Win rate lies. Hot streaks lie. Even six-month ROI can lie. The one number that doesn't lie — the one sharp books actually watch when they decide whether to limit your account — is whether you consistently beat the closing line.
Most bettors evaluate themselves the way casual fans evaluate teams: by results. They look at this week's record, this month's ROI, the unit total at the top of their tracker. Those numbers feel concrete because they're tied to dollars that moved. But they're also dominated by variance — and over short windows, variance can hide a losing strategy behind a winning record, or bury a winning strategy under a cold stretch. The only way out of that variance trap is a metric that doesn't depend on what happened after the ball was thrown, the puck was dropped, or the first pitch was thrown. That metric is closing line value.
Closing line value, almost always abbreviated CLV, measures one thing: the difference between the price you got on a bet and the price that bet closed at right before the game started. If you bet a moneyline at +130 and the closing price was +115, you got a better number than the market eventually settled on. That's positive CLV. If you bet the same team at +130 but the closing price drifted to +145, the market moved against your side after you bet. That's negative CLV — you paid more in expected value than the market thought the bet was worth by the time it locked.
The reason CLV matters is the closing line itself. The closing line is the most efficient price the betting market produces all week. By the time a game starts, the line has absorbed every public bet, every sharp bet, every injury report, every weather forecast, and every public adjustment from the books. There is more information packed into the closing price than at any point earlier in the cycle. That's why beating it consistently — paying a price the market would later regard as too generous — is the strongest available signal that you are picking winners faster than the market can.
Suppose you bet 100 NFL games at -110 across a season and went 53-47. That's a 53% win rate, slightly above the 52.4% breakeven on a standard -110 spread. You'd net about 0.3 units of profit on 110 units risked — a +0.3% ROI. Statistically that's a coin flip. You cannot tell from that result whether you had a real edge that variance compressed, or no edge at all and got lucky to finish above breakeven.
Now imagine across those same 100 games, you averaged +0.8 cents of CLV per bet — meaning the closing line was, on average, eight tenths of a cent worse than the price you got. That tells a different story. It says the market consistently agreed with you that the price you took was generous — across 100 independent events. That's not luck. That's evidence of edge.
The win rate and the ROI report on what happened. CLV reports on whether the price you paid was correct given everything the market eventually learned. The first is a noisy lagging indicator. The second is a noisy leading indicator. Both have noise, but only CLV is statistically robust enough to converge on a real answer over a couple hundred bets.
Every major sportsbook tracks CLV at the account level. Their internal model knows the closing line on every game they offered. When they look at a customer's bet history, they don't see "this person went 53-47" — they see "this person got an average of +1.2 cents of CLV across 400 bets." That second number is the one that gets accounts limited.
The reason is simple: a customer who beats the closing line consistently is, by the book's own internal valuation, taking value off the table the book did not intend to give. The book's prices are wrong relative to where they end up. A losing customer who beats CLV is more dangerous to a sportsbook than a winning customer who doesn't, because the losing customer is on the right side of long-run math — they'll eventually win — while the winning customer is on the wrong side and will eventually lose.
This is why sharp bettors who care about account longevity work harder on CLV than on raw P&L. The path that keeps an account alive longest is to bet large enough to matter and small enough not to alarm — while putting up CLV numbers that match the book's internal assessment of the bet, not the customer's external ROI.
The reason CLV is statistically powerful is that it removes variance from the result. When you bet a moneyline at +130 and the game finishes, the outcome is binary — won or lost. The variance on a single bet's result is enormous. To get a statistically meaningful read on whether your edge is real, you need hundreds or thousands of bets.
CLV doesn't have that problem. The CLV on a bet is fixed the moment the line closes. There's no "did the team win" component. Your CLV on a bet is whatever the closing line minus your taken price was — done, settled, recorded. That means CLV converges to its true average much faster than ROI. A bettor with a real edge will have a stable, slightly-positive CLV average across 100 bets. The same bettor's ROI across 100 bets is still a coin flip in disguise.
If you're not measuring CLV, you're flying blind on whether your picks are actually good. The mechanic is simple. For every bet you place, record:
Translating cents to implied probability is straightforward and what really matters statistically, but the cents version is fine for a first pass. The simple rule: across 100+ bets, if your average CLV is positive, you are beating the close. If it's neutral or negative, the market is closing where you bet — or moving against you.
For spread and total bets, the same logic applies in points instead of cents. You bet a team -3.5 and the line closed at -2.5. You got the better number. That's positive CLV. The conversion to implied probability uses the same logic as moneyline cents — every half-point of spread, on average, moves implied probability by a measurable amount. Want the math done for you? Plug your taken price and the closing line into the CLV Calculator and it returns CLV in cents, implied-probability delta, and a verdict on whether you beat the close.
The simple CLV calculation (your price minus the closing price) is good enough for the first 100 bets of any tracking practice. The more rigorous version — what sharp accountants of edge use — compares your taken price to the de-vigged closing line, not the raw posted close. The reason matters.
Suppose you take a team at +130. The market closes at +120 on the same team and -130 on the opponent. The naive CLV looks like 10 cents of pickup. But the de-vigged closing line on your side — what the market thinks the fair price is, with the book's hold mathematically removed — might actually be +128. Your real CLV is only 2 cents, not 10. The other 8 cents you thought you'd gained was vig, not edge.
