Explainer · 6 min read
Betting on favourites vs underdogs: what the data shows
"Always back the favourite, it's safer." "No, the money is in the underdogs." Both ideas are common, and both lose over time. This explainer shows what research and five seasons of football odds say about the favourite-longshot bias, how bookmakers spread their margin across prices, and what that means if you look for sure bets instead.
Updated 22 October 2026
Contents
The favourite-longshot bias
In 1949 the psychologist Richard Griffith looked at horse-racing results and noticed a pattern: longshots (big outsiders) won less often than their odds suggested, and favourites slightly more often. Someone betting on longshots therefore lost more per euro staked than someone betting on favourites. This pattern is called the favourite-longshot bias. It has since been found in many markets, including fixed-odds football betting in the UK (Cain, Law and Peel, 2000).
What the bias does not say matters just as much: it does not say favourites make money. It only says that underdog bets cost more on average. Both usually lose; one loses less. Some markets even show the opposite pattern: Woodland and Woodland (1994) found a slight reverse bias in US baseball betting.
The numbers from five leagues
We checked this ourselves with 8,907 matches from the Premier League, Bundesliga, Serie A, LaLiga and Ligue 1, seasons 2021/22 to 2025/26, using the bet365 and Pinnacle prices that football-data.co.uk records before each round. For every price we assumed a €1 bet on that outcome and grouped the results by odds range.
| Odds | Bets | Hit rate | Result bet365 | Result Pinnacle |
|---|---|---|---|---|
| 1.01–1.49 | 1,721 | 75.5% | −0.8% | +0.5% |
| 1.50–1.99 | 3,280 | 56.2% | −3.7% | −1.3% |
| 2.00–2.99 | 5,297 | 39.5% | −5.4% | −2.4% |
| 3.00–4.99 | 11,998 | 25.9% | −6.1% | −3.9% |
| 5.00–9.99 | 3,777 | 14.0% | −15.1% | −13.6% |
| 10.00 and above | 648 | 6.0% | −23.8% | −21.7% |
Bets and hit rate refer to the bet365 prices; a few matches are missing for Pinnacle. The pattern is clear: the higher the odds, the bigger the loss per euro staked. At odds of 10 and above almost a quarter of the stake was lost; on clear favourites only about 1%.
Two words of caution. First, the underdog figures swing a lot. In 2021/22 the 10.00+ range would even have returned +22.7%; in the four seasons after that it lost between 14% and 56%. With a few hundred bets, a handful of shocks decide the result. Second, the "+0.5%" for Pinnacle favourites is not a profit signal: with around 1,500 bets the statistical uncertainty is about ±1.5 percentage points, well above the figure itself. More on this in betting variance.
How bookmakers spread the margin
Every price contains the bookmaker margin. In our data it averaged 5.5% per match (1X2) at bet365 and 3.1% at Pinnacle. The table above shows that this margin is not spread evenly over the three outcomes. A simplified example:
- Fair odds: 1/0.70 = 1.43, 1/0.20 = 5.00, 1/0.10 = 10.00
- Offered odds with the margin loaded unevenly: 1.40 / 4.60 / 8.00. Together 1/1.40 + 1/4.60 + 1/8.00 = 105.7%, so a 5.7% margin.
- Expected result per €1: favourite 0.70 × 1.40 = 0.98 (−2%), draw 0.20 × 4.60 = 0.92 (−8%), outsider 0.10 × 8.00 = 0.80 (−20%)
The same overall margin hits you very differently depending on what you back. How to turn odds into probabilities is explained in implied probability.
Why the bias exists
Research mainly points to three reasons:
- A taste for long odds. Many people prefer small stakes on big payouts, much like a lottery ticket. Demand for outsiders pushes their prices down.
- Small probabilities are overestimated. Snowberg and Wolfers (2010) show with US horse-racing data that misjudged probabilities explain the pattern better than a pure love of risk.
- Protection against better-informed bettors. In Shin's model (1993), bookmakers protect themselves against customers with inside information by charging more margin on outsiders, where an unexpected winner does the most damage.
For you as a bettor the reason matters less than the consequence: the markets of major leagues are fairly efficient, and the bias is known and priced in. See how efficient betting markets are.
