Compare n' Bet™
Compare n' Bet™
Aggregate CLV tells you whether you're sharp on average. It does not tell you where the edge comes from. Breaking CLV down by sport, market, and timing is how you find out which parts of your strategy are real and which are noise.
You've been tracking CLV for a year. Your overall CLV is +1.2% across 800 bets. The math says you're a slightly sharp bettor, edge is real, keep going. But that single number is hiding everything interesting.
What if your NFL CLV is +3.2% and your NBA CLV is −0.4%? The aggregate looks fine but you have a real strategy in one sport and a coin flip in the other, and the coin flip is dragging down a much better edge. What if your CLV on player props is +4% and your CLV on spreads is +0.1%? Same situation. What if your CLV on bets placed within an hour of game time is +2.5% and on bets placed two days early is −1.5%? You'd want to know that.
Aggregate CLV gives you a yes/no answer to the question "do I have an edge." Attribution gives you the answer to "where is the edge actually coming from," which is the question you need to act on if you want to improve.
CLV per bet, in implied probability:
For most retail bettors, the per-bet standard deviation of CLV is in the 0.02 to 0.03 range (2 to 3 percentage points). At 100 bets, the standard error of mean CLV is around 0.0025, or 0.25 percentage points. That's small enough to detect a 1% true CLV at high confidence in a few hundred bets.
The mechanics are simple. For each bet, tag it with categorical metadata: sport, market type, bet timing, etc. Compute average CLV within each tag separately:
The catch: as you split your sample into more categories, the per-category sample size shrinks, and the standard error grows. A category with only 30 bets has SE around 0.005 (0.5 percentage points), which means you can't detect a CLV of 1% with confidence. A category with 100 bets pushes SE down to 0.003. A category with 300 bets gets SE to 0.0017.
Practical implication: don't slice too finely until you have meaningful sample. Start with 3 to 5 broad categories (sport groups, major market types) and only refine when you've accumulated enough bets per category to support real conclusions.
The dimensions that produce the most useful attribution for most bettors:
Sport. NFL, NBA, MLB, NHL, soccer, tennis, etc. The most basic split. Almost every retail bettor has different CLV across different sports, often dramatically different.
Market type. Spread, total, moneyline, player prop, team prop, futures. Edge often concentrates in one or two market types and is absent or negative in others.
Bet timing. Days/hours before kickoff. Bettors who shop lines closer to game time often capture better numbers. Bettors who bet early are sometimes giving up CLV by betting before the line has matured.
Book. Which book each bet was placed at. CLV by book tells you which operators give you the best prices on your typical bet types. Useful for deciding where to fund larger.
Bet type/structure. Single bet, parlay, SGP, hedge. Multi-leg bets have different CLV characteristics than singles, and understanding the difference matters.
Stake size. Small bets vs your average bet vs your big swing bets. If your CLV on big bets is worse than your CLV on small bets, you're systematically betting bigger when you're less confident, which is exactly the wrong direction.
Hypothetical bettor with 800 bets across one year, split between NFL and NBA. Aggregate CLV is +1.2%. Break it down:
The aggregate "+1.2% CLV" was hiding a much more useful story. NFL spreads and NBA props are real edge sources. NBA spreads contribute nothing positive. NFL totals are weak. The actionable read: this bettor should drop NBA spreads and probably dial down NFL totals, doubling down on their two real edges.
If they continued betting all four categories at equal volume, the aggregate edge stays around +1.2%. If they shifted volume toward the two strong categories and away from the weak ones, the aggregate edge improves to something closer to +2.2%, almost double.
If you split your bets into 20 categories, even if you have zero true edge in any of them, you'll see one or two with p-values below 0.05 just by random chance. This is the multiple comparisons problem and it's real for anyone slicing CLV finely.
Two ways to handle it:
Practically: if you split into many categories and find one or two with promising CLV but the rest show nothing, the promising ones are likely false positives. Real edges show up across multiple related categories or are large enough that even Bonferroni-corrected significance survives.
Rough benchmarks based on observed retail bettor data:
These thresholds are sport-independent. A +2% CLV bettor in NFL is in the same broad band as a +2% CLV bettor in tennis. The implied long-term ROI is similar regardless of sport.
The minimum viable structure for CLV attribution:
Don't try to compute closing lines manually. Subscribe to a data source or use a tool that pulls closing lines automatically. Manual closing-line entry is error-prone, time-consuming, and biased. You'll subconsciously round in directions that flatter your numbers.
Update the spreadsheet weekly or after each session. Don't update only at the end of the month or year. The point of the analysis is to inform real-time decisions about where to spend your betting attention.
Slicing too finely. Splitting 200 bets into 15 categories produces 15 noisy estimates and zero useful conclusions. Start with 3 to 5 broad buckets and only refine when each bucket has 100+ bets.
Cherry-picking the categories that look good. If you split bets 20 ways and report only the categories with good CLV, you're guaranteed to find some signal even if there isn't one. Always report all categories or at least account for multiple comparisons.
Confusing per-bet CLV with per-dollar CLV. If you bet bigger on some types of bets than others, average CLV per bet doesn't tell you about your dollar-weighted edge. Compute both, especially if your stake size varies meaningfully.
Not accounting for correlated CLV within a session. Multiple bets on the same Sunday's NFL slate have correlated outcomes (and correlated CLV signals, since they share information). Your effective sample size is smaller than your raw bet count when bets cluster.
Treating CLV as the only metric. CLV is a leading indicator. Realized ROI is the real bottom line. Use CLV to predict and diagnose. Use ROI to confirm. They should agree over a large enough sample.
Aggregate CLV is the basic check. Attribution is what tells you what to do differently. The bettor who runs +1.2% aggregate but has a +2.4% NFL spread edge and zero NBA edge should be betting a lot more NFL spreads and a lot less NBA. The bettor who runs +1.5% aggregate but loses CLV on bets placed two days early should be betting closer to game time. The bettor whose big bets have worse CLV than their small bets has a sizing problem to investigate.
None of this shows up in the aggregate. All of it shows up in attribution. If you're going to track CLV at all, attribute it. The extra work of tagging bets is small compared to the upside of knowing which strategies are real and which aren't.
The content on Compare n' Bet is published for educational and informational purposes. By reading this guide you acknowledge:
If you or someone you know has a problem with gambling, the National Council on Problem Gambling helpline is 1-800-522-4700 (US) or visit ncpgambling.org. International readers can find local resources at gamblingtherapy.org.