{
 "_FLOW": "A ledger of the archive's registered predictions and falsification conditions. TWO LAYERS. Layer 1 is a systematic extraction: 451 conditions harvested from the 326 deposits that carry a NAMED falsification section, classified by how each would resolve. Layer 2 is a hand-verified set of predictions whose outcome is known, weighted deliberately toward the ones that FAILED. A prediction ledger whose entries all came true would be evidence of a ledger written after the fact.",
 "dataset": "prediction-ledger",
 "version": "v1.0",
 "date": "2026-08-23",
 "method": "Structural, not semantic. A registered prediction is distinguished by carrying a RESOLUTION CONDITION, and resolution conditions live in named sections. Regex on prediction language returns 939 candidates across 415 deposits, most of them conditional prose. Regex on falsification-section HEADINGS returns 326 deposits and 359 sections, from which 451 individual conditions were extracted. The section is the discriminator.",
 "layer_1": {
  "deposits_with_falsification_section": 326,
  "sections": 359,
  "conditions_extracted": 451,
  "by_resolution_route": {
   "VERIFIABLE — resolves by checking a surface": 200,
   "EMPIRICAL — resolves on a study": 20,
   "PHILOLOGICAL — resolves against a source": 40,
   "UNCLASSIFIED — resolution route not stated": 164,
   "FORMAL — resolves by construction or proof": 18,
   "TEMPORAL — resolves on a dated re-measurement": 9
  },
  "findings": [
   "200 conditions are VERIFIABLE NOW — resolvable by checking a surface, a commit, a hash or a deposit. Anyone could resolve them today.",
   "164 conditions — more than a third — DO NOT STATE A RESOLUTION ROUTE. That is a defect class: a falsification condition that does not say how it would resolve is a gesture, not a condition. Naming it here is the first step to fixing it.",
   "Only 9 carry a clock. The archive registers many more conditions than it schedules."
  ]
 },
 "layer_2": {
  "hand_verified": 13,
  "by_status": {
   "RESOLVED — WRONG": 6,
   "RESOLVED — WRONG (withdrawn)": 1,
   "WITHDRAWN": 1,
   "OPEN": 5
  },
  "note": "SIX RESOLVED WRONG, ONE WITHDRAWN, FIVE OPEN. The failures are the point: a hypothesis that failed within an hour of being proposed, two pre-registered predictions falsified by the archive's own battery, a duplicate gate that returned a false negative, and three claims this archive made about its own capabilities that its own corpus refuted."
 },
 "the_credential": "An archive's registered predictions are worth exactly what its failed ones are worth. The most valuable rows here are RESOLVED — WRONG.",
 "coverage": "Layer 1 is a census of named falsification sections, not of all registered predictions — conditions stated in body prose without a section heading are not captured. Layer 2 is a curated set, not a sample.",
 "next": "Resolve the 200 verifiable conditions. Assign resolution routes to the 164 that lack one. Schedule the 9 that carry clocks.",
 "status_vocabulary": {
  "RESOLVED — WRONG": "The prediction was made, tested, and failed. Preserved.",
  "WITHDRAWN": "The claim was retracted by its author before or without testing.",
  "NARROWED — NOT WITHDRAWN": "The observation stands; its evidential weight was reduced by later findings. ADDED 2026-08-23 because the ledger had no way to record the difference between an observation and its interpretation, and conflating them would let explanation erase fact.",
  "OPEN": "Registered, not yet resolved.",
  "SUSPENDED — SUBSTRATE LOST": "The claim stands but its antecedent no longer obtains, because the infrastructure it presupposed was removed. Cannot be confirmed, falsified, or counted either way. ADDED 2026-08-23 on discovering that predictions registered under DOI-anchored infrastructure became untestable when the DOIs were severed five weeks later."
 }
}