KEY TAKEAWAYS
CONTENTS
The favorite-longshot bias is the observed tendency for bets on long odds (longshots) to return less per dollar, on average, than bets on short odds (favorites). Put differently, the price on a longshot usually sits further above its true chance of winning than the price on a favorite does. In practice, that means the sportsbook’s built-in margin is often not shared evenly across the outcomes of a market, and the biggest markup tends to land on the biggest payouts.
This matters because most bettors judge a price by its payout, not by its cost. A +1400 price looks exciting, but the question that matters is how much probability you are actually buying for that payout. Reading odds well means comparing the implied probability of a price with a realistic estimate of the true chance. This article explains where the favorite-longshot bias comes from, shows the math with clearly hypothetical odds, and is careful about what the research does and does not say. It describes a tendency in historical data, never a guarantee about the next bet.
What the Favorite-Longshot Bias Describes
The favorite-longshot bias describes a gap between prices and outcomes: when bets are grouped by their odds, the longshot group loses a larger share of money per dollar staked than the favorite group does. The pattern has been studied extensively in horse racing, where researchers Erik Snowberg and Justin Wolfers summarized it as a case where longshots are overbet while favorites are underbet.
Two vocabulary points keep the idea precise. Overbet means more money is wagered on an outcome than its true chance would justify, which pushes its price shorter than it would otherwise be. Underbet is the reverse. Neither word says the outcome is likely or unlikely to win. They describe the relationship between how much money a price attracts and how often the outcome actually happens.
Researchers disagree about the cause. One family of explanations says bettors enjoy risk and will accept a worse price for a shot at a large payout. Another says people misjudge small probabilities, treating a 5% chance as if it were larger. Snowberg and Wolfers tested both ideas against horse-racing exotic bets and reported evidence in favor of the misperception explanation. A third view holds that the sportsbook itself shapes the pattern by loading more margin onto longshots, because that is where it faces the most uncertainty and the least competition on price.
It also helps to separate this bias from variance. Longshots win rarely, so a longshot bettor sees long losing streaks even when the price is fair. The bias is something different: it says the price itself, not just the luck, tends to be worse on long odds.
How Margin Spreads Unevenly Across Outcomes
Every market priced by a sportsbook contains an overround, the amount by which the implied probabilities add up to more than 100%. If you want a refresher on measuring it across several outcomes, the guide to overround across every outcome in a market walks through the arithmetic. The key point here is that the total tells you the size of the margin but not where it sits.
There are two simple ways to picture the margin. In the even-spread picture, each outcome’s implied probability exceeds its fair probability by the same fraction. In the uneven-spread picture, the fraction is larger on the longshots. Compare the two by measuring the markup, which is the implied probability divided by the fair probability. A markup of 1.05 means you are paying 5% more than the fair price; a markup of 1.33 means you are paying 33% more.
Why would a sportsbook do this? A few reasons are commonly given. Longshot probabilities are harder to estimate precisely, so a wider cushion protects against model error. Bettors tend to be less price-sensitive on big payouts, so the extra cost goes unnoticed. And a small change in a tiny probability looks trivial in the price: moving a longshot from 5% to 6.7% implied probability is a small absolute step but a large relative one. These are explanations offered in the economics literature and by market observers, not facts about any specific sportsbook, and individual books price differently.
The same logic explains why the usual shortcut for removing margin can mislead. The standard no-vig calculation divides every implied probability by the same total, which quietly assumes the margin is spread proportionally. If the margin is actually heavier on longshots, that shortcut leaves the longshot’s no-vig probability too high and the favorite’s too low.
Worked Example: A Four-Outcome Market
Suppose a hypothetical sportsbook lists four outcomes in a single market, such as a small futures list. These prices are illustrative only, not real lines. To measure the markup we also need an assumed fair probability for each outcome, and in the real world nobody knows those exactly, so treat them as a teaching assumption: 50%, 30%, 15% and 5%.
| Outcome | Offered odds | Implied probability | Assumed fair probability | Markup (implied ÷ fair) | Expected return per $100 staked |
|---|---|---|---|---|---|
| A (favorite) | 1.91 (about -110) | 52.36% | 50% | 1.047 | $95.50 |
| B | 3.10 (+210) | 32.26% | 30% | 1.075 | $93.00 |
| C | 6.00 (+500) | 16.67% | 15% | 1.111 | $90.00 |
| D (longshot) | 15.00 (+1400) | 6.67% | 5% | 1.333 | $75.00 |
The four implied probabilities add up to 107.95%, so the overround is about 7.95 percentage points. Read the table from top to bottom: the markup climbs from roughly 1.05 on the favorite to 1.33 on the longshot, even though the same book posted all four prices.
