Favourite-Longshot Bias: What Academic Research Reveals

Favourite-longshot bias represents one of the most extensively documented phenomena in betting market research. Decades of academic study, across millions of races and multiple countries, consistently find the same pattern: longshots are overbet relative to their true probability of winning, while favourites are underbet. What this accumulated evidence tells us about where value really lies challenges assumptions that many punters never question.

The bias matters because it directly affects expected returns. If longshots consistently deliver worse value than their odds suggest while favourites deliver better value, betting strategy should respond. Yet casual punters routinely chase big-priced winners while dismissing short-priced selections as “bad value”—the opposite of what research indicates.

This guide explains what the favourite-longshot bias is, examines the research evidence from UK racing specifically, explores why the bias exists, and translates findings into practical betting implications.

What Is Favourite-Longshot Bias?

The favourite-longshot bias describes a systematic pattern in betting markets: the odds on longshots don’t fully reflect how rarely they win, while the odds on favourites don’t fully reflect how often they win. In statistical terms, expected returns decrease as odds lengthen. Backing every favourite produces better returns—or smaller losses—than backing every outsider.

Consider what fair odds would look like. If a horse has a 10% chance of winning, fair odds are 9/1. If a horse has a 2% chance, fair odds are 49/1. But research shows that horses at 9/1 win more often than 10% of the time, while horses at 49/1 win less often than 2% of the time. The market systematically misprices both ends of the spectrum—underestimating favourites, overestimating longshots.

The bias doesn’t mean favourites always win or longshots never win. Individual race outcomes vary enormously; longshots score at every meeting. The bias describes average returns across thousands of bets. A punter blindly backing every favourite would lose money to the bookmaker’s margin, but would lose less than a punter blindly backing every outsider at the same stakes.

The effect intensifies at extreme odds. Horses at 2/1 or 3/1 show modest bias; the mispricing is relatively small. Horses at 33/1 or 50/1 show substantial bias; the gap between implied probability and actual winning frequency is significant. The longer the odds, the worse the expected value—a pattern that directly contradicts the intuition that “bigger odds mean bigger value.”

Importantly, the bias persists despite being well-documented. Economic theory suggests that known inefficiencies should be arbitraged away—yet the favourite-longshot bias has been observed for decades across multiple betting markets. Its persistence suggests structural causes that knowledge alone doesn’t eliminate.

The Research Evidence

Academic research on betting markets stretches back to the 1940s, with the favourite-longshot bias identified early and confirmed repeatedly since. The consistency across studies, timeframes, and geographies provides confidence that the bias reflects genuine market structure rather than statistical noise.

UK-specific evidence comes from comprehensive studies of British racing. Research published in Management Science examined over 1.3 million horse starts across 127,313 races at 38 British tracks from 1994 to 2018. This enormous sample size—covering essentially the entire population of British racing over nearly 25 years—provides robust statistical power. The findings confirmed the favourite-longshot bias persists in British racing, though researchers noted the effect has weakened compared to earlier decades.

Earlier foundational work by Ziemba and Hausch established benchmark findings. Their research documented returns at different odds levels, finding that bettors received approximately 65% return at odds between 10/1 and 18/1, dropping to roughly 28% return at odds of 18/1 and above. These figures represent catastrophic value destruction—the longer the odds, the greater the loss rate.

At extreme odds, the bias becomes dramatic. Research has shown that horses at 100/1 have actual winning probabilities closer to 700/1—returning approximately 14 cents per dollar bet. The implied probability from 100/1 odds (roughly 1%) vastly overstates true chances. Punters attracted to huge potential payouts are systematically disadvantaged.

The bias has weakened over time, particularly since betting exchanges and online markets increased competition. Research by Smith and Vaughan Williams found the effect declining after 2000 compared to previous decades, likely due to sharper market pricing driven by professional participation and exchange efficiency. But the bias persists—it’s reduced, not eliminated.

International evidence shows similar patterns. American racetracks, Australian racing, and European markets all demonstrate favourite-longshot bias, suggesting the phenomenon reflects fundamental aspects of how humans bet rather than specific market structures.

Why the Bias Exists

Several psychological and structural explanations account for the bias’s persistence. Understanding why it exists helps punters recognise the forces that might affect their own betting behaviour.

The dream factor draws punters toward longshots. A £5 bet on a 50/1 shot could return £250; the same bet on a 2/1 favourite returns £15. The psychological appeal of life-changing returns from small stakes attracts money to longshots that pure probability calculations wouldn’t justify. Punters pay a premium for the entertainment value of dreaming about big wins.

Overconfidence in selection ability inflates longshot betting. If you believe your form analysis has identified a genuine 10/1 chance available at 33/1, the apparent value is irresistible. But this assessment requires your analysis to significantly outperform the market’s collective wisdom—a demanding standard that most punters overestimate their ability to meet. The confidence that makes longshots appealing is often misplaced.

Risk preferences may differ from rational expectation. Some punters genuinely prefer high-variance outcomes—they’d rather have a small chance of a large win than a larger chance of a small win, even at worse expected value. This preference is irrational in pure mathematical terms but reflects legitimate psychological variation.

Media and cultural factors reinforce longshot appeal. Racing coverage celebrates surprise winners; “shock 100/1 winner” makes headlines while “favourite wins” doesn’t. The availability heuristic means memorable longshot victories loom larger in memory than forgotten favourite wins, distorting perception of how often each outcome occurs.

Bookmaker margins are typically higher on longshots, compounding the bias. The same percentage margin on 2/1 versus 50/1 produces different absolute effects; structural pricing compounds the psychological factors pushing money toward outsiders.

Practical Implications for Bettors

The research carries direct implications for betting strategy. The favourite-longshot bias suggests that short-priced horses often offer better value than their odds appear to indicate, while long-priced horses offer worse value than their attractive returns suggest.

This doesn’t mean backing every favourite is profitable—bookmaker margins ensure aggregate losses regardless of strategy. But it does mean dismissing favourites as “bad value” simply because of short prices is wrong. A 2/1 shot that wins 35% of the time offers better expected returns than a 33/1 shot that wins 2% of the time, even though both are “correct” odds after margin.

Extreme caution is warranted on extreme longshots. Horses at 33/1 and above face severe negative expectation based on the documented bias. Unless your analysis provides genuine edge beyond market consensus—and honestly assessing whether it does—systematic longshot betting destroys bankrolls.

Place betting on shorter-priced horses may offer value that win betting on longshots doesn’t. The bias research focuses on win markets; place markets may show different characteristics, and horses that are sound each-way propositions at mid-range odds might outperform longshot win bets.

The bias has weakened but not disappeared. Modern markets are sharper than historical ones; the edge from simply backing favourites has compressed. But the fundamental pattern persists, and punters should incorporate it into their assessment of where value might actually exist.

The favourite-longshot bias provides empirically-grounded guidance for betting strategy: value is more likely to exist at shorter prices than the appeal of big wins suggests, and longshots systematically underperform their apparent attractiveness.

For the complete framework on identifying value in betting odds, see our comprehensive value betting guide. And for specific guidance on when odds-on selections offer genuine value, our guide to odds-on betting addresses the shortest end of the market spectrum.