The strict version: CLV compares the price you took against the closing price with the bookmaker's margin removed. A market closes with vig baked into both sides, so the honest benchmark is the de-vigged close, the market's implied fair probability at the moment betting stopped. Positive CLV means your number was better than that final fair price. Negative CLV means you paid more than the market ultimately said the bet was worth.
The CLV formula
CLV% = (decimal_taken / decimal_fair_close - 1) * 100decimal_taken is the decimal form of the odds you bet. decimal_fair_close is the decimal form of the closing price after de-vigging, meaning the bookmaker's margin has been stripped out so the two sides of the market sum to a probability of 100%. Skipping the de-vig step is the most common CLV mistake, and the worked example below shows how much it distorts the number.
Worked example: you take −105, it closes −120
The odds below are illustrative, chosen for clean arithmetic. They are not a recorded market. Suppose you bet a side at −105, and the market closes with your side at −120 and the other side at +100.
- Convert your price. −105 in decimal is 1 + 100/105 =
1.9524. - De-vig the close. Implied probabilities at close are 120/220 =
54.55%for your side and 100/200 =50.00%for the other. They sum to 104.55%; the extra 4.55 points are the book's margin. Normalizing, your side's fair closing probability is 54.55 / 104.55 =52.17%, which is fair decimal odds of 1 / 0.5217 =1.9167. - Score it. CLV% = (1.9524 / 1.9167 − 1) × 100 =
+1.86%.
You beat the fair close by 1.86 percent. Run the same numbers without the de-vig step, comparing −105 directly to the vigged −120 close, and the formula claims +6.49% (1.9524 / 1.8333). The difference is the bookmaker's margin, credited to you as if it were skill. Un-devigged CLV systematically flatters the bettor.
De-vigged CLV estimates the expected value of the bet
The formula is not arbitrary. Expected value per unit staked is p × d − 1, where p is the true win probability and d is your decimal odds. Take the fair close as the best available estimate of p, so that p = 1 / decimal_fair_close, and the expression becomes decimal_taken / decimal_fair_close − 1. That is the CLV formula. De-vigged CLV is an estimate of the bet's expected return in percent of stake, conditional on the close being the market's best probability estimate. In the example above, +1.86% CLV reads as roughly +1.86 cents of expectation per dollar staked at entry.
Why CLV reads skill faster than win rate
A win-loss record is a series of yes or no outcomes, and yes or no outcomes are extremely noisy. The arithmetic, using illustrative round numbers: break-even at the standard −110 price is 52.4%, because 110 / 210 = 0.524. A bettor whose true long-run win rate at −110 is 54% has durable edge. But the standard error of an observed win rate over n bets is roughly 0.5 / √n, which is about 2.2 percentage points at 500 bets. The 1.6-point gap between that winner and a break-even bettor is smaller than one standard error at 500 bets, so a full season of results cannot reliably separate skill from luck.
CLV replaces the yes or no with a measurement. Every bet produces a CLV observation whether it wins or loses, and the observation is a price differential rather than a binary outcome, so far less information is discarded per bet. The signal accumulates faster because each data point carries the size of your edge, not just the direction of one result.
The benchmark works because of what the closing line is: the last price of the market's life. By close, the line has absorbed injury news, lineup confirmations, weather, and the weight of informed money that the opening number had not seen. The close is not truth, but it is the most informed estimate the market produces. Beating it repeatedly means you are consistently ahead of the information, a definition of sharpness that does not depend on how any single game ended.
What closing line value does not tell you
- It does not grade a single bet. One CLV observation is noise, and the close itself is an estimate that can be wrong about any individual market. CLV becomes evidence only across a comparable sample.
- It does not certify the benchmark. Main markets at high-liquidity books close efficiently. Player props, minor leagues, and thin exchange markets can close at prices no informed bettor would take. CLV measured against a weak close inherits every bias in that close.
- It does not prove you could get the money down. CLV scores the price you took, not the stake the book allowed, the limits you hit, or whether the number was still available at size.
- It is not realized profit. De-vigged CLV estimates the expectation of your entries. Converting expectation into money requires volume, and the gap between the two over any finite sample is variance you have to be able to fund.
- It does not measure staking. Two bettors with identical CLV and different sizing discipline end up with very different bankrolls. CLV says nothing about bet size, correlation between positions, or drawdown tolerance.
- It is easy to corrupt with selective tracking. Graded on every bet, CLV is informative. Graded on the bets you chose to record, it is a highlight reel. The metric is only as honest as the sample behind it.
How CLV.gg measures closing line value
CLV.gg is named after the metric because measuring it is the product. The engine records offered prices across 35+ sportsbooks, exchanges, and prediction markets, builds a de-vigged sharp consensus, flags mispriced lines before they close, and then follows every detected edge to the close. Every settled bet is graded against where the market closed. On higher tiers, detection streams in real time at under 500 milliseconds. Coverage spans six core sports, soccer, basketball, tennis, baseball, football, and hockey, plus UFC, Formula 1, and NASCAR.
The full pipeline, including how the sharp consensus is weighted and de-vigged and how detections are tracked through settlement, is documented in the CLV.gg methodology. CLV.gg is analytics and education only. It is not a sportsbook and it does not sell picks.
Frequently asked questions
Is positive CLV proof of a winning bettor?
No. Positive CLV across a large, comparable sample is strong evidence that your entries beat the market's final prices, and it is one of the best indicators of long-term success. It is still an estimate against a benchmark. It says nothing about staking, limits, or whether the sample behind it is honest.
What is a good CLV percentage?
There is no universal threshold. Against a properly de-vigged close, any sustained positive figure means your entries carried positive expectation relative to the market's final estimate, and the percentage approximates expected return per unit staked. A sustained small number across many bets means far more than occasional spikes on thin markets.
Can a bettor beat the close and still lose money?
Yes, routinely, over short samples. A bet that beat the close still loses whenever the outcome lands against it, and expectation only converges toward results over volume. The reverse also happens: a bettor can run hot with negative CLV. That divergence is the reason the metric exists.
Should CLV be measured against my book's closing line or a sharp reference?
Against the sharpest available close, de-vigged. Your own book's close can lag the market or carry a skewed margin, and a benchmark is only useful if it is better informed than the price you took. CLV.gg grades against a sharp consensus built from multiple books for exactly this reason.
Does CLV work for player props and prediction markets?
The math transfers; the benchmark quality does not. Player props and thin exchange markets can close at inefficient prices, so CLV measured against those closes is weaker evidence. Prediction-market fills are quoted as probabilities and need fee adjustments before grading, which changes the formula's inputs but not its logic.
Why do sportsbooks care about a bettor's CLV?
An account that consistently beats the close is taking prices the market later agrees were too good, which makes its action predictive of line movement. Sportsbook risk teams can respond by reducing limits. Policies differ by operator and jurisdiction, and access to a price is never guaranteed.