Did Your Prices Beat the Close? Check a Whole Season at Once
Five illustrative bets show a raw gain of +2.73 percent that falls to -0.32 percent once the margin is removed.
Closing line value compares the price you took with the market's last price after its margin is removed. It is a measure of price, not a record of profit.
The calculator below takes a whole list of bets, strips the margin from every closing market and returns the average, the number of bets above the close and a 95 percent interval. You supply the closing prices yourself, from an account you are entitled to use.
- Source
- Your own bet records; Pinnacle Betting Resources articles for the method
- Version
- Articles dated 15 September 2025, 4 April 2019 and 10 November 2017
- Limit
- Proportional margin removal only; linked bets and selection are not corrected
Closing line value: the formula and three readings
The sharp betting hub explains why some analysts read the final price of a market as its best-informed one. This page turns that idea into arithmetic you can run on your own bets, and it keeps the arithmetic visible so two calculators can be compared line by line.
The convention is short. Invert every closing price of the market to get raw probabilities, add them up (that sum, S, sits above 1 whenever the market carries a margin) and divide each raw probability by S to get the fair probability. Closing line value is then the price you took multiplied by the fair probability of your outcome, minus 1. Take 1.95 against a close of 1.90 and 1.98: the inverses are 0.5263 and 0.5051, S is 1.0314, the fair probability of your outcome is 0.5103 and the value is 1.95 x 0.5103 - 1 = -0.49 percent.
Two other readings sit beside it. The raw ratio, 1.95 / 1.90 - 1, gives +2.63 percent because the closing margin stays inside the close, and the link between the two is (1 + value) = (1 + raw) / S, here 1.0263 / 1.0314 = 0.9951. The probability reading, fair probability minus the inverse of your price, gives -0.25 points (0.5103 minus 0.5128). The tool prints all three, since a figure without its convention cannot be compared with anyone else's.
Joseph Buchdahl, in an article on Pinnacle's site dated 4 April 2019, divides the price taken by a margin-free closing price, the same ratio as above, with a removal method that allows for the favourite-longshot bias and that the article does not spell out. In a piece dated 10 November 2017 on the same site he shows why raw figures mislead: an opening price of 5.00 that shortens to 4.90 looks like a 2 percent gain, yet once the margin comes out it carries no expected profit.
Each row of the tool is one bet. Choose two or three outcomes, type the price you took, every closing price in market order, and the number of the outcome you backed. A comma works as a decimal separator, a price of 1.00 or below is refused, and up to 60 bets fit on the page. Stakes are not used: the statement averages bets, it does not weight them.
Closing line value statement
Type one row per bet. The fields start empty and the grey figures are only placeholders, so every number in the result is yours.
Reads only the numbers typed above; it sends and stores none of them.
What does a wide interval tell you?
The statement reports the average value, the raw ratio beside it, the average gap in probability, how many bets landed above the close, the standard deviation, the standard error and a 95 percent interval from Student's t distribution. The interval is the average plus or minus t times the standard error. The multiplier t shrinks as bets are added: it is 2.0423 for 31 bets and much larger for 5, which is why a short record gets such wide limits.
When the interval contains zero, the tool says the record does not separate your prices from zero. When the whole interval sits above zero, the message describes these bets only. There is no universal number of bets at which the answer becomes reliable; the interval computed from your own bets is the answer, and the spread of your values sets its width. Five bets whose values have a standard deviation of 1 point give an interval of about plus or minus 1.2 points; the record in the table below has a half-width of 5.4 points. A record that mixes short prices and long ones stays wide.
The count of bets above the close is a separate reading. A record can sit at 12 of 20 above the close with an average near zero, because a few large negative values outweigh many small positive ones.
Where do you get closing prices you are entitled to use?
Pinnacle does not publish a public archive of past odds that we could find, and its API page says access is by application, while a README whose ownership no official link confirms dates a closure to the general public on 23 July 2025. Its terms prohibit scraping and automated access without written agreement, so this page points to no script, no data vendor and no export; the page on Pinnacle data access lays out who can ask for data and on which terms. Record closing prices yourself from an account you are entitled to use: copy the price of every outcome at the last moment the market is open, or save a capture of the market page, and label the source beside the row in your own notes.