This matters more on tighter markets. Major-sport moneylines (NFL, NHL, NBA) at sharp books de-vig to within a couple cents of the posted price, so the gap between naive and rigorous CLV is small. Player props, alternate lines, and same-game parlays at retail books carry 8–15% hold — the gap between posted close and de-vigged close can be enormous. A naive CLV calc on those markets routinely overstates real edge by 3–5 percentage points. If you're going to track CLV on prop bets at all, devig first. (See the No-Vig Fair Odds guide for the math.)
The academic and industry literature on CLV is consistent on one finding: CLV is a leading indicator of ROI, and ROI is not a leading indicator of anything. A bettor's average CLV across a meaningful sample (typically 200–500 graded bets) reliably predicts their long-run ROI within ~1–2 percentage points, while their actual short-run ROI can diverge from the long-run expectation by 5–10 percentage points purely from variance.
The reason is statistical: CLV is the average of a continuously-valued signal (closing price gain in cents) measured against a noiseless reference (the closing line itself is fixed). ROI is the average of a binary signal (the bet won or lost) measured against an enormous variance distribution. Even when CLV and true ROI move together — which the math says they must, eventually — CLV gets there with a small sample, while ROI gets there with a huge one. For a recreational bettor placing 200–400 bets a year, that's the difference between knowing whether your process works inside one season and never being statistically sure.
Sharp sportsbook risk teams have published similar findings internally for years: limiting accounts on positive-CLV behavior catches future winning bettors faster than limiting on ROI. They've been able to do this longer than most retail bettors have known what CLV was, which is part of why their account-limiting decisions can look unreasonable from the customer's side — the book is reacting to a signal the customer is rarely tracking.
Public estimates put a serious recreational sharp at around +2 to +3 cents of average moneyline CLV. A professional with access to multiple books and price-shopping discipline can sit at +4 to +6 cents. The vig on a standard moneyline market is about 4 to 5 cents total (split across both sides), which means a CLV of +2.5 cents is enough to overcome the vig and pay long-run profit.
Anything zero or negative is a signal to either tighten the strategy or stop. Sustained negative CLV with positive short-term ROI is almost always a luck story that mean-reverts. Sustained positive CLV with break-even short-term ROI is almost always a variance story that mean-reverts the other way — toward profit.
CLV is wired into our model as a kill criterion, not as a marketing metric. Every official pick is logged with the price we sent and the closing line on the same market. If a model — or a specific edge within a model (a goalie workload signal, say, or a specific F5 totals shape) — runs negative CLV across a meaningful sample, that signal gets retired even if the short-term ROI is positive. The logic is non-negotiable: a strategy with negative CLV will eventually lose, full stop. We're not interested in surfing variance until the wave breaks.
Conversely, when a strategy posts strong CLV but short-term ROI is flat or negative, the response is to size it normally and wait. CLV says the price was right, the result was unlucky, and the math will catch up. That's the discipline. Our published track record is built on it.
For all its power, CLV is not a silver bullet. A few things to keep in mind:
On thinly-traded games — small-conference college basketball, soccer leagues with no sharp action, exotic prop markets — the “closing line” doesn't get the same scrutiny. The closing price in those markets is closer to the opener: not a real consensus, just where the book set it. CLV signals there are noisier and less trustworthy.
CLV tells you whether you're beating the market price. It does not tell you whether your edge is large enough to justify the risk you're taking. A bettor at +1 cent CLV who bets too large still goes broke from variance. CLV is the picking metric. Unit sizing is the survival metric. They're separate.
In some markets, especially first-half NBA and overnight college lines, the “close” is set well before sharps engage. CLV calculated against those soft closes overstates real edge. Pros adjust by using a later reference price — for example, the price 30 minutes before game start at a sharp book — rather than the official close.
Closing line value (CLV) is the difference between the price you got on a bet and the price the same bet closed at right before kickoff/tipoff/puck drop. Positive CLV means you got a better number than the market eventually settled on — an indication you bet a price the market regarded as too generous. Across a large enough sample, average CLV converges to a reliable measure of edge.
Subtract your taken price from the closing price. On a moneyline bet, CLV is measured in cents (e.g., taken +135, closed +120 → +15 cents of CLV). On spread/total bets, it's measured in points. The implied-probability version — what statistical analysis uses — converts both prices to probability and takes the difference. Our CLV Calculator runs both versions.
For evaluating whether your strategy works, yes — over realistic sample sizes (under 1,000 bets). CLV converges to truth faster than ROI does because it removes the variance of individual game outcomes. ROI is the metric that pays your bills; CLV is the metric that tells you whether your bills are paid by skill or by luck. Both matter; they're not interchangeable.
Recreational sharps typically target +2 to +3 cents of average moneyline CLV. Professionals with multi-book line shopping discipline can reach +4 to +6 cents. Anything zero or negative across 200+ bets is a sign the strategy isn't beating the market. Crucially: any positive CLV after 200+ bets is a real edge, even if small.
Books track CLV internally on every account because it predicts future profitability faster than P/L does. A customer with sustained positive CLV is, by the book's own internal model, taking value the book didn't intend to give — even if that customer is currently losing money to variance. Limiting positive-CLV accounts is a defensive risk-management decision: it stops the long-run bleed before the variance washes out.
Yes — and this is actually the most common state for serious bettors during their first season. Positive CLV with negative ROI is a variance story: the prices you took were good, the games went the wrong way, the math will catch up if you keep doing the same thing. The reverse (positive ROI with negative CLV) is a luck story that mean-reverts toward losses. The first is sustainable; the second isn't.
For informational and entertainment purposes only. Past performance does not guarantee future results. Sports betting involves risk — never bet more than you can afford to lose. Please gamble responsibly.