S1 (a1): feature-promise audit — no claim survives unverified

PROMISE-AUDIT.md: every /pricing claim → verified/built/reworded.
BUILT (was vapor): alt line ladder + edge ranking (same-features regrade
at shifted lines, Desk-gated at the API), quarter-Kelly (engine quantile
P(win) x real captured odds — either missing → no sizing), free-tier
kill-condition locked previews. FIXED (was false): analyst 15/day cap vs
the Founder 'Unlimited reads' promise → analyst unlimited; every '40+
factors' claim (real count: 22 named features) reworded truthfully in 7
files. VERIFIED: cascade alerts (real, wired), phi correlation,
leg history, cross-book comparison, WC soccer, real-time feed.
Locked by tests/unit/promiseAudit.test.js. Jest now ignores
.claude/worktrees (parallel agents' suites no longer leak into runs).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Kev
2026-07-11 14:24:41 -04:00
parent e4d2e79f95
commit d242b11b4b
20 changed files with 296 additions and 30 deletions
+1 -1
View File
@@ -19,7 +19,7 @@ export default function AboutPage() {
<p className="mono" style={{ fontSize: 14.5, color: 'var(--text-1)', lineHeight: 1.7 }}>
VYNDR is prop intelligence a proprietary terminal that reads the context the books hide.
Beyond the box score: form, matchup, usage, rest, the cascade when a star sits. Then a single
grade, earned across 40+ factors, that tells you whether the number is worth taking.
grade, earned across every factor the model weighs, that tells you whether the number is worth taking.
</p>
<p className="mono" style={{ fontSize: 14.5, color: 'var(--text-1)', lineHeight: 1.7 }}>
We are an analytics tool, not a sportsbook. We don&apos;t take wagers. We hand you the read and
+1 -1
View File
@@ -5,7 +5,7 @@ import SectionHead from '@/components/vyndr/SectionHead';
import TerminalInput from '@/components/vyndr/TerminalInput';
const FAQ: { cat: string; q: string; a: string }[] = [
{ cat: 'GRADES', q: 'What does a VYNDR grade mean?', a: 'A grade is our confidence that a prop hits, weighed across 40+ factors. A+ is the rarest and strongest; D means stay away. The letter is earned, not opinion.' },
{ cat: 'GRADES', q: 'What does a VYNDR grade mean?', a: 'A grade is our confidence that a prop hits, weighed across every factor the model computes for that prop. A+ is the rarest and strongest; D means stay away. The letter is earned, not opinion.' },
{ cat: 'GRADES', q: 'What is a kill condition?', a: 'A red flag we detect that can sink a prop regardless of the stats — a blowout risk, a minutes cap, a brutal matchup. Analyst and Desk see them in full.' },
{ cat: 'READS', q: 'How many free reads do I get?', a: 'Five reads every calendar month on the Free tier. The counter resets at the start of each month.' },
{ cat: 'READS', q: 'What counts as a read?', a: 'Grading one prop. Viewing the same prop twice in a session only counts once.' },
+3 -1
View File
@@ -44,7 +44,8 @@ interface ScanResponse {
confidence?: number;
sample_size?: number;
factors?: Record<string, string>;
alt_lines?: { line: number; grade: string; hit_rate?: number; edge_pct?: number }[];
alt_lines?: { line: number; grade: string; hit_rate?: number; edge_pct?: number; base?: boolean }[];
kelly?: { pct: number; quarter: number; full: number; odds: string };
kill_conditions?: { code: string; reason: string }[];
reasoning?: string;
historical_hit_rate?: number;
@@ -733,6 +734,7 @@ export default function ScanPage() {
sample_size: result.sample_size,
factors: result.factors,
alt_lines: result.alt_lines,
kelly: result.kelly,
kill_conditions: result.kill_conditions,
tier,
// Session 44 — forward the engine's intel fields so the grade
+1 -1
View File
@@ -14,7 +14,7 @@ const PRICES: Record<Cadence, { display: string; sub: string; checkoutParam: str
const EARLY_FEATURES = [
'Unlimited reads',
'Full factor analysis (40+ signals)',
'Full factor analysis — every signal the model weighs',
'Kill conditions surfaced inline',
'Cascade alerts when lineups shift',
'Parlay leg history with grades',