Methodology

How UMI reads the market

Transparent, rule-based intelligence over public market data. Every read is reactive — it explains what the data is doing, never predicts, and never asserts causality.

1 · Principles

Reactive, rule-based market intelligence.
No investment advice.
Macro and event causality require editor confirmation.
Public APIs expose curated, redacted DTOs only.

2 · Source hierarchy

Price truthMulti-venue spot composite (Binance/Coinbase/Kraken/OKX), outlier-filtered

Market contextBreadth, derivatives positioning, macro (FRED), on-chain (DeFiLlama)

Crowd oddsPrediction markets (Polymarket, Manifold, Kalshi)

3 · How signals are graded

Each engine is scored out-of-sample (walk-forward Information Coefficient + Deflated Sharpe) and fused by measured skill — so the read says which signals actually predict, not just what they assert.

4 · Daily cycle

A cycle closes at 09:00 GST. Reads are reactive to the data available at cutoff; nothing auto-publishes — causality and editorial framing require human review.

5 · Limits & disclaimers

Automated signals from public data. Not a forecast and not investment advice. Market-implied probabilities and price reactions are reported as facts; causality is never asserted. Covered assets: BTC, ETH. Methodology version: umi-v1-scaffold.

Automated, rule-based market intelligence derived from public data sources. Reactive signals, not a forecast and not investment advice. Market-implied probabilities and price reactions are reported as facts; causality is never asserted.