No vibes. No screenshots. No survivor bias. Every public X stock pick, backtested at $1,000 per call against publicly available prices.
Pull every $TICKER cashtag tweet from the picker's
timeline. X API only — no scraping, no DMs, no private groups.
Each tweet goes through Claude Haiku 4.5 — bullish, bearish, or neutral — with a 2–6 word snippet capturing the call.
$1,000 long on every bullish call. $1,000 short on every bearish call. Marked to today's close. Real receipts, hypothetical money.
A 0–100 composite of four weighted signals — keeps the leaderboard honest.
Share of directional calls that finished green. A 70% win rate at random scale is the cleanest skill signal we have.
How much of the picker's history sits in a single ticker. A 100-call track record on one stock is a coin-flip dressed as a portfolio.
Anchored on X account creation date (or oldest call when unavailable). New accounts get penalised; survivorship bias is real.
Share of total simulated P&L driven by the single best ticker. Catching one 10-bagger isn't the same as a repeatable process.
Each profile shows which signals fired and the full rationale. 0–33 is low risk, 34–66 medium, 67–100 high.
Transparency is the product. Here's what's out of scope on purpose.
Equities only. Crypto tickers are pulled from the archive but filtered from every user-facing surface. Options flow is not in scope.
The simulation is gross of all real-world friction. Real execution would not match these numbers.
Only public X posts. Nothing scraped from gated platforms or paid signal services.
Past performance — real or simulated — does not predict future results. Ledger is informational only and is not financial advice.
$200/month. Cancel anytime from the Stripe portal.