Find Value Bets: Check Any Price Against the Sharp Market
You know the theory. Here is how to check a price against one that already removed its margin.
A value bet pays more than the true chance of winning, nothing more mystical than that.
Probability times odds, minus 1: that is the whole formula, and it only works if the probability is trustworthy. The steadiest source is a sharp bookmaker's own price with its margin stripped out, treated as the fair number to test everything else against. On a market priced 1.60, 4.00 and 6.50, the true chance behind that 6.50 outsider lands close to 14.5 to 15 percent once the margin is removed, not the 15.4 percent its raw price implies. Test whatever price you are actually offered against that fair number before trusting a hunch, and expect the method used to remove the margin to matter more on a long price than a short one.
How to actually find a value bet, step by step
Expected value is simple math: the probability you believe an outcome carries, times the odds on offer, minus 1. If that number comes out positive, the price is value, in theory. The whole exercise falls apart on the word "believe", because most bettors do not have a model producing probabilities, they have a hunch dressed up as a percentage. The workaround that professional bettors actually use skips the guessing: borrow the probability implied by a sharp bookmaker's own price, strip out the book's margin, and treat what is left as the fair number to test everything else against. That sharp-book price is basically the answer to which bookmaker actually has the best odds in the first place: a book pricing tighter to the true chance of an outcome, instead of every line carrying the same padded margin.
Four steps, no software required. Pick a reference market from a book known for tight, fast-moving prices. Remove its margin to get a fair probability for each outcome. Multiply that fair probability by whatever price you are actually being offered elsewhere. Subtract 1. Positive means the price you found beats what the sharp market thinks is fair. Negative means it does not, whatever your gut says about the match.
Where that fair number is supposed to come from
A raw price never equals a true probability, because a bookmaker builds in a cut. Add up the implied probabilities of every outcome in a market (1 divided by each price) and the total comes to more than 100 percent; the excess is the margin, sometimes called the overround. Removing it, known as devigging or computing no-vig odds, is where the actual disagreement lives: there is more than one honest way to decide which outcome absorbs how much of that margin. The margin calculator runs that first step, the plain overround, on any market you enter; this page goes a step further and tests a specific price against it.
Three assumptions do most of the work in practice. Spread the margin in proportion to each price, so bigger favorites lose a slightly bigger slice. Spread it as a flat, equal amount off every probability instead. Or solve for a single power that shrinks every price by the same ratio until the total lands on exactly 100 percent. None of them is correct in any provable sense; they are three different guesses about how a sharp book actually builds its price, and what actually makes a book sharp in the first place is a separate question worth its own read.
We treat the proportional method as the default on this page, because it tracks how margin tends to get reported: shaded a little more toward outsiders than favorites, not because it hands out the friendliest number on any given bet. Use whichever assumption you trust more. The point is picking one and staying consistent with it, not switching formulas until one flatters the bet you already wanted to place.
Same devigged market, three different verdicts
Take a three-way market priced 1.60, 4.00 and 6.50. The implied probabilities add up to about 102.9 percent, so there is roughly 2.9 percent of margin sitting in that market. Strip it out under each of the three methods and the fair probability behind that 6.50 outsider moves around: about 14.95 percent under the proportional method, 14.50 percent under the power method, 14.42 percent under the additive method. Small differences in probability, but they compound once you test a price against them.
Say a local book offers 7.00 on that same outsider. Multiply each fair probability by 7.00 and subtract 1: the proportional method calls it plus 4.7 percent value, the power method plus 1.5 percent, the additive method barely plus 1.0 percent. Same bet, same price, three different verdicts depending on an assumption about margin, not about the match. Run the same exercise on a near-even market instead (prices of 1.80 and 2.10, a price of 2.30 tested against the 2.10 side) and the gap nearly disappears: plus 6.15 percent, plus 5.87 percent and plus 5.74 percent, a spread of four tenths of a point instead of more than three and a half. The method barely matters on a coin flip. It matters a lot once the price gets long, which is exactly the outsider case above and exactly what the chart below makes visible.
Expected value by margin removal method
Published formula, calculatedView the data as a table
| Margin removal method | Outsider price: 6.50 reference, 7.00 tested | Near-even price: 2.10 reference, 2.30 tested |
|---|---|---|
| Proportional | 4.70% | 6.15% |
| Additive | 1.00% | 5.87% |
| Power | 1.50% | 5.74% |
The reference market and the outsider price above match one of the preset buttons on the calculator further down this page, labelled "Try a 3-way market", so you can load the same numbers and switch methods yourself instead of trusting the paragraph above at face value.
Pick the wrong reference price and the whole exercise collapses. A book padding every line by 8 percent because it is built for recreational turnover is not a usable reference, however confidently it prices a match, and neither is an average taken across several random odds-comparison sites, since averaging soft prices together does not manufacture a sharp one. A reference price that is a day old has usually drifted from wherever the market has since moved too, so the closer to kickoff you pull it, the more it is actually worth trusting.
