Market Intelligence

Market fitness scoring, resolution settlement, event calendar, and universe management.

Market Intelligence helps the LLM decide which markets to trade, when to enter, and when to exit - the research layer that feeds the autonomous decision loop.

Market Fitness Scoring

Before deploying a strategy on a market, assess whether it’s a good fit.

Scoring Dimensions

Dimension	What It Measures	Source
Liquidity	Orderbook depth, spread, daily volume	Exchange orderbook + trade data
Edge	Mispricing vs. LLM forecast	LLM signal engine + oracle
Strategy Fit	Is this market suitable for MM / directional / arb?	Microstructure analysis
Risk	Correlation with portfolio, tail risk	Position data + correlation matrix
Time Value	Time to resolution vs. capital lockup cost	Resolution calendar + yield curve

# Rank markets by fitness
scores = market_scorer.rank(
    markets=discovered_markets,
    strategy_type="market_maker",
    portfolio=fund.positions(),
    top_n=10,
)
# [{"market": "will-x-win", "score": 0.87, "liquidity": 0.92, "edge": 0.75, ...}, ...]

Market Resolution Settlement

Prediction markets have a unique lifecycle: they resolve to 0 or 1. This system automates end-of-life handling.

Settlement Flow

1. Market Active

   Strategy is trading the market normally.

2. Resolution Detected

   Exchange confirms the market has resolved to an outcome.

3. P&L Booked

   Position settles at 0 or 1. Realized P&L recorded.

4. Capital Freed

   Released capital returns to the fund pool for reallocation.

5. Post-Analysis

   Was the edge real? Update models for similar future markets.

Step	What Happens
Resolution Detection	Poll exchange for resolved markets
P&L Booking	Position settles at 0 or 1, realized P&L recorded
Capital Recycling	Freed capital returns to fund pool
Attribution	Resolution P&L tracked separately from trading P&L
Post-Analysis	Was our edge real? Update models for future markets

Near-Resolution Handling

Markets approaching resolution need special treatment:

• High confidence: hold to resolution (maximize edge capture)
• Low confidence: exit early (reduce variance)
• Liquidity drying up: exit before orderbook disappears
• Cancel open orders: prevent getting filled at bad prices near resolution

Event Calendar Intelligence

Track catalysts that move markets and use them for timing.

# Upcoming catalysts
catalysts = calendar.upcoming(days=7)
# [
#   {"event": "FOMC Decision", "date": "2026-03-18", "markets_affected": 12},
#   {"event": "CPI Release", "date": "2026-03-20", "markets_affected": 8},
# ]

# Score markets by catalyst proximity
scored = calendar.catalyst_score(markets)

Market Universe Management

The set of markets the fund actively tracks and considers for trading.

Universe Lifecycle

1. All Markets

   Full set of markets across all connected exchanges.

2. Discovery Filter

   Apply minimum liquidity, volume, and fitness score thresholds.

3. Active Universe

   Markets the fund is actively tracking and considering for strategies.

4. Strategy Assignment

   Assign the best-fit strategy template to each market.

5. Trading

   Deploy and run strategies. Remove markets that resolve, lose liquidity, or get restricted.

Action	Trigger
Add market	New market discovered with fitness score above threshold
Remove market	Market resolved, liquidity dropped, or compliance restricted
Promote market	Fitness score improved, assign strategy
Demote market	Fitness score declined, reduce exposure

# Configure market universe
universe = MarketUniverse(
    exchanges=["polymarket", "kalshi"],
    min_liquidity=1000,
    min_volume_24h=5000,
    excluded_categories=["adult"],
    max_markets=100,
)
universe.refresh()  # Scan and update

MCP Tools

Tool	Description
scan_opportunities	Scan all markets, return ranked by fitness
research_market	Deep analysis of a single market
resolution_status	Check resolution status of active markets
market_universe	Current universe with fitness scores