat(ms-ai-architect): Fase 0 steg 3 — base-rate-rapport + GATE: BYGG Fase 3 (scoped/hybrid) [skip-docs]

compute-base-rate.mjs (tynn CLI over testet lib/base-rate.mjs) -> base-rate-report.{json,md}. 373 paastander: verifiserbar feilrate 13,4 % (40/299), Wilson [10,0 %, 17,7 %]; unsourced 19,8 %. staleness-recall 0/40 = 0 % -> korrekthets-judge er eneste mekanisme som fanger feilene. Feil konsentrert i fetchbare typer (sku 36 %, version 25 %, tpm 20 %); pris 74 % unsourced -> operatoer-gated. GATE besluttet i plan-doc: BYGG Fase 3 scoped til fetchbare claim_types (taxonomy/status/version/tpm/sku/region); pris ikke maskinverifiserbar.
2026-06-26 17:15:54 +02:00 · 2026-06-26 17:15:54 +02:00 · 49fd18c6d2
commit 49fd18c6d2
parent 6985ed1f5d
4 changed files with 512 additions and 0 deletions
--- a/docs/ref-kb-workflow-plan-2026-06.md
+++ b/docs/ref-kb-workflow-plan-2026-06.md
@ -65,6 +65,17 @@ _Fase 0-verifiseringen er bevisst utsatt til en uthvilt sesjon (retnings-avgjør

 **Verifisering (testbart, fra planen + locked):** ≥45 volatile-filer merket; Wilson-bånd beregnet; gull-sett + sample-frame lagret som artefakter; eksplisitt tall for «feil uten lastmod-endring»; begge nye scripts har failing-test-først (Iron Law).

+### Fase 0 → GATE (BESLUTTET 2026-06-26): **BYGG Fase 3 — scoped/hybrid**
+
+Rapport: `scripts/kb-eval/data/base-rate-report.{json,md}` (deterministisk fra `gold-correctness-set.json`, 373 påstander, via testet `lib/base-rate.mjs`). Beslutningen følger gate-kriteriet (over) + per-claim_type-konsentrasjonen:
+
+1. **Feilraten klarer baren.** Verifiserbar feilrate **13,4 % (40/299)**, Wilson 95 % **[10,0 %, 17,7 %]**. Nedre bånd 10 % ≫ «~few %» ⇒ «lav feilrate ⇒ hopp over Fase 3»-grenen er utelukket med 95 % konfidens.
+2. **Det billige alternativet fanger ingenting.** Staleness-recall = **0/40 = 0 %**. Ingen av de 40 reelle feilene har `lastmod_changed=true` (false=37, null=3) — de er bakt inn ved fil-skriving, eller MS Learn sitemap_lastmod er for grovt til å fange endringen. En korrekthets-judge er ENESTE automatiske mekanisme som fanger dem. **CAVEAT løst:** registry `last_poll=2026-06-23` er ferskere enn 32/40 fildatoer (ville fanget en post-fildato sitemap-endring om den fantes); de 8 feilene i filer dat. 2026-06-24 er ~2 dager gamle ⇒ kan ikke ha drevet ennå. `true=0` er reelt signal, ikke stale-registry-artefakt.
+3. **Feilene konsentreres der en fetch-judge virker.** Per claim_type: **sku 36,4 % (8/22), version 25,0 % (8/32), tpm 20,0 % (5/25)** — alle nær fullt hentbare (sku 0 unsourced, version 1, tpm 5). Høyt utbytte + høy rekkevidde.
+4. **Pris er ute av scope — av data, ikke antakelse.** 76 pris-påstander, **56 unsourced (74 %)**. «0/20 = 0 %» pris-feilrate er en **falsk null** (74 % kunne ikke sjekkes — ikke at de er korrekte). JS-rendrede Azure-prissider beseirer `microsoft_docs_fetch` ⇒ en fetch-basert judge når dem heller ikke. Pris er 56 av 74 unsourced (76 %); resten av korpuset er 18/297 unsourced (**6 %**) ⇒ den uverifiserbare massen er **isolert til pris**, ikke spredt.
