Systematic Edge Detection
Edge detection is what separates systematic trading from casual speculation. Instead of browsing markets and making gut-feel decisions, you define your view through a thesis, compute implied prices, and let the system find mismatches.
The Edge Detection Pipeline
- Thesis computation: Your causal tree produces thesis-implied prices for the root event
- Market mapping: The system identifies all contracts related to your thesis (across Kalshi and Polymarket)
- Price comparison: Current market prices are compared against thesis-implied prices
- Friction adjustment: Spread, fees, and estimated slippage are subtracted from theoretical edge
- Ranking: Opportunities are ranked by executable edge × liquidity score
Cross-Venue Edge Detection
One of the most powerful forms of edge detection is cross-venue — finding the same event priced differently on Kalshi vs. Polymarket. If Kalshi prices a recession at 28% and Polymarket at 35%, there's at least a 7-point arbitrage opportunity (minus cross-venue friction).
Edge Decay
Edges are perishable. The typical lifecycle:
- New information creates a mispricing (minutes)
- Active traders spot and start trading it (minutes to hours)
- Price converges toward fair value (hours to days)
- Edge disappears completely
The heartbeat's frequency determines how quickly you catch edges. Faster heartbeats catch shorter-lived opportunities.
Filters and Thresholds
Not every edge is worth trading. The sf edges command supports filters:
--min-edge 5: Only show edges above 5 points--min-liquidity B: Only show markets with B-grade or better liquidity--category economics: Only show economic contracts