The last column converts that markup into money. A $10 stake on outcome D at 15.00 pays $150 if it wins, which is $140 of profit. If the true chance is 5%, the average payout is 0.05 × $150 = $7.50 per $10 staked, a long-run cost of $2.50. A $10 stake on A at 1.91 pays $19.10, and at a 50% true chance the average is $9.55, a cost of $0.45. The longshot costs more than five times as much per dollar in this illustration.
Now apply the proportional no-vig shortcut by dividing each implied probability by 107.95%. You get 48.50%, 29.88%, 15.44% and 6.18%. Compared with the assumed fair values, the favorite is understated (48.50% versus 50%) and the longshot is overstated (6.18% versus 5%). The shortcut is still useful as an approximation, but it can bias your estimate of the longshot’s true chance upward.
How to Interpret the Bias Without Overreading It
The pattern is most reliable when the market has many outcomes. In the 2021 literature review by Philip Newall and Dominic Cortis, longshot bias was found most often in markets with several possible outcomes, such as horse racing and tennis. For markets with only two outcomes, such as many team-sport moneylines, their review reports that studies more often found the reverse, a bias toward favorites. The authors also note that publication bias and small samples limit what can be concluded, and that they believe the evidence fits both biases being present in the average bettor.
That nuance changes how you should use the idea. With many outcomes the pattern is more plausible. If you are looking at a long list of outright winners, or a futures market with dozens of entries, it is reasonable to expect the longest prices to carry the highest markups. If you are looking at a single two-way game line, the same assumption is far less safe.
It also matters whether the bias is a feature of the sportsbook’s pricing or of bettor behavior. Both can produce the same table. Bettors pile onto longshots, so the book shortens those prices to manage risk. Or the book posts a deliberately wide longshot price because it knows few customers will compare it. From the reader’s side, the practical conclusion is the same: compare the implied probability with your own honest estimate, not with the payout.
Finally, remember that your estimate of a true probability is itself uncertain. If your estimate of the longshot is 5% but it could reasonably be 3% or 7%, the markup of 1.33 is not a precise measurement. Treat the markup as a rough guide to relative cost, not as an exact reading.
Common Mistakes and Misconceptions
Mistake one: treating the bias as a betting system. Learning that longshots tend to cost more does not mean favorites are safe or profitable. In the example, the favorite still carries a 4.5% expected cost. A favorite bet is still a negative-expectation bet if the price exceeds the true chance, and favorites still lose often enough to cause real losses.
Mistake two: assuming every longshot is overpriced. The bias describes averages across many bets. An individual longshot can be fairly priced or even underpriced, and no single result can confirm or refute the pattern. A longshot winning once says nothing about the average.
Mistake three: confusing the bias with the vig. The vig or overround is the total margin; the bias is about how that margin is distributed. Two markets can have the same overround but very different distributions.
Mistake four: carrying the pattern across every market. As the literature review notes, results differ by sport, market size and number of outcomes. Applying a horse-racing finding to a two-team game line is a common overreach.
Mistake five: chasing a longshot to recover losses. Because longshots lose most of the time, using them to win back earlier losses magnifies risk. Never stake money you need for essentials, and never raise stakes to recover a loss.
Where the Bias Shows Up on a Real Odds Board
You are most likely to meet the bias in outright and futures lists, where a handful of contenders are priced short and a long tail of outsiders is priced at very large numbers. Adding up the implied probabilities of the whole list often shows an overround far larger than on a single game line, and much of that excess sits in the long tail.
It also appears when bettors combine legs. Because each leg carries its own margin, stacking several longshots compounds the cost, which the article on how parlay vig compounds on every leg quantifies. A high-payout ticket is attractive precisely because it is built from long prices.
A simple habit helps: convert the price to an implied probability, and ask whether that probability is plausible given what you know about the event. If a +1400 price implies about 6.67%, would you honestly say the outcome happens about one time in fifteen? That question is more useful than the size of the potential payout, and it keeps the decision about price versus probability rather than hope versus fear.
Related Concepts and Next Steps
If the idea of prices drifting away from true probabilities interests you, the next logical step is understanding why prices become accurate at all. The article on market efficiency in sports betting explains how competition and information push prices toward fair values, and where that process falls short. To see how human psychology shapes what bettors choose to back, read about common cognitive biases in sports betting. Together with the margin and probability articles linked above, these form a clear path from reading a single price to judging a whole market.
Frequently Asked Questions
