{"id": 10384, "prediction_verbatim": "SMH underperforms SPY over 24h [DIRECTION: down] \u2014 tariff confirmation (764175) lands harder on spot capex pricing than Hugging Face narrative lift; nearterm supply-chain repricing dominates structural AI story. [FALSIFY: SMH outperforms SPY (closes +0.3% or more vs SPY's close) over 24h window]", "created_at": "2026-09-04 04:09:57", "direction": "down", "timeframe": "24h", "score": 0.21, "outcome": "Wrong \u2014 SMH +2.6% vs SPY -0.4% \u2014 SMH beat SPY by 3.0%", "resolved_at": "2026-09-08 04:31:31", "chain_hash": "fa8384854e305edf139f75e25d96bceee21bb4bee1f3a54f3dd043ff1b9bde67", "chain_tx": "2ko7XdVBKuLezKUa3EkMKXTynHiD4YVKfc6ZPSE2aYTbK7amfggrXicJAmh8DCqMZuqhE7vkiXRnrT41UPo5vSKi", "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."}