Why "always favourites" is not enough
Favourites lose the least, but they still lose. At bet365 the best range came out at −0.8%. In Austria the betting fee can come on top: if a bookmaker charges 5% on top of the stake, −0.8% quickly becomes around −5.5%.
A 75% hit rate feels safe, but it is not. One in four favourites under 1.50 did not win. At odds of 1.30 you need three winners to cover one loser (3 × €0.30 = €0.90 profit against a €1 loss), and even then you are 10 cents short. Why "safe bets" don't exist is covered in can you win at sports betting long term?
Why "always underdogs" is expensive
Underdog strategies live on the hope of the big hit, and sometimes it comes. But on average, odds of 5.00 and above were the most expensive way to bet in our data, at −13% to −24%. On top of that come long losing runs: with a 6% hit rate, 30 losses in a row are not unusual (0.9430 ≈ 16%).
Anyone looking for underdogs with "value" needs a model that is better than the market. That is hard, as building your own betting model shows.
What it means for sure bets
In a sure bet you cover every outcome at different bookmakers. Whether the favourite or the underdog wins then makes no difference to the result, as long as all bets stand at the planned odds. The bias still matters, though, when you choose which sure bets to take:
- Price gaps are bigger at long odds. Because bookmakers load a lot of margin there, their outsider prices differ more (9.50 at one, 11.00 at another). That is why many sure bets include an outsider leg.
- Stakes become lopsided. At 1.25 and 6.00 (together 96.7%), €300 splits into €248.28 and €51.72. Both outcomes return about €310.33, so roughly +€10.33.
- The void risk is uneven. If the outsider bet is voided as an obvious price error, the big favourite bet (€248.28) is left on its own. With balanced prices such as 2.10 and 2.05 the leftover risk would be much smaller. What to do then is explained in voided bets and palpable errors.
For beginners this means: sure bets with prices not too far apart are usually calmer. A sure bet with odds of 15 on one side is just as good on paper, but more fragile if something goes wrong.
Frequently asked questions
Is it better to bet on favourites or underdogs?
Neither wins on its own over time. Favourites lost less per euro staked in our data, about 1 to 4 percent at bet365, while odds of 5.00 and above lost 13 to 24 percent. Both remain below zero after the bookmaker margin.
What is the favourite-longshot bias?
It is the observation that outsiders win less often than their odds suggest and favourites slightly more often. It was first described by Richard Griffith in 1949 for horse racing and has since been found in many betting markets, including football.
Are bets on big favourites safe?
No. Even at odds below 1.50, about one in four favourites did not win in our data. At odds of 1.30 you need more than three winners to cover a single loss.
Does the favourite-longshot bias matter for sure bets?
Not for the outcome, because a sure bet covers every result. It matters for selection: big price gaps often appear on outsiders, which makes stakes lopsided and leaves a larger position exposed if the outsider bet is voided.
Sources
- Griffith, R. M. (1949): Odds adjustments by American horse-race bettors. American Journal of Psychology 62(2), 290–294.
- Shin, H. S. (1993): Measuring the incidence of insider trading in a market for state-contingent claims. Economic Journal 103(420), 1141–1153.
- Woodland, L. M. and Woodland, B. M. (1994): Market efficiency and the favorite-longshot bias: the baseball betting market. Journal of Finance 49(1), 269–279.
- Cain, M., Law, D. and Peel, D. (2000): The favourite-longshot bias and market efficiency in UK football betting. Scottish Journal of Political Economy 47(1), 25–36.
- Snowberg, E. and Wolfers, J. (2010): Explaining the favorite–long shot bias: is it risk-love or misperceptions? Journal of Political Economy 118(4), 723–746.
- football-data.co.uk: results and odds data (Premier League, Bundesliga, Serie A, LaLiga, Ligue 1, seasons 2021/22 to 2025/26; our own analysis)
As of October 2026. Past results do not predict future results. The odds data are closing-round prices as recorded by football-data.co.uk.
Not betting, legal or tax advice. 18+. Betting involves risk and no profit is guaranteed: odds change, bets get voided and accounts get limited. Only bet money you can afford to lose. Our responsible gambling page lists free help, and the disclaimer applies.