One positive number does not prove you have an edge
Finding a positive number once proves almost nothing. A 3 percent edge at odds of 2.00 (a 51.5 percent chance of winning, priced at 2.00) is a real advantage over a long enough run, but the standard error on your results after just 100 bets is about 10 percent, wider than the edge itself. After 1,000 bets that error narrows to roughly 3.2 percent, still close enough to the edge to leave real doubt. Only somewhere around 4,400 bets does a genuine 3 percent edge sit two standard errors clear of zero, the rough point where most people would call a result more than luck.
(Horse racing bettors run into this earlier than most, since a single card can hand you a dozen qualifying prices in one afternoon. That speed helps you reach 4,400 bets faster, and it is exactly as much of a curse for anyone tempted to judge an edge after one good weekend.)
Ten winning weekends in a row can happen to someone with no edge at all, and ten losing weekends can happen to someone with a genuine one. Whether the underlying math even supports betting at real volume is its own separate question, one that the numbers on whether betting actually pays off answers with more room than this page has.
A positive number on a screen and money in an account are two different things, separated by variance and by whoever is on the other side of the bet. Keep stakes small and flat while the sample builds; flat stakes tell you about the edge, oversized ones mostly tell you about your nerves. And do not expect every book to sit still if you are actually right. Bookmakers built on recreational turnover watch for exactly this pattern, consistent value-priced stakes on the same kinds of markets, and why a winning account gets capped in the first place is usually a pricing decision on their side, not a rules violation on yours. None of this is a promise that the edge is real or that it survives contact with your own discipline. It is arithmetic on a price, nothing more.
Test your own price against the sharp market
Enter a reference market of two or three outcomes, the price you actually want to test, and a method, and the value bet calculator below runs the exact same subtraction used in the examples above, live, including what the other two methods would have said about the same price.
A positive percentage in the EV column is the case for the bet, measured against whichever method you picked at the top, nothing more. Check the two other rows in the same table before acting on it. If all three agree the price falls short of fair, that verdict is worth trusting more than a single method that happens to say what you wanted to hear.
Do you actually need to pay for a scanner?
Software exists that automates exactly the comparison above across hundreds of markets a minute, watching a reference book and flagging any price elsewhere that clears it. What that software cannot do: stop a book from limiting your account once your stakes start looking like exactly what they are, or stop a price from moving between the moment a scan flags it and the moment your bet actually goes through. Automating the comparison does not remove either risk, it just finds more chances to run into them, faster.
Run the calculation above by hand for a few weeks before reaching for a paid subscription. The three-line arithmetic any scanner is built on is free, and doing it yourself for a month tells you whether your reference book is actually showing a gap in your sport before you pay anyone to automate it.
Frequently asked questions
How do I find value bets without a betting model?
Borrow a probability instead of guessing one. Take a market from a bookmaker known for tight, fast-moving prices, remove its margin, and use what is left as the fair chance of each outcome. Multiply that fair chance by any price you are actually offered elsewhere and subtract 1; a positive result is the case for value.
How do you calculate the value of a bet?
Expected value equals the true probability of an outcome, multiplied by the decimal odds on offer, minus 1. On a coin-flip market priced at 2.10, a true 50 percent chance gives 0.50 times 2.10 minus 1, or plus 5 percent. The formula never changes; what changes is how much you can trust the probability you fed into it.
What does removing the bookmaker's margin (devigging) actually mean?
Add up 1 divided by every price in a market and the total comes to more than 100 percent; the excess is the margin, sometimes called the overround. Devigging, or computing no-vig odds, spreads that excess back out of the prices so what remains behaves like a real probability. There is a page where you can work out a market's overround yourself on any prices you enter.
Does the method used to remove the margin change the answer?
Barely, on a near-even price. A lot, on a long one. Across the three common assumptions (proportional, additive and power), a coin-flip market moves by a few tenths of a percentage point either way; a genuine outsider price can swing by three or four points depending on which method you picked, which is exactly the gap the chart on this page measures.
How many bets do I need before I know I have a real edge?
More than most people run in a season. A real 3 percent edge at odds of 2.00 still carries a standard error of about 10 percent after 100 bets, 3.2 percent after 1,000, and only settles down to about 1 percent after 10,000. Roughly 4,400 bets is the point where a genuine 3 percent edge sits clearly clear of zero rather than inside the noise.
Do I need to pay for a value betting scanner?
Not to start. Paid software automates the same comparison across many markets at once, which saves time once you already trust your reference book, but it does not remove the two risks that matter most: a bookmaker limiting your account once your pattern of staking gives you away, and a price moving between the moment it looks like value and the moment your bet is actually placed.
Is value betting different from arbitrage and from closing line value?
Yes, on both counts. Value betting tests one price against a fair number before you bet; betting every outcome instead of picking one side is a different approach built to lock in a return regardless of the result. closing line value is neither of those: it checks, after the fact, whether the price you actually took held up against where the market closed.