CLV: How to Track It, Read It, and Stop Lying to Yourself
How to Track CLV: Formulas, a Logging Template, and the Traps
Most bettors judge themselves the way drunk poker players judge their reads. They look at the last 20 results, decide they are either a genius or cursed, and change everything. That is not analysis. That is vibes with a spreadsheet.
This page is the workshop manual. If you want the case for why closing line value is the metric that matters, and why sportsbooks limit it rather than winners, start there. What follows assumes you’re sold and want to actually measure it: the math for spreads, totals and moneylines, how to strip the vig out, a log you’ll keep filling in, spreadsheet formulas, and the situations where your CLV number is lying to you.
Two Kinds of CLV You’ll Be Logging
CLV is the gap between the price you bet and the price the market closed at. For tracking purposes it comes in two flavors, and your log needs a column for each.
- Line CLV (points)
Did you beat the closing spread or total by points or half points? - Price CLV (odds)
Did you beat the closing odds on the same line, meaning a better payout for the same outcome?
In the NFL the point usually matters more because of key numbers. In MLB moneylines the cents are the whole game.
CLV for Spreads and Totals
Spread CLV (points)
Example: you bet Team A -2.5. The market closes Team A -3.5.
You beat the close by a point. In the NFL that point crosses 3, the most common final margin in the sport, so it is worth far more than a point in the middle of nowhere. In the NBA it is usually smaller but still meaningful.
The formula, from your side of the bet:
Spread CLV = your spread minus the closing spread
Bet -2.5, closes -3.5: (-2.5) – (-3.5) = +1.0. Bet +3.5, closes +2.5: 3.5 – 2.5 = +1.0. Positive means you beat it.
One thing the points column can’t tell you on its own: if the price also moved, a point of line CLV can be partly cancelled. Taking -2.5 at -105 and seeing it close -3.5 at +110 on the same side still beats the close, but by less than the raw point suggests. That is why you log the odds next to the line, every time.
Total CLV (points)
Example: you bet Under 47.5. It closes 46.0. You beat the close by 1.5 points. For an Over, flip the sign: bet Over 44, closes 45.5, also +1.5.
Totals cluster rather than having one magic number, so a half point is not worth the same everywhere. In NFL totals, 41, 43, 44, 47 and 51 land more often than their neighbors because scoring comes in 3s and 7s. Even so, a totals hook is almost never worth as much as a spread hook through 3 or 7. Don’t record them as equivalent in your head just because the columns look the same.
CLV for Moneylines
Moneyline CLV is purely about price. Same team, different payout, different bet. Log it in implied probability, not cents.
Convert American odds to implied probability
For plus odds:
Implied probability = 100 / (odds + 100)
For minus odds:
Implied probability = |odds| / (|odds| + 100)
Example: you bet +145 and it closes +130.
+145 = 100 / 245 = 40.82%
+130 = 100 / 230 = 43.48%
You own a bet that needs 40.8% to break even while the closing market priced it at 43.5%. That 2.7-point gap is your CLV before de-vigging, which comes next.
Why “cents” is a bad unit
Bettors love saying they beat the close by 10 cents. Ten cents is not a fixed amount of value. Going from +100 to +110 changes the break-even from 50.0% to 47.6%, a 2.4-point gain. Going from -300 to -290 changes it from 75.0% to 74.4%, barely half a point. A bettor who lives on big favorites can report great “cents” CLV and have almost nothing. Probability is the only unit that means the same thing at every price.
De-Vig the Close or Your Numbers Are Flattering You
A closing line includes the book’s margin. Compare your price to it raw and you overstate your edge, sometimes badly.
The simple fix is to take both sides’ implied probabilities, add them up, and divide each side by the total.
No-vig probability for your side = your side’s implied probability / (sum of both sides)
Worked example. You bet the dog at +125 (break-even 44.44%). The market closes favorite -130 / dog +110.
-130 = 56.52%. +110 = 47.62%. Sum = 104.14%.
No-vig dog = 47.62 / 104.14 = 45.73%.
Against the raw close, your CLV looks like 47.62 – 44.44 = 3.2 points. Against the no-vig close it is 45.73 – 44.44 = 1.3 points. Same bet, less than half the edge. Across a season, the difference between those two numbers decides whether you think you’re a winning bettor.
Proportional de-vigging like this is the standard starting point. It slightly misallocates the margin on big favorite/underdog splits, where books tend to load more of it onto the longshot. For most bettors that is a rounding error. The no-vig betting guide covers the method in more depth.
How to Track CLV Without Hating Your Life
You have two realistic options. Betstamp grades bets against the close automatically, reports CLV against both the book you used and the best available price, and de-vigs for you. It’s what we use for every BetPredictionSite pick. The other option is a spreadsheet you control. Either works. What doesn’t work is filling anything in the next day, when you “totally remember” the number you got.
Minimum viable log (6 columns)
- Date and time placed
- Sport and game
- Market (spread, total, moneyline)
- Your bet (line and odds)
- Closing line and odds
- Result
Enough to see whether you’re beating the market at all.
The full log (the version that teaches you something)
- Date
- Time placed
- Sport and league
- Game
- Market type (spread/total/ML)
- Side (team/over/under)
- Bet line (if applicable)
- Bet odds
- Stake (units)
- Book
- Closing line
- Closing odds, both sides (you need the other side to de-vig)
- Close source (which book or consensus; we suggest a Pinnacle/Bookmaker consensus)
- CLV points (spreads/totals)
- CLV probability, no-vig
- Flag (scratch, void, late news, see below)
- Notes (injury, weather, steam, limit, why you bet)
The notes column is where you improve. “Why did I lose CLV here?” usually has an answer, and the answer is usually “I bet it after it moved.”
