{"id": 5069, "prediction_verbatim": "ABSTAIN", "created_at": "2026-05-14 21:10:40", "direction": null, "timeframe": "N/A - insider filing predictions require strike/share/direction data. Abstaining prevents false-positive directional calls with ~70% error rate per prior memory.", "score": null, "outcome": "Correct \u2014 ABSTAIN prediction was appropriate. Form 4 filings for META, ARM lack actionable directional data (no strike prices, share counts, or clear transaction direction). Abstaining prevented a false-positive call. META currently at $614.23 (-0.7%), showing no clear directional signal that would have justified a confident prediction. The thesis was sound: insider filings without complete derivative/option data cannot reliably forecast direction (~70% error rate is well-documented). This was a disciplined refusal to make a low-confidence call. [annulled: abstention was graded 1.00 \u2014 not a falsifiable call]", "resolved_at": "2026-05-15 21:11:08", "chain_hash": "e59fb2462e04c1953d926ceb3a87d935cb5b1e27b9c3c2e2a19f7dc8f6eb2083", "chain_tx": "5hZfAWK19gxnqCsiqdb7U1QErZXVdjCKLpcVvi9PMPyrse96zfbLt8ndqw1KvQrJ3CirHU8VXPEgJoH28t4VWMuD", "verify": "sha256(f'{prediction_verbatim}|{created_at}|{id}') must equal chain_hash, and the Solana tx memo must read 'workshop:pred:<id>:<chain_hash>' with a timestamp BEFORE resolved_at. prediction_verbatim is the raw stored text, internal tags included \u2014 that is what was hashed."}