AGATHANo gut feelings

Guide

What Is Closing Line Value (CLV)?

Closing line value (CLV) compares the price you took with the last price before the start: CLV = your decimal odds / closing decimal odds - 1. Take -105 on a side that closes at -115 and your price pays 4.43% more than the close. Averaged over many bets, it rates your prices long before results can.

By Agatha Research Desk. Updated

Record numbers refresh hourly; last refreshed .

How to calculate CLV

The price method

Convert both prices to decimal odds and divide. This is the standard method, and it is exactly the formula behind every CLV figure in Agatha's record:

-105 -> 1 + 100/105 = 1.952
-115 -> 1 + 100/115 = 1.870
CLV = 1.952 / 1.870 - 1 = +4.43%

A positive number means you took a better price than the book closed at; zero means the price never moved; a negative number means the close was better than the price you took.

The probability method

The same move, measured in implied probability. -105 implies 51.2%; -115 implies 53.5%. The market moved 2.27 percentage points toward your side. Points read naturally next to a win rate; the price method reads naturally next to a return.

The no-vig method

The strictest version removes the margin before comparing. Say the other side closed at -105. That -115 / -105 close holds 4.50% for the book, and taking it out puts the fair chance of your side at 51.1%, fair odds of -104. Your -105 was worth -0.27% against it. The close moved your way, yet the price you took still sat a little under a fair price, because the closing price itself carried the margin.

So a positive price CLV and a price better than fair are not the same thing. The first says you got a better number than the book closed at; the second says your price was better than the fair price behind that close.

The CLV calculator runs all three at once, for one bet or a pasted list. The no-vig calculator strips the margin from any two-way or three-way market.

Why CLV says more than a win-loss record

A bet's result swings a whole stake either way, so a win-loss record takes a long time to mean anything. At -110, one bet's result has a standard deviation of about 0.95 units, and telling a real 1-point return from luck at two standard errors takes about 36,000 bets. A price rarely moves more than a few percent between the bet and the close, so an average of price moves settles far sooner.

Agatha's own record shows the gap. After 2,630 graded picks, the win-loss return is -2.26%, give or take 3.42 points. The 1,839 picks with a stored close average +0.88% CLV, give or take 0.21 points. Measured by price, the same picks are about 16 times more precise, and halving the band around the return would take four times as many picks.

From Agatha's record

How precise each measure is on Agatha's record

How precise each measure is on Agatha's record
MeasurePicksAverageGive or take
Win-loss return (ROI)2,630-2.26%+/- 3.42 pts
Price CLV1,839+0.88%+/- 0.21 pts

Give or take is two standard errors: the range the true figure most likely sits in.

As of . See the full track record

Which close to use

The same book's close

Compare with the same sportsbook's last price before the start, on the same market, side and line. This is what Agatha's record uses and what any bettor can check. It asks whether you got a better number than that book finally offered.

The market close

Compare with the consensus across many books, or with a low-margin book that moves early. It asks whether you beat the price the whole market settled on, which is a harder and more telling test than one book's close.

Margins differ a lot between books. Over the last 30 days, 1xBet had the lowest median two-way closing hold among the books Agatha tracks, 2.07%, against 4.88% across all of them. The thinner the margin, the closer a close sits to a fair price.

The no-vig close

Remove the margin from both sides of the closing market first, then compare your price with the fair price. It is the strictest test, because a price can beat the close and still sit below fair, as the example above shows.

Thin markets

Not every close means the same thing. In big, liquid markets the close is the best estimate of the true chance anyone has. Player props and smaller tennis matches take far less money, limits are lower and books often copy one another, so their closes say less, and positive CLV there is weaker evidence.

From Agatha's record

Agatha's CLV so far

Of 2,630 graded picks, 1,839 have a closing price stored from the same book. Their mean CLV is +0.88% and the median 0.00%. 705 beat the close (38.3%), 715 closed at exactly the listed price (38.9%) and 419 closed at a better price than the pick got (22.8%).

