Live Numbers, Not a Pitch

This page pulls straight from the same detection engine that feeds every customer's alerts. No backtest, no cherry-picked window — the numbers below update every few minutes as the agent scans the live market.

What this is not: a claim that a 30-day soak has been "validated." We're publishing this as a continuous, honest window into detection quality while that validation period runs. If a number below looks thin, it's because the sample is thin right now — not because we've hidden a better one.

Median net edge, per detected opportunity

Net edge is the spread after both platforms' fees — the number that decides whether an opportunity is real arbitrage at all. This is a rate, not a dollar amount; see "realizable profit" below for what it's worth at actual order-book depth.

Not enough data in the current window yet.

Realizable profit per opportunity

Net edge multiplied by how many contracts the thinner side of the order book can actually absorb — what a customer could realistically capture, not the edge on an infinite order size.

Not enough data in the current window yet.

Days until resolution

How long the capital sits before the arbitrage settles — the shorter leg wins, since the position closes the moment either market resolves. This is why net edge alone doesn't tell you the annualized return.

Not enough data in the current window yet.

Right now

Live in the feed

—

Last 24 hours

—

Hours with zero detections

— / 24h

Inventory rotation, last 7 days

Whether new pairs show up daily, or the feed is the same handful of markets on repeat. "New" means it hasn't appeared in the previous 6 days — a pair silent for a week is fresh inventory for capital that's been sitting idle.

Distinct pairs, 7 days

—

New, last 24h

—

Recurring, last 24h

—

30-day window

Accumulated dry-run P&L (simulated fills, not live trading), how many of the last 30 calendar days had at least one detection, and matcher precision from every pair a human has actually reviewed — approved on review, versus rejected onto the blocklist.

Dry-run P&L, 30d

—

simulated fills only — never live capital

Days with a detection

— / 30

Matcher precision

—

no manual reviews yet

What this maps to for a typical, non-selective customer

—

At the median edge and median holding period above, computed from the live numbers on this page. Customers who are selective about which alerts they act on may see more or less, depending on the spreads and holding periods they pick. Full math in how much you can make.

Learn more: