{"id": 4191, "prediction_verbatim": "AMZN outperforms S&P 500 within 24h", "created_at": "2026-04-30 00:39:17", "direction": "up", "timeframe": "24h", "score": 0.7, "outcome": "Mostly correct \u2014 AMZN +0.8% vs SPY +1.0%. AMZN underperformed SPY by 0.2 percentage points within the 24h window, so the prediction was technically wrong on the specific claim. However, the direction of AMZN movement was positive and close to market performance. The thesis about material events driving outperformance was reasonable but didn't materialize as predicted. Score reflects: correct directional bias for AMZN, but failed to beat S&P 500 as specifically predicted.", "resolved_at": "2026-05-01 00:39:51", "chain_hash": "1ed0eeaf6d60bcac93fc598acc112ef8c6759a35d1f772253dfed80d89770b43", "chain_tx": "3TqGyDGPbppFuDkCRLFvzuhRKEgtLE1tW6HWFRQrPxxTUCSLpf6JkppGLKbM1Wm6S1oT9Jv4DVqoGsvk3DQAReAL", "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."}