Price Gap Detection

Finding the Edge

The gap detection engine is where simulation meets execution. After MiroFish produces its probability estimate for a market, the system compares it against the current market price on Kalshi.

The 15% Threshold

Not every divergence is tradeable. Small gaps can be explained by bid-ask spreads, transaction costs, and normal market noise. AI Predicted Wins only triggers a trade when the gap between the AI estimate and the market price exceeds 15%.

This conservative threshold ensures:

• Statistical significance — A 15% gap is unlikely to be explained by random noise alone
• Margin of safety — Even if the AI estimate is slightly off, there’s still a profitable edge
• Transaction cost coverage — Kalshi’s fee structure is accounted for in the minimum gap requirement

Gap Analysis in Practice

Consider a Kalshi market priced at $0.35 (35% implied probability). If MiroFish estimates the true probability at 55%, that’s a 20% gap — above the 15% threshold. The system would buy contracts at $0.35, expecting them to resolve closer to $0.55.

Conversely, if a market is priced at $0.70 and MiroFish estimates 50%, the system would take the opposite side — selling (or buying “No”) at the inflated price.

Continuous Re-evaluation

Gap detection isn’t a one-shot process. As market prices move and new information becomes available, the system continuously re-evaluates open positions. If a gap narrows below the threshold after a trade is placed, the position is flagged for review.