Spreadsheet formulas you can paste
Assuming your bet odds are in H2, your side’s closing odds in L2 and the other side’s closing odds in M2 (Excel or Google Sheets):
Bet implied probability:=IF(H2>0,100/(H2+100),-H2/(-H2+100))
No-vig closing probability for your side:=IF(L2>0,100/(L2+100),-L2/(-L2+100))/(IF(L2>0,100/(L2+100),-L2/(-L2+100))+IF(M2>0,100/(M2+100),-M2/(-M2+100)))
Price CLV = no-vig closing probability minus bet implied probability. Format it as a percentage. Positive is good.
Which close, and when
Pick one closing source and never switch it. Best: a sharp market close. Second best: a consistent consensus from a reputable odds screen. Worst: whatever your favorite soft book closed at, because soft books close lazily and it will flatter you.
Timing matters too. Grab the close as near to the start as you can: kickoff, tip, first pitch. A number taken 90 minutes early isn’t the close in the NBA, where late scratches routinely move a line two points in the last half hour.
Bets That Should Not Count
Some rows will wreck your average if you let them in unflagged.
MLB pitcher changes. If you bet with listed pitchers and the starter changes, most books void the bet. Delete the row. If you bet “action” and the ace gets scratched, the line can move 60 cents against you on information nobody had when you bet. That’s real money lost but not a process error. Flag it and look at your CLV both with and without these.
Late NBA scratches work the same way in both directions. A star ruled out 20 minutes before tip can hand you four points of CLV you did nothing to earn. Flag those as well, or one lucky Tuesday will prop up a month of mediocre entries.
Weather in outdoor sports is the gray area. Betting an Under before a forecast turned to 25 mph wind is only skill if the forecast was already readable when you bet. Be honest in the notes column.
What Good CLV Looks Like by Sport
Track average no-vig CLV first and the percentage of bets beating the close second. The beat rate is the number people quote. The average is the one that pays. Neither means much before a few hundred bets.
NFL
The most efficient mainstream market. Edges are small and key numbers dominate. Regularly beating the close on NFL sides, even by small margins, is meaningful. The flip side is that you need multiple seasons of NFL sides to reach a sample worth trusting, since there are only 272 regular-season games a year.
NBA
An information market. Injury reports and load management cause violent moves, so being early matters more and CLV swings are bigger than in the NFL. Be honest about where yours comes from. If your edge is being quicker on public injury news, it’s real, but it’s fragile and it’s the first thing a book notices.
NCAAB
Softer, with more stale numbers outside marquee games. CLV is easier to get, but the close is less trustworthy because limits are lower and information is patchier. A Tuesday night Horizon League close is not a Saturday Big Ten close. See why CLV isn’t equal in every market for how much to discount.
MLB
CLV lives in moneyline cents, and this is where the probability conversion earns its keep. Small edges repeated over 2,430 games add up. Pitcher confirmations and lineup cards drive the moves, so the time-placed column matters more than any view you have about momentum.
Traps That Make Bettors Think They Are Sharp
Steam chasing
If you bet after the move, you’re not following sharps. You’re buying their leftovers at a worse price, and your CLV column will say so. You might win the bet. Long term, you’re paying extra for your pride.
Fake CLV from a slow book
Sometimes you beat the close because your book was slow or shading for local liability, not because you predicted anything. That’s still usable money, but it comes with low limits, bet delays and eventual restrictions. The book column tells you whether your process scales or you just found a window that will close.
Self-generated CLV
In thin markets, your own bet can move the line. It then “closes” in your favor because you pushed it. Props and small-conference totals are where this happens. If your best CLV comes from markets where you’re a big share of the handle, discount it heavily.
Using the wrong close
Grade against a soft book’s close and you look better than you are. Grade against a sharp close and it’s harder but honest. Pick the source that makes you slightly uncomfortable.
Parlays and SGPs
Parlay CLV is messy because pricing isn’t linear, and same-game parlays include correlation adjustments that differ by book. If you insist on tracking it, log each leg’s price at bet time against its own close, and don’t pretend the result is as clean as a straight bet.
The Weekly Review (15 Minutes, No Excuses)
Once a week, pull three things:
- Average no-vig CLV by sport and market
- Percentage of bets beating the close
- Your 10 worst negative-CLV bets and why they happened
Then read them honestly:
Positive CLV, losing money
Variance. Don’t torch a process that is buying good numbers because the universe flipped tails ten times. Keep the stake size where it is.
Negative CLV, making money
A heater. Enjoy it and don’t believe it. Fix your entry timing and price discipline before regression shows up with a baseball bat.
Flat CLV
You’re betting too late, betting only the sharpest markets, or not shopping. Line shopping fixes the last one faster than anything else on this page. Reading line movement and knowing when the hook is worth it fix the others.
Using Projections as the Trigger
The cleanest way to stop forcing bets is to decide your fair number before you look at the market, then only bet when the market is wrong by enough. Our daily projections give you that number for MLB, NBA and NFL. Our AI sports picks are the plays that cleared the threshold, posted in the morning and graded on Betstamp against the close, with every result on the results page.
Run your own entries through the same log. A projection only helps if the price you actually got beats the close. If your projection-triggered bets keep losing CLV, the model isn’t what’s broken. Your timing is.
