Most people think "arbitrage betting" means finding two books that disagree by accident. That's not really where the money is anymore — books are faster than that now.

The real edge is in lag between market tiers. Institutional/exchange-style markets (the ones sharp syndicates and pro traders operate on) react to information almost instantly — injury news, weather, sharp money, lineup changes. Retail sportsbooks, by contrast, are optimized for volume and UX, not speed. Their trading desks re-price on a delay, especially on lower-liquidity markets (player props, secondary markets, smaller leagues).

That delay is measurable. When you diff a retail line against the sharp consensus in real time, the gap between them is the +EV signal — no guessing, no "feel," just a statistical mispricing that closes itself out within minutes as the retail book catches up.

A few things we've learned building infrastructure around this:

1. Speed beats sophistication. A dumb diff calculated in under a second beats a brilliant model calculated in 30 seconds. The edge decays fast.

2. Headless > dashboards. If a human has to look at a screen to catch the signal, you've already lost the trade. Anything worth acting on should hit your system via webhook, not a UI.

3. Line count matters more than book count. Coverage across more markets (not just more sportsbooks) is what actually surfaces volume of opportunities — most operators over-index on adding books and under-index on adding markets.

We've been heads-down building exactly this — a real-time data layer that does the diffing and pushes the edge straight to your infrastructure. If you're already trading this space and want to compare notes on methodology, happy to talk shop in the comments.