+
+**Scope for Fase 3-judgen:** kun fetchbare claim_types — `taxonomy|status|version|tpm|sku|region` (297 påstander, 94 % hentbare, bærer 100 % av de verifiserbare feilene). `claim_type=price` flagges «ikke maskinverifiserbar» ⇒ **operatør-gated, ikke judge-gated**. Forutsetningene i Fase 3-seksjonen (autoritets-backfill, full frontmatter/Fase 2, judge-kalibrering mot dette gull-settet før tallene stoles på) gjelder uendret før judgen tas i bruk. Gull-settet (373 påstander) er kalibreringssettet.
+
 ---

 ## Fase 1 — Trygge, retnings-uavhengige grep (kan kjøres parallelt med Fase 0)
--- a/scripts/kb-eval/compute-base-rate.mjs
+++ b/scripts/kb-eval/compute-base-rate.mjs
@ -0,0 +1,102 @@
+#!/usr/bin/env node
+// compute-base-rate.mjs — Fase 0, steg 3 glue: turn the gold correctness set into
+// a defensible base-rate report (the directional input for the Fase 3 gate).
+//
+// All non-trivial math (verifiable error rate, Wilson bands, the staleness-
+// catchable vs judge-unique split) lives in tested lib/base-rate.mjs. This CLI is
+// thin wiring: read gold-correctness-set.json -> computeBaseRate() -> write a
+// machine-readable .json and a human-readable .md. Same shape as build-gold-set.mjs.
+//
+// Usage:  node scripts/kb-eval/compute-base-rate.mjs [--write]
+//   (default: print the overall + per-stratum summary; --write persists
+//    data/base-rate-report.json and data/base-rate-report.md)
+
+import fs from 'node:fs';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { computeBaseRate } from './lib/base-rate.mjs';
+
+const __dirname = path.dirname(fileURLToPath(import.meta.url));
+const DATA = path.join(__dirname, 'data');
+
+const gold = JSON.parse(fs.readFileSync(path.join(DATA, 'gold-correctness-set.json'), 'utf8'));
+const claims = gold.claims;
+const report = computeBaseRate(claims);
+
+const pct = (x) => `${(x * 100).toFixed(1)}%`;
+const band = (w) => `[${pct(w.low)}, ${pct(w.high)}]`;
+
+// One row of the per-dimension tables. b = a finalized bucket from the lib.
+function row(key, b) {
+  const v = b.byVerdict;
+  return `| ${key} | ${b.total} | ${v.correct} | ${v.outdated} | ${v.wrong} | ${v.unsourced} | ${b.errors}/${b.verifiable} = ${pct(b.errorRate)} | ${band(b.errorRateWilson)} | ${b.errorsJudgeUnique} |`;
+}
+const HEAD =
+  '| key | n | correct | outdated | wrong | unsourced | err-rate (verifiable) | Wilson 95% | judge-unique |\n' +
+  '|---|---|---|---|---|---|---|---|---|';
+
+// Sort dimension entries by descending total so the heaviest buckets read first.
+function table(title, byKey) {
+  const rows = Object.entries(byKey)
+    .sort((a, b) => b[1].total - a[1].total)
+    .map(([k, b]) => row(k, b));
+  return `### ${title}\n\n${HEAD}\n${rows.join('\n')}\n`;
+}
+
+const o = report.overall;
+const md = `# Base-rate-rapport — Fase 0 (KB korrekthet)
+
+_Generert deterministisk av \`compute-base-rate.mjs\` over \`gold-correctness-set.json\`. Tall fra testet \`lib/base-rate.mjs\` (15 tester). Ikke rediger for hånd — regenerer._
+
+**Gull-sett:** ${claims.length} påstander · metode: ${gold._meta?.method ? 'se gold-correctness-set.json `_meta.method`' : 'n/a'}
+
+## Verdict-vokabular
+
+- **correct** — en hentet learn.microsoft.com-side oppgir den påståtte verdien
+- **outdated** — hentet kilde viser en annen, erstattet verdi (tidsdrift)
+- **wrong** — hentet kilde motsier påstanden; den var aldri korrekt
+- **unsourced** — ingen hentbar MS Learn-side oppgir verdien (kan ikke verifiseres)
+
+«Reelle feil» = outdated + wrong. **unsourced er IKKE en feil** — det er den uverifiserbare massen (priser på JS-rendrede Azure-sider som en fetch-basert judge heller ikke når). Den verifiserbare feilraten ekskluderer derfor unsourced fra nevneren.
+
+## Overall
+
+| metrikk | verdi |
+|---|---|
+| Påstander totalt | ${o.total} |
+| correct / outdated / wrong / unsourced | ${o.byVerdict.correct} / ${o.byVerdict.outdated} / ${o.byVerdict.wrong} / ${o.byVerdict.unsourced} |
+| Reelle feil (outdated+wrong) | ${o.errors} |
+| Verifiserbare påstander (nevner) | ${o.verifiable} |
+| **Verifiserbar feilrate** | **${o.errors}/${o.verifiable} = ${pct(o.errorRate)}** |
+| Wilson 95 % | ${band(o.errorRateWilson)} |
+| Unsourced-andel | ${o.unsourced}/${o.total} = ${pct(o.unsourcedRate)} |
+| Feil staleness-loopen fanger (lastmod_changed=true) | ${o.errorsStalenessCatchable} |
+| **Feil kun en korrekthets-judge fanger (judge-unique)** | **${o.errorsJudgeUnique}** |
+
+> **Gate-kritisk:** \`errorsJudgeUnique\` = reelle feil hvis siterte kilde-lastmod IKKE endret seg etter fildato — den eneste klassen en korrekthets-judge fanger som den eksisterende staleness-loopen bommer på. Staleness-recall på de reelle feilene = ${o.errors ? `${o.errorsStalenessCatchable}/${o.errors} = ${pct(o.errorsStalenessCatchable / o.errors)}` : 'n/a'}.
+
+## Per stratum
+
+${table('Stratum', report.byStratum)}
+## Per skill
+
+${table('Skill', report.bySkill)}
+## Per claim_type
+
+${table('Claim type', report.byClaimType)}
+`;
+
+if (process.argv.includes('--write')) {
+  const jsonOut = path.join(DATA, 'base-rate-report.json');
+  const mdOut = path.join(DATA, 'base-rate-report.md');
+  fs.writeFileSync(jsonOut, JSON.stringify({ _meta: { source: 'gold-correctness-set.json', claim_count: claims.length }, ...report }, null, 2) + '\n');
+  fs.writeFileSync(mdOut, md);
+  console.log(`wrote ${jsonOut}`);
+  console.log(`wrote ${mdOut}`);
+} else {
+  console.log(`claims: ${o.total} | correct=${o.byVerdict.correct} outdated=${o.byVerdict.outdated} wrong=${o.byVerdict.wrong} unsourced=${o.byVerdict.unsourced}`);
+  console.log(`verifiable error rate: ${o.errors}/${o.verifiable} = ${pct(o.errorRate)}  Wilson ${band(o.errorRateWilson)}`);
+  console.log(`unsourced: ${o.unsourced}/${o.total} = ${pct(o.unsourcedRate)}`);
+  console.log(`staleness-catchable=${o.errorsStalenessCatchable}  judge-unique=${o.errorsJudgeUnique}`);
+  console.log('(dry run — pass --write to persist base-rate-report.json + .md)');
+}
--- a/scripts/kb-eval/data/base-rate-report.json
+++ b/scripts/kb-eval/data/base-rate-report.json
@ -0,0 +1,333 @@
+{
+  "_meta": {
+    "source": "gold-correctness-set.json",
+    "claim_count": 373
+  },
+  "overall": {
+    "total": 373,
+    "byVerdict": {
+      "correct": 259,
+      "outdated": 29,
+      "wrong": 11,
+      "unsourced": 74
+    },
+    "errors": 40,
+    "verifiable": 299,
+    "unsourced": 74,
+    "errorsStalenessCatchable": 0,
+    "errorsJudgeUnique": 40,
+    "errorRate": 0.13377926421404682,
+    "errorRateWilson": {
+      "p": 0.13377926421404682,
+      "low": 0.0998039899711332,
+      "high": 0.17704568986149216
+    },
+    "unsourcedRate": 0.19839142091152814
+  },
+  "bySkill": {
+    "ms-ai-advisor": {
+      "total": 79,
+      "byVerdict": {
+        "correct": 59,
+        "outdated": 7,
+        "wrong": 2,
+        "unsourced": 11
+      },
+      "errors": 9,
+      "verifiable": 68,
+      "unsourced": 11,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 9,
+      "errorRate": 0.1323529411764706,
+      "errorRateWilson": {
+        "p": 0.1323529411764706,
+        "low": 0.07122163958543326,
+        "high": 0.2328027696704853
+      },
+      "unsourcedRate": 0.13924050632911392
+    },
+    "ms-ai-engineering": {
+      "total": 86,
+      "byVerdict": {
+        "correct": 54,
+        "outdated": 9,
+        "wrong": 4,
+        "unsourced": 19
+      },
+      "errors": 13,
+      "verifiable": 67,
+      "unsourced": 19,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 13,
