$ qubo learn signals/04

Reading Lens reliability scores

Every Lens analysis on a panel of assets is scored after the fact: did the price follow the main scenario, an alternative one, or none? Over time this builds a track record.

A hit rate is meaningful only with its sample size. 7 hits out of 10 says little; 140 out of 200 says more. Qubo publishes reliability only once enough analyses have been scored.

Compare the hit rate with the stated probabilities. If scenarios given 60% come true about 60% of the time, the probabilities are well calibrated. If they come true 40% of the time, they are overconfident.

Past reliability does not guarantee future reliability. Markets change, and a tool that worked in one regime can do worse in another.

// rule to remember

// your turn

Open the Lens reliability page and note the number of scored analyses and the hit rate of the main scenario.

Educational content only. Not investment advice. Past data does not predict future results. Risk disclosure