Methodology

The edge model

Edge is the distance between what a contract is worth and what the market is charging for it. We estimate the first, measure the second honestly, and show you the gap.

Fair value, then mid, then the gap

Most "signals" stop at a number on a chart. The edge model is deliberately simpler and more honest: a model fair value, a depth-aware market mid, and the difference between them. Everything else is plumbing in service of those three quantities.

01

Estimate fair value

Each contract gets a model probability built from the data that actually drives it — order flow, the crypto and equity underlyings it tracks, and historical resolution behaviour.

02

Read the market mid

We take the real, depth-aware mid from the live order book — not the last print — so the comparison reflects where you could actually transact.

03

Edge = fair value − mid

The gap between the two, expressed in price and in expected value per contract. Positive edge means the market is paying you to take the model side.

04

Size against the book

Edge is only tradable if the book can absorb it. We net edge against available depth and slippage so the number you see is the number you can realistically capture.

What the edge model is not

  • Not a price target. Fair value is a probability estimate with error bars, not a promise of where the contract settles.
  • Not a backtested win rate. We don't publish a single accuracy figure, because one number across thousands of heterogeneous markets would be more misleading than useful.
  • Not free money. Edge net of depth and slippage is frequently small or zero. The model's job is to tell you when it isn't.

Where the model is least confident

Thin books, ambiguous resolution criteria, and long-dated contracts all widen the error on fair value. Rather than hide that, the model surfaces a resolution-risk haircut and flags low-depth markets so you can discount the edge yourself. The terminal shows the inputs, not just the verdict.