Closing line value asks one question about a bet: was the price you took better or worse than the market's last price before the game started?
That last price is called the closing line. It reflects everything the market learned between the moment you placed your bet and the start of the game: injuries, lineups, weather and money coming in on each side. If you consistently get better prices than the close, you are reading the market earlier than it settles. That is a question about price. It is not a question about whether any single bet won.
We take prices from our own odds store as a series of dated snapshots. For each bet, the close is the last snapshot for that game and market taken before the game started, and it counts only if that snapshot was taken within an hour of the start. We use the time the snapshot was taken, not the time a book last changed its number, so a line that simply never updated does not count as the close.
The reference close is Pinnacle's, with its margin removed. We take both sides of the market at the close, turn each price into a probability, and scale the two so they add up to one. That gives a fair probability for your side with the bookmaker's cut taken out. When you bet at one of the other books we hold, you can also get a second score against that book's own close.
You upload a file of your logged bets. Each row carries the date, the sport, the game, the market (moneyline, spread or total), the side, the point where there is one, the price you took in American odds and the book you used.
Each bet is matched to its game, market and side at the close. For spreads and totals, the close has to exist at the exact point you bet; we never estimate a nearby number.
The score is a percentage. It is what your price would be worth at the fair close probability for your side: positive means your price beat the close, negative means the close was a better price than the one you took. It measures price, never results.
A bet we cannot match exactly gets a reason instead of a score, and it is left out of your averages. The reasons are: the game was not found, the game matched more than one entry, the date is outside what we hold, the side was not offered at the close, the point was not available at the close, only one side of the market was priced at the close, there was no close within an hour of the start, or the row could not be read, which includes a sport we do not cover. Rows past the free allowance are counted and shown as locked instead of scored.
For each bet you see its score, or its reason if it could not be scored. Your scores are also broken down by sport, by market, by book and overall, as the number of scored bets and the average score. Any group with fewer than five scored bets shows its count and no average, because an average over a handful of bets is noise.
You do not see the closing price or the fair close probability for any bet. A file with a result or outcome column, or with any header other than those eight columns, is refused, and nothing about outcomes is ever stored. The tool does not grade wins or losses.
A positive average over a small number of bets is noise. The score measures the prices you took against a market close, not the quality of any single pick, and it says nothing about future results.
Coverage as of September 15, 2026 is MLB.
Upload your file at defintel.net/clv. Free accounts get the first 20 rows scored.
Five small tools at defintel.net/calculators: implied probability, book hold, closing line value, price gap, and how bets and contracts settle.
Educational analysis only. Not betting, financial or investment advice.