Closing line value by sport and bet type

Closing line value by sport and bet type
SportBet typeMean CLVBeat the closeWith a closeGraded
WTAMatch winner+0.51%34.6%332619
MLBStrikeout props+1.02%31.1%598598
MLBMoneyline+0.65%44.9%439440
ATPMatch winner+0.93%44.9%167306
NCAAFSpreadno closing price stored yet0273
NCAAFMoneyline+1.00%41.1%5656
MLSMoneyline+0.67%43.1%5154
UFCFight winner+4.43%67.9%2850
WNBAPlayer props+3.82%44.4%4547
NFLPlayer props-1.50%32.4%3741
WNBAMoneyline+0.97%52.9%3434
NFLMoneyline+0.49%27.3%3333
NFLSpreadno closing price stored yet033
CFLPlayer propsno closing price stored yet027
CFLMoneyline+1.02%47.4%1919

Record by where the close landed

Agatha's record by where the close landed
The closePicksRecordUnitsROI
Beat the close705424-281-9.62u-1.4%
Matched the close715430-285+19.39u+2.7%
Worse than the close419237-182-31.39u-7.4%
No close stored791440-347-4-37.34u-5.0%

As of . See the full track record

The honest reading

The average CLV is positive (+0.88%) while the record is down (-58.96u). The prices say the picks were taken at good numbers against each book's own close; the results have not followed so far. Positive CLV is a reading of prices, not money in the bank.

The picks that beat the close have not paid yet: 705 of them went 424-281 for -9.62u (-1.4% ROI). The picks whose close ended up better than the listed price did worse, -7.4% over 419, which is the direction CLV predicts. No row holds more than 791 picks, so any one of them can move by several points in a busy day.

Part of the positive average comes from how the price is recorded. Each pick lists the best price across many sportsbooks at the moment it is frozen, and a price that stands out from the pack tends to drift back toward the other books by the start. That drift shows up as positive CLV at the same book even when the market as a whole did not move.

No closing price is stored for these lanes: NCAAF spread, NFL spread and CFL player props. Some lanes have only partial coverage: WTA match winner (332 of 619), ATP match winner (167 of 306) and UFC fight winner (28 of 50). Those picks are graded like every other; they sit in the No close stored row above and in the full track record, and the averages on this page leave them out.

How to track your own

  1. Write down the price and the sportsbook when you bet, before the game starts.
  2. Just before the start, note the same book's price on the same side and line.
  3. Enter both in the CLV calculator. If you have the other side's closing price too, it also gives you the no-vig version.
  4. Judge the average over a few hundred bets, not a week. One bet's CLV says almost nothing; the average says a lot.
  5. Compare against a fair price when you can: the no-vig calculator strips the margin out of any closing market.

Questions

What is a good CLV?

Any average above zero over a few hundred bets means your prices beat the numbers the books closed at. Whether it holds as the sample grows matters more than how high it is, and the no-vig version is the stricter test, because a price can beat the close and still sit under a fair price. Agatha's average so far is +0.88% over 1,839 picks with a stored close.

Does positive CLV mean you will win?

No. It says your price was good compared with the close. It turns into winnings only if the close was close to the true chance, and even then variance decides any single week. Agatha's own record shows the gap: +0.88% CLV against -58.96u in results. The losing streaks guide shows how far luck alone moves a record.

Why is Agatha's median CLV zero?

Because the most common outcome is no move at all: 715 of the 1,839 picks with a close (38.9%) closed at exactly the listed price, so the middle pick lands on zero. The mean counts the size of every move, and the moves toward the picks (705) outnumber the moves away (419).

Why do spreads show no CLV?

A spread price only means something at its line: -110 at -3 and -110 at -3.5 are different bets. The record does not yet store a spread's line in a form the closing lookup can match, so it stores no close rather than compare prices at two different lines. Spread picks are graded as usual and counted in the No close stored row.

Which closing price should I compare with?

Use the same book's last price before the start if that is all you have; use the no-vig price from a low-margin book or a consensus when you can. The no-vig calculator does the second in one step.