
A football tip can lose after its odds shorten. Another can win after its price drifts. Neither result tells you, on its own, whether the original quote was competitive.
Closing line value, usually shortened to CLV, compares an earlier price with a defined closing price for the same selection. It adds a useful question to a tipster review: did the quoted odds hold up against the market near kick-off?
The comparison needs care. A price drop can include bookmaker margin, the closing quote may be stale, and readers may never have been able to obtain the advertised odds. A useful audit keeps those issues visible alongside results. You can do the exercise with a paper record; there is no need to place bets to test the arithmetic.
Decide what counts as the close
For this article, the closing price means the last usable pre-match quote captured before the actual start. Record its source and observation time. A snapshot taken ten minutes before kick-off is a ten-minute snapshot, even if your file calls the column “close”.
Choose that convention before inspecting the results. Otherwise, it becomes tempting to use whichever bookmaker or timestamp makes each selection look strongest.
The selection must also remain the same. Compare a home win in the 90-minute 1X2 market with that exact outcome, including the same settlement rules. A price on qualification is a different contract. So is a home handicap of −0.5 compared with −0.75.
That distinction matters with goal totals. Over 2.5 at 1.95 and over 3.0 at 1.95 share a price but have different settlement possibilities. A simple odds ratio cannot measure the value of that line change. Keep such entries separate unless your method explicitly models the different payouts.
Start with the price ratio, then name its limits
A straightforward measure is the raw price comparison:
Raw price CLV = (quoted decimal odds ÷ closing decimal odds − 1) × 100
Suppose a fictional home-win tip was quoted at 2.10 and the same outcome closed at 2.00. The raw comparison is (2.10 ÷ 2.00 − 1) × 100 = +5%.
The earlier price offers a higher gross return for the same stake if that outcome wins. That is all the calculation establishes. It does not establish a 5% chance of winning, five percentage points of forecasting accuracy, or a guaranteed 5% profit.
A one-unit bet at 2.10 returns 2.10 units if it wins, including the stake, and returns zero if it loses. Its net result is therefore +1.10 or −1 unit. The CLV figure stays the same whichever result occurs.
This separation is useful when reviewing a close match. You can record the result and still assess the pre-match price without rewriting the explanation around a late goal. The related guide to football odds movement deals with a different question: what evidence supports an explanation for the movement?
Account for the margin before calling it an edge
Bookmaker prices contain an overround. In a complete 1X2 market, add the three inverse decimal prices to see how far the quoted probability total exceeds 100%. Smarkets explains this calculation. The overround is a property of the quoted book, not a measurement of the bookmaker’s eventual profit on that match.
Here is a wholly hypothetical closing market, used only to show the arithmetic:
| 90-minute outcome | Closing odds | Raw implied probability |
|---|---|---|
| Home win | 2.00 | 50.00% |
| Draw | 3.60 | 27.78% |
| Away win | 4.00 | 25.00% |
| Total | — | 102.78% |
One simple adjustment divides each inverse price by their total. This is proportional normalization: it assumes the excess is distributed in proportion to the raw implied probabilities.
For the home win:
Adjusted closing probability = (1 ÷ 2.00) ÷ [(1 ÷ 2.00) + (1 ÷ 3.60) + (1 ÷ 4.00)] = 48.65%
The corresponding adjusted closing odds are approximately 2.056. Against that benchmark, a quote of 2.10 gives (2.10 ÷ 2.056 − 1) × 100 ≈ +2.16%, rather than the raw +5%.
The sign can change too. A quote of 2.02 beats the unadjusted 2.00 close by 1%, but sits below the adjusted 2.056 benchmark. On this method its adjusted comparison is about −1.73%.
Proportional normalization is a transparent assumption, not a discovery of the match’s true probabilities. Other margin-removal methods can give different answers, particularly across favourites and outsiders. Name the method and preserve the original prices so somebody else can reproduce the calculation.
For a binary win-or-lose selection, if you choose to treat the adjusted closing probability as a probability estimate, the expression accepted odds × estimated probability − 1 gives an estimated return per unit before costs. Its usefulness depends on that probability estimate. Pushes, half-wins and split handicaps require their own payout calculations.
Our overround explanation provides the background if the probability conversion is unfamiliar.
Audit the reader’s price as well as the tipster’s quote
A service may publish at 2.10 while a reader, opening the selection later, can obtain only 1.98. Against the same adjusted close of 2.056, those are different records:
| Price being reviewed | Decimal odds | Adjusted comparison |
|---|---|---|
| Tipster’s publication quote | 2.10 | +2.16% |
| Reader’s available price | 1.98 | −3.68% |
These are invented prices, not OptaTips performance data. They illustrate why a favourable advertised record may fail to describe the experience available to a reader.
