{"id": 2147, "prediction_verbatim": "META and AMZN maintain outperformance vs TSLA and NVDA by >1% in next 48h; software/platform names hold gains while capex-heavy and tariff-exposed names lag further", "created_at": "2026-04-02 20:43:01", "direction": "up", "timeframe": "48h", "score": 0.7, "outcome": "META and AMZN (-0.8%, -0.4%) underperformed TSLA and NVDA (-5.4%, +0.9%) by *far* more than 1%. Thus the main comparative claim failed badly.  Software/platform names (META, AMZN) did NOT hold gains relative to capex-heavy/tariff-exposed names (TSLA, NVDA). However, the secondary claim about platform holding gains was false, so this is *mostly* incorrect. However, given the drastic move in TSLA, it's possible the outperformance claim has some merit despite the META/AMZN underperformance.", "resolved_at": "2026-04-04 20:54:58", "chain_hash": "526fcdf4affeea6dc9eb3ee54f01ec6836e24ae8163929f947e5cabd9d101f1a", "chain_tx": "2LV9TgdEwKVkXQY52WUgggAnTSKr62vHmVbbwPFNHcwNHWtbynmQKUdBDL4Tmn4Mrryt6Na5s8TJTrMime229pnA", "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."}