What +EV Means in Player Props (and What It Doesn't)
"+EV" gets thrown around as if it were a synonym for "good bet." It is narrower than that, and a little humbler. This article explains what expected value actually is, how a tag like the +EV badge on ArchivePicks is built, and where the idea stops being useful.
Expected value in one paragraph
Expected value (EV) is the average result of a bet if you could place it over and over. Suppose a bet pays 2.00 in decimal odds (+100 in American odds) and wins with a true probability of 51%. Over a long run you would expect to get back 0.51 x 2.00 = 1.02 for every 1.00 staked, so the EV is +0.02, or +2% of the stake. If the true chance were 49%, the same price would be about -2%.
The catch is in the phrase "true probability." Nobody knows it for a player prop. Every EV figure is therefore a statement about a price measured against somebody's estimate of the odds, and the estimate is where all the uncertainty lives.
Where the estimate comes from
There are two common ways to get one.
- Your own model or judgment. You decide a player clears 64.5 yards 58% of the time and compare that with the price. The odds calculators include an expected value tool that works this way: you enter the price and your own win chance. A positive result only means your estimate beats the price, not that the bet will win.
- The market's view. Take the prices other books post on the same number, remove their margin, and treat the result as the market's estimate of the chance. A price that pays more than that estimate implies is "+EV against the market."
ArchivePicks uses the second approach, because it needs no opinion about any player. It does not predict how anyone will perform.
Removing the margin
A book's two prices on one prop add up to more than 100%. A -115 over and -105 under, for example, imply roughly 53.5% and 51.2%, about 104.7% together. The extra is the margin, often called the vig or juice. To get a fair chance, you scale the two numbers back so they total 100%. In that example the fair over chance is about 51%. The glossary has short definitions of implied probability and related terms, and the calculators page has a no-vig tool that does this for you.
A worked example (hypothetical numbers)
These numbers are made up to show the arithmetic.
Three books post the same receiving-yards line. Two of them, with their margin removed, put the over at 51% and 52%. The third, call it Book C, has the same number but pays +100 on the over and -130 on the under.
- Average the other books' fair over chance: (51% + 52%) / 2 = 51.5%.
- Book C's over pays 2.00 in decimal terms. EV = 0.515 x 2.00 - 1 = +0.03, or about +3% of the stake.
- Book C's under pays about 1.77 in decimal terms. The under's fair chance is 48.5%, so EV = 0.485 x 1.77 - 1 = about -14%.
So in this example Book C's over is the one flagged, and the under is not. The flag says Book C's over price is more generous than the rest of the market implies. It does not say the player will go over.
How ArchivePicks builds the tag
The tag follows the same steps, with some guardrails:
- Each book's two prices on the same number are turned into a fair over chance by removing the margin proportionally.
- A book's price is judged against the average of the other books, never against itself.
- At least two other books must price both sides of that number. One book's line is an opinion, not a market, so a number with fewer than three books on it is never flagged.
- The edge has to be at least 1% to show. Smaller than that is rounding noise.
- The board shows the side (O or U), the percentage and the book. A "+EV price" filter on the NFL and NBA boards lets you list props that carry the tag and sort by the edge.
The hover text on the tag says the same thing: the price beats the market's view, not that the bet wins. The calculators use the same formulas.
What +EV cannot tell you
- The market can be wrong together. If every book shades a number the same way, an average of them inherits the error.
- Stale prices look like edges. Books refresh at different speeds. A price that has not caught up with news can look generous for a moment and be gone before you could use it. Prices on ArchivePicks refresh every couple of minutes and can already be out of date. See comparing a prop across books.
- One result proves nothing. A +3% edge still loses most of the time on a near coin-flip prop, and even a long run of results is a small sample. See sample size and hit rates.
- It ignores the player. A market-based tag knows nothing about injuries, role or matchup. That context is what the hit rates, head-to-head and defense data on the player page are for.
FAQ
Is +EV the same as a pick? No. ArchivePicks does not give picks or projections. The tag describes one price relative to other prices, nothing more.
Why is there no +EV tag on a prop? Either no price clears the 1% bar, fewer than three books post that number, or the sport is not one where the tag runs. It is limited to NFL and NBA props.
Does a +EV tag mean the bet wins more often than it loses? No. It is about the payout versus the estimated chance. A price can be +EV and still lose most of the time.
ArchivePicks is for information only. It does not take bets or tell you what to bet. Must be of legal age. Gambling problem? Call 1-800-GAMBLER.