On this page
What closing-line value is
The closing line is the last price before a game starts. By then the market has absorbed the news, the injury reports and the money from the sharpest bettors. It's the best estimate the market will ever make. Closing-line value (CLV) compares the price you took with the closing fair price. If you regularly get better prices than the close, you're probably beating the market.
Why it matters more than results early on
Results are noisy. A bettor with a real 3% edge on even-money bets wins about 51.5% of the time; over 200 bets, normal luck can easily show them losing. CLV measures the price, not the coin flip, so it settles much faster. A long run of positive CLV with poor results usually means bad luck; a long run of negative CLV with good results usually means good luck that won't last.
How to measure it
Measure CLV against the fair closing price, with the vig removed, not against the closing price at your own book. Otherwise the vig makes every bet look worse than it was.
CLV (points) = closing fair probability − 1 ÷ d_you_took
CLV (EV) = closing fair probability × d_you_took − 1
Our collector keeps the last pregame snapshot before each start as the closing line. The closing fair is that snapshot's devigged consensus.
Worked examples
Took +110 (2.10); closing fair 50.5% → CLV +2.88 points; CLV EV = 0.505 × 2.10 − 1 = +6.05%
Took −105 (1.9524); closing fair 52.0% → CLV +0.78 points; CLV EV = +1.52%
Took −110 (1.9091); closing fair 50.0% → CLV −2.38 points; CLV EV = 0.50 × 1.9091 − 1 = −4.55%
The third bet paid the full vig on a market that closed at a coin flip. Win or lose, it was a worse price than the close.
The formula our engine uses
Every flagged edge is graded with the same two numbers, computed from the net price at first sight (d_flag) and the closing fair probability (p_close):
CLV (points) = p_close − 1 ÷ d_flag
CLV (EV) = p_close × d_flag − 1
Flagged at +120 (2.20), closing fair 47%: CLV = 0.47 − 0.4545 = +1.55 points; CLV (EV) = 0.47 × 2.20 − 1 = +3.40%
Same bet, closing fair 44%: CLV = −1.45 points; CLV (EV) = −3.20%
CLV (EV) answers a direct question: if the closing price was right, what was this bet worth when we flagged it?
Why CLV settles faster than results
At −110, one bet's result is either +0.91 or −1 units. Its standard deviation is about 0.95 units, almost the whole stake. Averaged over N bets, the standard deviation of your ROI is roughly 0.95 ÷ √N:
| Bets | Standard deviation of ROI |
|---|---|
| 100 | 9.5% |
| 500 | 4.3% |
| 1,000 | 3.0% |
A true 3% edge is only one standard deviation from zero after 1,000 bets. CLV doesn't carry that coin-flip noise. Each bet's CLV is a price difference of a few points, not a swing of a whole unit. For illustration, if your CLV varied by 3 points from bet to bet, the average over 100 bets would be pinned to within about 0.3 points. That's why a few hundred bets of CLV can say what thousands of bets of results can't.
How we'll use it
Every edge our tools flag is logged at first sight with the price a member could have taken, then graded at the close with its CLV and later its result. Coming soon The public track record will show settled results only, never live edges. It will report CLV and ROI by EV bucket (0–2%, 2–4%, 4–6%, 6–10% and 10%+), by devig method and by venue, losses included. If the higher EV buckets don't show higher CLV, that will be visible too. Coming soon The bet tracker will measure your own bets the same way; + Track already saves prices on this device.
Measure your own CLV today
- Write down the price you took and the book.
- Just before the start, record the market. Take two-sided prices from a few books, ideally sharp ones, on the same line. The Odds Screen's Fair column gives you the consensus fair directly, with your devig method and weights.
- Compute the CLV with the formulas above, or with the no-vig and EV calculators: the closing fair probability as the win probability and your price as the odds.
- Keep going. Track the average over every bet, not just the memorable ones.
If your line moved to a different number by the close (you took −3 and it closed −3.5), compare against the closing fair at your number, which the fair ladder provides, not the closing price at the new number.
Common mistakes when measuring CLV
- Using the closing price at your own book. It still includes that book's vig, so every bet looks worse than it was. Use the fair close.
- Ignoring fees. On an exchange, measure CLV from your net price after fees, not the quote.
- Comparing different lines. If you took −3 and the market closed at −3.5, compare with the closing fair at −3, not the closing price at −3.5.
- Cherry-picking. Measuring only the bets you remember turns CLV into a story. Measure every bet.
Limits of CLV
- Thin closes. A prop priced by two books closes on little information. CLV against a thin close is a weak signal.
- The close isn't truth. It's the best available estimate, not the true probability. Over large samples it's a good yardstick, not a perfect one.
- Timing. Betting early captures more CLV when you're right and more risk when news breaks against you.
- Limits. Books watch CLV too. Consistently beating the close is a common reason accounts get limited.
In build
This tool is in build on the same tested engine. The app shows what is coming and what already runs.