CLV Audit

Where the number we own came from

One CLV figure answers “did the market come to us?” and hides the more useful question: did we get a good price because we waited, or because the market moved onto our side afterwards? We have prices at three moments — the first quote we ever saw (open), the price we froze (lock), and the last price before tip (close) — so we can split it, and the split is exact: waiting + drift = the raw move from our price to the close.

In plain words. Waiting paid: how much cheaper the price we locked was than the first price on the board — a positive number means not betting the opener saved us that much, in probability points. Market came to us: after we locked, how much worse the same bet got for anyone taking it later — that is closing-line value, and positive means we own the better number. Market drift: which way the price moved across the whole day whatever we did — negative means the market walked away from our side, so our number came from timing, not from the market agreeing with us. The three reconcile by construction: waiting + drift = market-came-to-us, exactly, on these rows.

Why three of these columns say “raw”. A quoted price contains the book’s margin; a fair probability doesn’t. We freeze both sides when we lock and the close is two-sided, so close-vs-lock can be de-vigged — that is the de-vigged CLV column, and it is the number we stand behind. We never captured the opening quote for the other side, so anything involving the open cannot be de-vigged. That matters and it isn’t neutral: the opener is thin and juiced, so part of “waiting paid” is the juice compressing rather than the number moving — real money we didn’t pay, but not evidence that the market agrees with us. The last column shows the same rows measured both ways so you can see the size of that correction, and we never average a raw figure and a de-vigged one together.