Prediction markets are an unusual kind of signal — instead of measuring something that already happened, they price what a community of bettors collectively expects to happen. Palanor treats them as first-class signals on the lattice, the same way it treats FRED economic series or composite Indices.
How markets reach your signal library
Two paths.
The first is **automatic discovery**. A daily cron walks Kalshi's open events catalog — typically a thousand or more on any given day — and drops the obvious noise (sports games, entertainment microbets, celebrity casting decisions). The survivors get scored by Numen for stewardship relevance on a 0-to-1 scale. The top fifteen by score get promoted into the signal catalog with an attribution noting they were Numen-discovered. Markets whose close-time has passed retire automatically; new ones flow in on the next run.
The second is **user-added**. On the Signals page, the + Add a prediction market link opens a page where you paste a Kalshi market URL. Palanor parses the ticker, validates the market exists, pulls its current price, and adds it to your organization's signal lattice with you marked as the adder. Your org is auto-subscribed so it appears in your Weave immediately.
What Numen scores for
The relevance scoring rubric is explicit:
1.0 Critical macro instrument — Fed decisions, CPI, recession calls, major GDP or employment data, US elections, debt ceiling 0.8 Major capital markets or regulatory — Treasury yields, equity index thresholds, FDA approvals, major IPOs, central bank policy 0.6 Strategic geopolitical or structural — major elections worldwide, climate-goal deadlines, energy transition, AGI/AI milestones 0.4 Sector-relevant — specific company IPOs, CEO succession at major institutions, regulatory rulings 0.2 Tangential — long-horizon technology or exploration timing 0.0 Irrelevant — sports outcomes, celebrity gossip, entertainment casting, game scores
Each scored market also gets a one-sentence rationale stored alongside it, so the curation is legible — you can see why Numen elevated each one.
What happens to the prediction-market values
Each market's price is the implied YES probability, normalized to the 0-to-1 range. The existing kalshi-sync cron pulls fresh prices hourly for every active market on the catalog. The probability lands in signal_observations the same way any other signal value does, which means prediction markets participate in scenario probability scoring, schema scoring, Lattice Canvas topology, and any future Index that wants to incorporate forward belief on named events.
Why Kalshi specifically
Kalshi is US-regulated by the CFTC, focused on macro and economic markets, exposes a public read-only API, and settles in USD. It's the right fit for an enterprise-stewardship lens. Polymarket can be added as a second source adapter later without rewriting the data model — the source field on signals_catalog already accommodates both.
A note on liquidity
Discovery scores markets for relevance, not for liquidity. A high-scoring market with no trading activity will show in your library but won't have observations until someone trades it. We surface the score and the close-time so you can decide whether to subscribe; longer-horizon markets (closing in 2028, 2030, 2050) often have thin order books.
Promotion is liquidity-weighted: discovery only promotes markets with open positions and a usable price. The composite score blends relevance (75 percent weight) with a log-scaled liquidity boost (25 percent), so a relevance-0.7 market with 10,000 open interest can beat a relevance-0.9 market with zero. Markets without a usable price are dropped from promotion entirely.