Prediction Market Edge Detection: How to Find Mispriced Contracts on Kalshi

## What Is an Edge? An edge is the difference between what you think a contract is worth and what the market is pricing it at. If you believe "US recession by 2026" has a 70% probability, and Kalshi prices it at 35¢ (35%), you have a +35¢ edge on YES. If your model is right, buying at 35¢ for something worth 70¢ is a good trade. The hard part isn't the math — it's systematically building a model, scanning markets, and accounting for execution costs. ## Step 1: Build Your Model You need a probability model for the events you're trading. SimpleFunctions does this with causal trees: sf create "Fed cuts rates below 4% by December 2026" The engine creates a tree of sub-claims (inflation trajectory, employment data, financial stress signals) and assigns probabilities. These propagate up to a thesis confidence that implies prices for related contracts. ## Step 2: Scan for Divergences sf edges Output: Ticker Mkt Thesis Edge Spread Liq Signal Gas .90 Dec 5¢ 60¢ +55¢ 6¢ low WATCH Fed cut Apr 5¢ 50¢ +46¢ - - CONSIDER WTI T180 Dec 20¢ 58¢ +38¢ 1¢ high CONSIDER KXRECSSNBER-26 35¢ 72¢ +37¢ 1¢ high CONSIDER CPI May >0.4% 30¢ 60¢ +30¢ 3¢ med CONSIDER **Theoretical edge** = thesis price - market price. But you can't trade theoretical edge. ## Step 3: Calculate Executable Edge Executable edge accounts for the bid-ask spread. If the edge is +38¢ but the spread is 10¢, your real edge is +28¢ after crossing the spread. sf liquidity --topic oil This shows you: - **Bid/Ask**: where you can actually buy and sell - **Spread**: your execution cost - **Depth**: how many contracts you can trade at these prices - **Slippage**: what happens when you try to fill 100+ contracts A 55¢ edge on a market with 6¢ spread and low depth might be worse than a 30¢ edge on a market with 1¢ spread and high depth. ## Step 4: Monitor and Update Edges aren't static. The heartbeat engine checks every 15 minutes: - **Price changes**: did the market move toward or away from your thesis? - **News**: did something happen that changes a causal node probability? - **Milestones**: is there an upcoming data release (CPI, unemployment, FOMC) that could resolve uncertainty? # What changed since yesterday? sf feed --hours 24 # Performance over time sf performance ## Common Mistakes **1. Ignoring liquidity.** A huge edge on a market with no depth is worthless. Always check `sf liquidity` before entering. **2. Overconcentrating.** Five oil contracts is one oil bet. Check `sf dashboard` for risk concentration. **3. Not stress-testing.** Run `sf whatif` before sizing up. How much do you lose if your thesis is wrong? **4. Confirmation bias.** The heartbeat engine evaluates all evidence, not just confirming evidence. But you should still actively seek disconfirming information. ## Getting Started npm install -g @spfunctions/cli sf setup sf create "your thesis" sf edges # find edges sf liquidity # check execution quality sf whatif # stress test Full documentation: [simplefunctions.dev/docs](https://simplefunctions.dev/docs)