+      "errorRate": 0.19402985074626866,
+      "errorRateWilson": {
+        "p": 0.19402985074626866,
+        "low": 0.11705063857457099,
+        "high": 0.3041933762415822
+      },
+      "unsourcedRate": 0.22093023255813954
+    },
+    "ms-ai-governance": {
+      "total": 76,
+      "byVerdict": {
+        "correct": 52,
+        "outdated": 5,
+        "wrong": 2,
+        "unsourced": 17
+      },
+      "errors": 7,
+      "verifiable": 59,
+      "unsourced": 17,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 7,
+      "errorRate": 0.11864406779661017,
+      "errorRateWilson": {
+        "p": 0.11864406779661017,
+        "low": 0.058675082779140006,
+        "high": 0.22523875773415053
+      },
+      "unsourcedRate": 0.2236842105263158
+    },
+    "ms-ai-infrastructure": {
+      "total": 36,
+      "byVerdict": {
+        "correct": 29,
+        "outdated": 1,
+        "wrong": 2,
+        "unsourced": 4
+      },
+      "errors": 3,
+      "verifiable": 32,
+      "unsourced": 4,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 3,
+      "errorRate": 0.09375,
+      "errorRateWilson": {
+        "p": 0.09375,
+        "low": 0.032400962626319516,
+        "high": 0.24218499335778831
+      },
+      "unsourcedRate": 0.1111111111111111
+    },
+    "ms-ai-security": {
+      "total": 96,
+      "byVerdict": {
+        "correct": 65,
+        "outdated": 7,
+        "wrong": 1,
+        "unsourced": 23
+      },
+      "errors": 8,
+      "verifiable": 73,
+      "unsourced": 23,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 8,
+      "errorRate": 0.1095890410958904,
+      "errorRateWilson": {
+        "p": 0.1095890410958904,
+        "low": 0.0565860416101191,
+        "high": 0.20162825897706282
+      },
+      "unsourcedRate": 0.23958333333333334
+    }
+  },
+  "byStratum": {
+    "volatile": {
+      "total": 331,
+      "byVerdict": {
+        "correct": 222,
+        "outdated": 28,
+        "wrong": 10,
+        "unsourced": 71
+      },
+      "errors": 38,
+      "verifiable": 260,
+      "unsourced": 71,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 38,
+      "errorRate": 0.14615384615384616,
+      "errorRateWilson": {
+        "p": 0.14615384615384616,
+        "low": 0.1083692375274728,
+        "high": 0.1942426326249217
+      },
+      "unsourcedRate": 0.21450151057401812
+    },
+    "control": {
+      "total": 42,
+      "byVerdict": {
+        "correct": 37,
+        "outdated": 1,
+        "wrong": 1,
+        "unsourced": 3
+      },
+      "errors": 2,
+      "verifiable": 39,
+      "unsourced": 3,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 2,
+      "errorRate": 0.05128205128205128,
+      "errorRateWilson": {
+        "p": 0.05128205128205128,
+        "low": 0.014177657646399527,
+        "high": 0.1688593904563791
+      },
+      "unsourcedRate": 0.07142857142857142
+    }
+  },
+  "byClaimType": {
+    "version": {
+      "total": 33,
+      "byVerdict": {
+        "correct": 24,
+        "outdated": 6,
+        "wrong": 2,
+        "unsourced": 1
+      },
+      "errors": 8,
+      "verifiable": 32,
+      "unsourced": 1,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 8,
+      "errorRate": 0.25,
+      "errorRateWilson": {
+        "p": 0.25,
+        "low": 0.13252243982621553,
+        "high": 0.4210689177024662
+      },
+      "unsourcedRate": 0.030303030303030304
+    },
+    "tpm": {
+      "total": 30,
+      "byVerdict": {
+        "correct": 20,
+        "outdated": 4,
+        "wrong": 1,
+        "unsourced": 5
+      },
+      "errors": 5,
+      "verifiable": 25,
+      "unsourced": 5,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 5,
+      "errorRate": 0.2,
+      "errorRateWilson": {
