Ledger / Prediction #3781 · token · 1H

Will $Claude AI still have sustained attention 1H from now (still trading and at least 2× its launch market cap, or graduated)?

GENESIS said 11% → FADES INCORRECT

What GENESIS knew, before

Launched 25m ago. X link, website. No hot term.

Made2026-10-14 09:30 UTC

Resolves2026-10-14 10:30 UTC

Evidence cutoff2026-10-14 09:29 UTC

Data sourcespumpfun.launch_feed

GENESIS version0.1.0

Enginetoken-heuristic-v0

Scoring methodtoken-sustained-v1

Active capabilities6

Evidence (exact record)

{
  "model": {
    "version": "token-heuristic-v0",
    "contributions": {
      "hotTerm": 0,
      "twitter": 0.4,
      "website": 0.3,
      "telegram": 0,
      "description": 0.2,
      "serialCreator": 0
    },
    "rawProbability": 0.11461549863791237
  },
  "features": {
    "mint": "DBJVEf9LzDDacYLNuKbEjfjMqhEhCfxFVgcNZkRpuaU8",
    "name": "Claude AI",
    "symbol": "Claude AI",
    "hotTerms": [],
    "ageMinutes": 25,
    "hasTwitter": true,
    "hasWebsite": true,
    "hasTelegram": false,
    "descriptionLength": 129,
    "marketCapAtObservation": 334169.3128694043,
    "creatorLaunchesInContext": 1
  },
  "selection": "uniform-random among launches aged 10-30 min",
  "candidates": 309,
  "calibration": null
}

Expected

{
  "direction": "fades",
  "probabilitySustained": 0.11461549863791237
}

Engine token-heuristic-v0

P(a token launched 10-30 minutes ago still has sustained attention at the horizon). Uses only what GENESIS saw at launch: social links, description, creator behaviour, whether it rides a hot term. Hand-set weights around a ~5% prior.

What happened, after

Graded2026-10-14 10:30 UTC

StatusSCORED

OutcomeSUSTAINED

ClassificationINCORRECT

Brier0.784

Log loss2.166

Graderjudge-v1

Outcome data

{
  "graduated": true,
  "replyCount": null,
  "lastTradeAt": "2026-10-14T09:13:42.000Z",
  "marketCapUsd": 1928.381160791251,
  "athMarketCapUsd": 340357.1786629401,
  "tradedInFinal15m": false,
  "launchMarketCapUsd": 334169.3128694043,
  "marketCapThresholdUsd": 668338.6257388087
}

Post-outcome analysis · judge-v1

Predicted 11% that $Claude AI would sustain attention for 1H. Outcome: SUSTAINED (market cap $1,928, graduated). INCORRECT.

— twitter was right but too weak on $Claude AI (1H). twitter contributed +0.40 toward yes, the true outcome, but could not overcome the prior.

— The prior was wrong on $Claude AI (1H). Every active signal pointed the right way; the base assumption outweighed them.

— Overconfident: 89% and wrong on $Claude AI. Confidence this high should be wrong about 1 time in 4 or less; track this bucket.

3 memory records written (sealed: written while memory was short-context only).

Integrity

Hash840164051968d006fdf2706a26d8915ca73ff2389b9f91ba9c27d03d1359f636

Previous hash32375f071ba0decd60a4a254c7ebff38300df6366087eecbf637434877762ffc