Closing line value: what it measures & what research says
Last reviewed · Editorial policy
Closing line value compares the price taken on a bet with the last price before the market closed, & academic studies of betting & prediction markets show both why that yardstick is used & where it falls short.
What CLV is
Closing line value, often shortened to CLV, compares the price taken on a bet with the last price available before the market closed. If the price taken was better than the closing price, the bet is said to have beaten the close.
The comparison works through implied probability, meaning the chance of winning that a price suggests. On Kalshi, a US prediction market, contracts are quoted with prices ranging from 1 cent to 99 cents, according to research by Constantin Bürgi, Wanying Deng & Karl Whelan of University College Dublin. They explain that someone who buys a contract for 70 cents can win the 30 cents put up by the counterparty.
Take a hypothetical case. Someone buys a Yes contract at 40 cents, an implied 40% chance, & the market closes at 50 cents, an implied 50% chance. The buyer paid less than the closing consensus, so the position has positive CLV. Had the market closed at 30 cents, the CLV would be negative.
The sources used for this guide do not set out a standard definition or formula for CLV. The example is an illustration of the idea, not a figure from the research.
Why the close is the benchmark
The UCD team’s working paper on Kalshi, which carries a July 2025 date on its series cover & September 2025 on its title page, examined how the accuracy of prices changes as markets near their close, using transaction-level data on 46,282 contracts from 12,403 events.
The authors collected the final traded price as each market closed &, where available, earlier prices at 24-hour intervals up to 10 days before. They found that Kalshi’s prices are informative & “improve in accuracy as markets approach closing”.
Two caveats apply. Kalshi is a prediction market covering culture, weather, sports, company announcements, financial markets & politics, so its findings are not a study of bookmakers’ sports odds. And “closing” there means the close of trading on a contract.
Older work points the same way. Justin Wolfers & Eric Zitzewitz, in a 2006 NBER working paper, found that for a broad class of models prediction market prices are usually close to traders’ average beliefs, & that prices typically provide “useful (albeit sometimes biased) estimates of average beliefs about the probability an event occurs”.
How to calculate it
The first step is to turn both prices into implied probabilities. On a contract priced in cents, the price maps directly to a percentage. The UCD paper notes that, before fees, a contract costing X cents must win X percent of the time to break even.
The second step is to account for the operator’s margin, because a quoted price is not a neutral probability. Kalshi charges fees, so the UCD authors say breaking even requires contracts to win more often than their price implies.
Bookmakers build in a commission known as the vig. Steven Levitt’s 2004 Economic Journal paper gives a typical example: bettors pay 110 units if a bet loses but are paid only 100 units if it wins.
The sources reviewed here do not describe a method for removing that excess from a closing price.
The last step is to express the gap as a number. In the hypothetical example above, the close implied 50% against 40% paid. That is 10 percentage points, or 25% relative to the price paid. The sources reviewed here do not set a convention for which form to use.
CLV vs results
One result says little about whether a price was good. The UCD paper notes that, before fees, a contract costing X cents must win X percent of the time to break even, so a 50 cent contract priced at that break-even rate still loses half the time.
So a bet that beat the close can lose, & a bet that did not can win. CLV grades the price, while the result grades a single outcome.
Levitt contrasts sports betting with roulette, keno & slot machines, where “the law of large numbers dictates profits for the house”. The same logic implies that a pricing edge shows up only across many bets.
The study covered in the next section judged its strategy across thousands of games. The sources do not say how many bets are needed before either CLV or results become reliable.
Evidence that beating consensus prices pays
A study posted on arXiv, titled “Beating the bookies with their own numbers – & how the online sports betting market is rigged”, tested whether football bookmakers’ own prices could be used against them.
Instead of building a forecasting model, the authors “exploited the probability information implicit in the odds publicly available in the marketplace to find bets with mispriced odds”. They report that the strategy was profitable in three settings: a 10-year historical simulation using closing odds, a 6-month simulation using minute-to-minute odds, & a 5-month period staking real money with bookmakers.
The authors say they made their code, data & models public. They conclude that the football betting market is inefficient & that bookmakers “can be consistently beaten across thousands of games”.
The abstract does not use the term closing line value. Its approach rests on a related idea: the market’s consensus price as the yardstick for judging any one bookmaker’s odds.
Bookmakers & successful bettors
The same arXiv study describes the authors’ betting experience to show how the industry “compensates these market inefficiencies with discriminatory practices against successful clients”. The abstract does not detail what those practices were.
Levitt’s paper explains the incentive. A small number of bettors with positive expected profits “could prove financially disastrous to the bookmaker”, he writes, because they could build large bankrolls or sell their information to others.
Using about 20,000 wagers on the National Football League, placed by 285 bettors in a high-stakes handicapping contest at an online sportsbook, Levitt found that bookmakers are more skilled than bettors at predicting games. He found they also exploit bettor biases by setting prices that deviate from the market clearing price, the price that would balance supply & demand.
His example is a bookmaker who knows local bettors prefer the local team & skews the odds against it. Levitt notes a limit: bettors who know the correct price can profit if a posted price strays too far from the true odds.
Limits of CLV
A closing price can be informative & still biased. The UCD study found a favourite-longshot bias on Kalshi, meaning cheap contracts win less often than their price implies & expensive ones win slightly more often. Buyers of contracts under 10 cents lost over 60% of their money, while contracts above 50 cents earned a small positive return.
Wolfers & Zitzewitz call prediction market prices “sometimes biased”. Beating the close therefore means beating a yardstick that may itself be off, especially at long odds.
Which side of the market a trader is on also matters. On Kalshi, makers post offers & takers accept them. The UCD paper finds favourite-longshot patterns for contracts bought by both makers & takers, but more pronounced for prices accepted by takers. It also reports that the average return on Kalshi contracts was minus 20% before fees & minus 22% after fees.
In some markets the close moves little. In Levitt’s American football data, the posted price changed an average of 1.4 times per game in the five days before kick-off. The Tuesday spread was within one point of the kick-off spread in 90% of games.
Questions
What does closing line value measure?
It compares the price taken on a bet with the last price before the market closed. A price better than the close is said to have beaten it.
Why use the closing price rather than an earlier one?
A study of Kalshi by University College Dublin economists found that contract prices improve in accuracy as markets approach closing. That study covered a prediction market, not bookmakers’ sports odds.
Can a bet beat the close & still lose?
Yes. The UCD paper notes that, before fees, a contract costing X cents must win X percent of the time to break even, so a 50 cent contract priced at that rate still loses half the time.
Is the closing price an unbiased probability?
Not necessarily. The Kalshi research found a favourite-longshot bias in Kalshi’s prices, & Wolfers & Zitzewitz describe prediction market prices as useful but “sometimes biased” estimates.
Do bookmakers treat winning customers differently?
The authors of the arXiv football betting study report “discriminatory practices against successful clients”. The abstract does not set out what those practices were.
Sources
Sources checked 9 October 2026. Information only. 18+.
Discussion
Ask a question, add useful context or share a source. Keep it relevant & respectful.
Add a comment