+        "p": 0.2,
+        "low": 0.08860454100652485,
+        "high": 0.3913133553653825
+      },
+      "unsourcedRate": 0.16666666666666666
+    },
+    "region": {
+      "total": 20,
+      "byVerdict": {
+        "correct": 15,
+        "outdated": 2,
+        "wrong": 0,
+        "unsourced": 3
+      },
+      "errors": 2,
+      "verifiable": 17,
+      "unsourced": 3,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 2,
+      "errorRate": 0.11764705882352941,
+      "errorRateWilson": {
+        "p": 0.11764705882352941,
+        "low": 0.03287908001292092,
+        "high": 0.3433684249770991
+      },
+      "unsourcedRate": 0.15
+    },
+    "status": {
+      "total": 70,
+      "byVerdict": {
+        "correct": 58,
+        "outdated": 4,
+        "wrong": 3,
+        "unsourced": 5
+      },
+      "errors": 7,
+      "verifiable": 65,
+      "unsourced": 5,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 7,
+      "errorRate": 0.1076923076923077,
+      "errorRateWilson": {
+        "p": 0.1076923076923077,
+        "low": 0.05315354431925606,
+        "high": 0.20601533031468616
+      },
+      "unsourcedRate": 0.07142857142857142
+    },
+    "price": {
+      "total": 76,
+      "byVerdict": {
+        "correct": 20,
+        "outdated": 0,
+        "wrong": 0,
+        "unsourced": 56
+      },
+      "errors": 0,
+      "verifiable": 20,
+      "unsourced": 56,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 0,
+      "errorRate": 0,
+      "errorRateWilson": {
+        "p": 0,
+        "low": 0,
+        "high": 0.16113012549493322
+      },
+      "unsourcedRate": 0.7368421052631579
+    },
+    "taxonomy": {
+      "total": 122,
+      "byVerdict": {
+        "correct": 108,
+        "outdated": 7,
+        "wrong": 3,
+        "unsourced": 4
+      },
+      "errors": 10,
+      "verifiable": 118,
+      "unsourced": 4,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 10,
+      "errorRate": 0.0847457627118644,
+      "errorRateWilson": {
+        "p": 0.0847457627118644,
+        "low": 0.046682191106169606,
+        "high": 0.14899481904471476
+      },
+      "unsourcedRate": 0.03278688524590164
+    },
+    "sku": {
+      "total": 22,
+      "byVerdict": {
+        "correct": 14,
+        "outdated": 6,
+        "wrong": 2,
+        "unsourced": 0
+      },
+      "errors": 8,
+      "verifiable": 22,
+      "unsourced": 0,
+      "errorsStalenessCatchable": 0,
+      "errorsJudgeUnique": 8,
+      "errorRate": 0.36363636363636365,
+      "errorRateWilson": {
+        "p": 0.36363636363636365,
+        "low": 0.19732972772607899,
+        "high": 0.5704865065628195
+      },
+      "unsourcedRate": 0
+    }
+  },
+  "_verdicts": [
+    "correct",
+    "outdated",
+    "wrong",
+    "unsourced"
+  ]
+}
--- a/scripts/kb-eval/data/base-rate-report.md
+++ b/scripts/kb-eval/data/base-rate-report.md
@ -0,0 +1,66 @@
+# Base-rate-rapport — Fase 0 (KB korrekthet)
+
+_Generert deterministisk av `compute-base-rate.mjs` over `gold-correctness-set.json`. Tall fra testet `lib/base-rate.mjs` (15 tester). Ikke rediger for hånd — regenerer._
+
+**Gull-sett:** 373 påstander · metode: se gold-correctness-set.json `_meta.method`
+
+## Verdict-vokabular
+
+- **correct** — en hentet learn.microsoft.com-side oppgir den påståtte verdien
+- **outdated** — hentet kilde viser en annen, erstattet verdi (tidsdrift)
+- **wrong** — hentet kilde motsier påstanden; den var aldri korrekt
+- **unsourced** — ingen hentbar MS Learn-side oppgir verdien (kan ikke verifiseres)
+
+«Reelle feil» = outdated + wrong. **unsourced er IKKE en feil** — det er den uverifiserbare massen (priser på JS-rendrede Azure-sider som en fetch-basert judge heller ikke når). Den verifiserbare feilraten ekskluderer derfor unsourced fra nevneren.
