Closing line value: did you get a good price?

Imagine buying the same train ticket for £20 before its price rises to £25. You secured a better deal than someone buying later. CLV asks a similar question about a bet: did you take better odds than the final comparable price before the event? The difference is that neither price tells you whether the bet will win.

For a back bet, larger decimal odds pay more for the same winning outcome. Taking 2.10 before the identical market closes at 1.95 is a positive raw price comparison. Taking 1.80 before it closes at 2.00 is negative.

These figures are different definitions, not conflicting answers. Label which one you use. For a fair-close calculation, first collect every mutually exclusive outcome in the closing market, remove the margin, then compare your taken odds with the resulting fair price.

For proportional removal, divide each implied probability by their total. If the selected outcome has closing probability p after that adjustment, fair odds are 1/p and fair-close CLV is (taken odds × p − 1) × 100. Other de-vig methods can give different estimates, especially on uneven markets.

Swipe the table sideways to see every column.

A well-traded closing market has had time to react to news and betting activity. That makes it a useful comparison point. It is still a market estimate, not the true probability written down somewhere.

Repeated positive fair-close CLV can support the case that your process finds useful prices. It does not guarantee profit, prove a model is calibrated, or tell you how much to stake. Small samples, correlated bets, low limits and weak reference markets all reduce the strength of the evidence.

A winning bet can have negative CLV; a losing bet can have positive CLV. Record both facts. Do not call every profitable bettor lucky, or every bettor beating the close skilled, from one short run.

Do not rush into any early price simply to beat a future number. Early markets may offer smaller limits, wider margins or incomplete team news. Decide whether the currently available price still meets your assessment.

1. Save the bet at placement

   Selection, line, bookmaker, taken odds, stake and timestamp. Keep free bets and boosts separate.

2. Choose the close in advance

   Use a consistent reference and cut-off before the start. Record the actual capture time and all sides needed to remove margin.

3. Match the contract

   Use the same handicap, total, player threshold and settlement rules. Mark missing or mismatched records rather than forcing a comparison.

4. Review coverage as well as CLV

   Show how many bets have a usable close. Label equal-bet and stake-weighted averages separately; neither is automatically the realised ROI.

Over 2.5 shots and over 3.5 shots are different bets. If the threshold moved, comparing their decimal prices directly is meaningless. Keep a closing quote for the original line, or record the line movement separately.

Likewise, an in-play capture is not a pre-match close. A stale quote from hours earlier is a last observed price; label it that way. If the opposing side is unavailable, you can still show raw price movement, but you cannot reconstruct a full two-way fair close from one price alone.

For player props, confirm the player, threshold and rules for participation. For tennis, check how each bookmaker settles retirements and walkovers. Apparently better odds may cover a different contract.

If ROI and CLV diverge for a long period, investigate the timestamps, market quality, settlement and selection process. Do not automatically explain everything as bad luck.

Take a small batch of your recent bets and compare the actual taken prices with a consistent close. Write down the definition and missing-data count. That gives you a useful starting point without pretending the metric can predict the next result.