What Is CLV.gg? Inside the Real-Time Betting Intelligence Platform

Most betting tools describe a market that has already moved.

They poll the books every minute or two, assemble a table, and email a summary. By the time a bettor reads it, the sharp price that made a number interesting has corrected and the soft book that was slow to follow has caught up. The tool was accurate. It was just describing the past. In a market where the edge is a brief disagreement between books that resolves in seconds, accurate-but-late is the same as wrong.

CLV.gg is the platform we built at ixprt to solve that specific problem. It is a real-time sports-betting intelligence system: it reads the covered market at once, estimates what each outcome is worth, and surfaces the gaps between that estimate and the prices on offer, fast enough that the gaps are still open when a bettor sees them. This post is the plain-language tour: what it is, why the speed is the whole point, what it finds, and how deep the book coverage runs.

What CLV.gg actually does

Strip away the framing and the platform does three things in a loop, continuously, across every market it tracks.

First, it reads prices from every tracked book at once. Second, it converts those prices into a single estimate of each outcome's true probability, weighting the sharpest books most heavily and removing the bookmaker margin. Third, it compares that estimate against every price currently on offer and flags the ones that are better than they should be. The output is not a prediction about who wins the game. It is a measurement: here is what this outcome is worth by the model, here is a book offering more than that, take it before it moves.

That last clause is the hard part, and it is where most of the engineering lives.

Why speed is the product

An edge in betting is not a durable property of a market. It is a momentary disagreement between books, and disagreements get resolved.

When a sharp book moves its line, the slower books follow within seconds to minutes. During that window, a slow book is offering a price the market has already decided is wrong. That window is the edge. A tool that refreshes once a minute will frequently miss it entirely, and when it does catch one, it is often showing a bettor a price that no longer exists by the time they click through.

CLV.gg is built around that constraint. Prices arrive over live websocket connections rather than periodic polling, so the true-price estimate updates the instant a book moves, tick by tick. Detection runs in under a second, which is the difference between a signal a bettor can act on and a screenshot of something that already happened. The internal phrase for it is that detection latency is the product: finding an edge is only useful if you find it while it is still takeable.

The six signals it surfaces

Every opportunity CLV.gg flags is a version of the same thing, a price that is out of line with the fair value, but they take recognizably different shapes. The platform surfaces all six in one feed.

The first five are about deciding what to bet. The last, closing line value, is about confirming the process works, by grading every bet against where the market actually closed rather than against whether it won.

How deep the book coverage runs

A true-price estimate is only as good as the venues behind it, so coverage is not a vanity metric here. It is the foundation.

CLV.gg reads more than a dozen books and venues, and they are not treated as interchangeable. Sharp books, principally Pinnacle and Circa, run low-margin, high-limit models that welcome winning bettors, which is why their prices lead the market and serve as the anchors for the true-price estimate. Recreational sportsbooks, the large consumer brands most bettors know, are slower and shade their lines toward public money, which is exactly why they are where the takeable edges open. Betting exchanges like Betfair price by participants trading against each other rather than a bookmaker margin. And prediction markets such as Kalshi and Polymarket are treated as first-class venues for the events they cover, not afterthoughts.

The breadth matters in both directions. The sharp and exchange venues sharpen the estimate of what a price should be. The slower books are where reality lags that estimate long enough to bet. A platform that covered only one kind of venue would either have a weak anchor or nowhere to find the edge. Underneath all of it sits years of historical pricing, billions of price points, which is what keeps the consensus grounded in how these markets actually behave rather than a single live snapshot.

Where it stands

CLV.gg is not a roadmap. It is shipped and live in market at clv.gg, with real users across live markets every day. For anyone who wants to see the output before taking a word of this on faith, the public sample edges page shows real opportunities the detector caught on completed events, each with the books involved, the sharp-consensus read, and how the bet actually settled.

It is the live, in-market piece of the larger system ixprt builds: data in, structure on top, decisions out. CLV.gg is what that looks like when the decision has to be made in under a second, against a market that is moving while you watch it.