+
+## Overall
+
+| metrikk | verdi |
+|---|---|
+| Påstander totalt | 373 |
+| correct / outdated / wrong / unsourced | 259 / 29 / 11 / 74 |
+| Reelle feil (outdated+wrong) | 40 |
+| Verifiserbare påstander (nevner) | 299 |
+| **Verifiserbar feilrate** | **40/299 = 13.4%** |
+| Wilson 95 % | [10.0%, 17.7%] |
+| Unsourced-andel | 74/373 = 19.8% |
+| Feil staleness-loopen fanger (lastmod_changed=true) | 0 |
+| **Feil kun en korrekthets-judge fanger (judge-unique)** | **40** |
+
+> **Gate-kritisk:** `errorsJudgeUnique` = reelle feil hvis siterte kilde-lastmod IKKE endret seg etter fildato — den eneste klassen en korrekthets-judge fanger som den eksisterende staleness-loopen bommer på. Staleness-recall på de reelle feilene = 0/40 = 0.0%.
+
+## Per stratum
+
+### Stratum
+
+| key | n | correct | outdated | wrong | unsourced | err-rate (verifiable) | Wilson 95% | judge-unique |
+|---|---|---|---|---|---|---|---|---|
+| volatile | 331 | 222 | 28 | 10 | 71 | 38/260 = 14.6% | [10.8%, 19.4%] | 38 |
+| control | 42 | 37 | 1 | 1 | 3 | 2/39 = 5.1% | [1.4%, 16.9%] | 2 |
+
+## Per skill
+
+### Skill
+
+| key | n | correct | outdated | wrong | unsourced | err-rate (verifiable) | Wilson 95% | judge-unique |
+|---|---|---|---|---|---|---|---|---|
+| ms-ai-security | 96 | 65 | 7 | 1 | 23 | 8/73 = 11.0% | [5.7%, 20.2%] | 8 |
+| ms-ai-engineering | 86 | 54 | 9 | 4 | 19 | 13/67 = 19.4% | [11.7%, 30.4%] | 13 |
+| ms-ai-advisor | 79 | 59 | 7 | 2 | 11 | 9/68 = 13.2% | [7.1%, 23.3%] | 9 |
+| ms-ai-governance | 76 | 52 | 5 | 2 | 17 | 7/59 = 11.9% | [5.9%, 22.5%] | 7 |
+| ms-ai-infrastructure | 36 | 29 | 1 | 2 | 4 | 3/32 = 9.4% | [3.2%, 24.2%] | 3 |
+
+## Per claim_type
+
+### Claim type
+
+| key | n | correct | outdated | wrong | unsourced | err-rate (verifiable) | Wilson 95% | judge-unique |
+|---|---|---|---|---|---|---|---|---|
+| taxonomy | 122 | 108 | 7 | 3 | 4 | 10/118 = 8.5% | [4.7%, 14.9%] | 10 |
+| price | 76 | 20 | 0 | 0 | 56 | 0/20 = 0.0% | [0.0%, 16.1%] | 0 |
+| status | 70 | 58 | 4 | 3 | 5 | 7/65 = 10.8% | [5.3%, 20.6%] | 7 |
+| version | 33 | 24 | 6 | 2 | 1 | 8/32 = 25.0% | [13.3%, 42.1%] | 8 |
+| tpm | 30 | 20 | 4 | 1 | 5 | 5/25 = 20.0% | [8.9%, 39.1%] | 5 |
+| sku | 22 | 14 | 6 | 2 | 0 | 8/22 = 36.4% | [19.7%, 57.0%] | 8 |
+| region | 20 | 15 | 2 | 0 | 3 | 2/17 = 11.8% | [3.3%, 34.3%] | 2 |
+