Honest, independent calibration of prediction market alpha. We publish what didn't work as carefully as what did.
Point-in-time test of whether 'follow informed wallets' actually works in the Polymarket cheap band. The result: our wave-34 gate produced a 9% hit rate. The wallets it filtered OUT hit 54%. We dropped the gate.
Our flagship backtest looked great at +76% per trade. Then we applied fees, split it by category, and checked temporal stability. Here's the version we can defend.
Our live signal detector fires on whale BUYs at price 0.20–0.30. In the last 14 days, the entire Polymarket platform saw 16 such trades — and zero in the last 48 hours. A structural finding about why the strategy is rare to test.
Every closed stop-out in our paper portfolios over the last two weeks. Did the 50% stop save us money or destroy reversible losses? The answer: saved $6.4K, cost zero.
Audits are written from live data on ohh.bet. The receipts for every number cited live on /transparency.
RSS