The label matters because a Pinnacle closing line, a broker's close and an exchange's last matched price are three different references, and the tool cannot tell them apart. Thin markets can make a weaker reference, so check how a line was settled and how many bettors were on it before you trust its close; how far a Pinnacle line tends to move before it gets there is the subject of the page on odds movement. Both sides of the market are needed, two outcomes or three, because a margin cannot be removed from one price alone.
The margin method is the same proportional one used for comparing a price between two routes. A power or Shin method would give another fair price, mostly on long prices, so a record full of outsiders deserves a second look at the method. Keep the record complete and unflattering, the way a desk keeps a blotter: the tool is only as honest as the list it is given.
Five bets, worked by hand
The table reproduces the tool's arithmetic on five bets. They are illustrative, not real odds, and they come from no bookmaker, broker or exchange. Bet 3 is a three-way market in which outcome 2, the draw, was backed; the other four are two-way markets.
| Bet | Price taken | Closing prices | Outcome backed | Sum of inverses | Fair probability | Value | Raw ratio |
|---|---|---|---|---|---|---|---|
| 1 | 2.10 | 1.95 / 1.95 | 1 | 1.0256 | 0.5000 | +5.00% | +7.69% |
| 2 | 1.91 | 1.87 / 2.02 | 1 | 1.0298 | 0.5193 | -0.82% | +2.14% |
| 3 | 3.40 | 2.40 / 3.30 / 3.10 | 2 | 1.0423 | 0.2907 | -1.15% | +3.03% |
| 4 | 2.50 | 2.60 / 1.55 | 1 | 1.0298 | 0.3735 | -6.63% | -3.85% |
| 5 | 1.80 | 1.72 / 2.25 | 1 | 1.0258 | 0.5668 | +2.02% | +4.65% |
Average value -0.32%, average raw ratio +2.73%, standard deviation 4.31 points, standard error 1.93 points, t 2.7764 for four degrees of freedom, 95 percent interval -5.67% to +5.04%, 2 of 5 bets above the close. Typed into the tool, the same rows return the same figures.
Four of the five bets look good in the raw column, and bet 4, a price of 2.50 against a close of 2.60, is the only visible miss. After the margin comes out, three bets sit below the close and the picture reverses. Bet 3 shows the sign flipping: +3.03 percent raw, -1.15 percent once the draw is priced against the other two outcomes. Five bets settle nothing.
Averaged, the five bets sit slightly below the close, and the interval is wide enough to hold prices well below the close and well above it.
Does positive closing line value mean you made money?
A positive average means the prices you took were better than the closing price on average in this record. It does not say you will win, that the next bets will behave the same way, or that the closing price you entered is right. One analyst, Pinho, reported in December 2025 that 6,806 football value bets measured against Pinnacle's close returned 1.9 percent when 4.2 percent was expected, a single source that has not been replicated.
Pinnacle's September 2025 article presents positive closing line value as a signal of good decisions (Pinnacle Betting Resources, 15 September 2025); this page treats that as the operator's statement and has not verified it. Buchdahl's 2019 article on the same site treats the efficiency of the closing line as a hypothesis and lists lucky or mistaken bettors as one explanation for profitable records that fail to beat it.
How far profit can drift from price over a run of bets is a separate calculation, and it belongs to the question of whether betting is profitable; this page does not simulate it. Pinnacle also says, in the same 15 September 2025 article, that consistently achieving closing line value is a key factor in whether a recreational sportsbook limits an account. That is the operator's own claim, and why bookmakers limit winners sets out the mechanism, while how Pinnacle treats winners covers its own side. For where the margin itself comes from, see how Pinnacle odds and margins are read.
Five ways the interval can mislead
The interval is computed from the bets you typed and from nothing else. Each of these sits outside it:
- Linked bets. The same match, the same day or the same model shrink the real information, so the standard error comes out too small.
- Selection. A record that holds only the best bets, the latest bets or one book describes that sample and leaves the rest of your practice out.
- One closing source. The reference is a single book at a single hour, and a less informed close is a weaker yardstick.
- The removal method. A different method gives a different fair price; the interval covers only the spread of your bets and ignores that choice.
- The shape of the values. Student's interval assumes values close to a normal curve, and long prices skew them.
A range that includes zero does not show that your prices are ordinary, and a range above zero does not show that later bets will beat the close.