Joseph Buchdahl’s 2016 analysis of tipster records examines both closing-price comparisons and the difficulty of obtaining fast-moving advised odds. It supplies a useful distinction: evaluating a published quote and evaluating whether followers could obtain it are separate tasks. Its historical examples do not validate any service’s current record.
Keep publication time, advertised bookmaker and original quote in one set of fields. Keep the reader’s observation time, available price and any actual accepted price in another. If availability is unknown, mark it unknown. Do not replace a missing value with the tipster’s advertised odds.
An observation on a screen also does not prove a requested stake would have been accepted. Record a known limit or rejection where relevant, without assuming that missing information means unrestricted availability. The guide to reviewing a tip after an odds change explains how a changed quote affects the original assessment.
Keep missing and awkward entries in the audit
Take every selection in a stated period. Retain losses, withdrawn selections, cancellations and entries without a usable closing comparison. Each needs an explicit status; it should not quietly disappear from the denominator.
Suppose a fictional month contains 100 published selections, but you can verify closing prices for only 72. Report coverage as 72 of 100. Any average CLV describes those 72 entries. It is not automatically a result for the full month, especially if the missing entries cluster in particular leagues or markets.
Choose the summary calculation too. An equal-weight average describes the average recorded selection. A stake-weighted average gives larger stakes more influence. Neither should silently replace the other. Keep quoted-price and reader-price summaries separate.
A positive average can be concentrated in a few unusual quotes. Check the median, the distribution of prices and the largest contributors. A record of 100 selections from 100 different fixtures is also different from 100 selections concentrated in ten matches. Shared fixtures and markets can create overlapping evidence.
There is no universal number of tips that turns CLV into proof of skill. Market quality, price range, selection dependence and the size of the measured difference all affect interpretation. An audit should reveal those conditions rather than hide them behind a “verified” label.
The closing feed needs an audit of its own
A column labelled “closing odds” can look authoritative while containing delayed quotes. As checked on 4 November 2026, Football-Data’s data documentation warns that its Pinnacle feed became unreliable from 23 July 2025, affecting both pre-closing and closing prices.
That warning concerns the named feed and period; it does not establish that every other provider is current or every older row is correct. Check the documentation for the actual dataset you use. A familiar bookmaker name is insufficient evidence of a fresh observation.
Ask when the quote was captured, whether the market was open, and whether all outcomes came from the same snapshot. A home price from one bookmaker, the best draw from another and an away price captured earlier do not constitute one bookmaker’s closing book.
Similarly, do not select the most favourable closing source separately for each row after seeing the numbers. Define the reference source, acceptable snapshot age and missing-data treatment in advance. Keep exceptions visible. Where no comparable close exists, “unavailable” is a more honest conclusion than a reconstructed number.
Read CLV beside returns and the underlying selections
A closing comparison is only as useful as its benchmark. It can offer evidence that the earlier price was competitive with a later market. It cannot certify the true probability of an outcome or establish that the person quoting it will perform the same way next season.
Keep the settled financial result alongside it. State stakes, returns, refunds and costs separately, and distinguish a selection-only record from a reader’s result after relevant fees. A positive closing comparison and a losing run can coexist; so can a winning run and unfavourable closing comparisons.
When those figures disagree, inspect the entries. Were the quotes obtainable? Did the comparison mix different lines? Were costs omitted? Did one match account for several supposed pieces of evidence? The guide to comparing tipster records covers returns, sample size and costs in more detail.
For an OptaTips review, start with the dated entries you can actually inspect. This article does not establish that every profile supplies a closing-price archive. If the required source and timestamp are absent, the CLV question remains unanswered.
What a useful review should leave on the page
A concise audit note might read:
“Hypothetical example: 100 selections published in the stated month; 72 had comparable pre-match closing snapshots. The review uses one fixed source and proportional margin adjustment. Publication-price comparisons and reader-price observations are reported separately. Missing snapshots remain listed; settled returns and fees are recorded independently.”
The note is useful because another reader can see the scope and challenge the method. Add the actual figures only when the underlying evidence supports them.
When reviewing a tipster’s CLV claim, ask for the original quote, the exact comparison market, the closing source and time, the margin method, and coverage of the full record. If those details are missing, a positive percentage leaves too much unexplained. A modest, reproducible finding is more useful than a confident number you cannot check.
For adults aged 18 and over. This article explains record analysis, not a recommendation to place a bet. Historical results and price comparisons do not guarantee future profit.
댓글 0
아직 댓글이 없습니다.