Favourite and Longshot Bias

There’s an irony to an online article beginning with a warning about the dangers of believing what you read online.

Nonetheless, when it comes to the theory of favourite-longshot bias, there is a lot of bunkum written online, with many sites misrepresenting what it is, and its likely explanations. This is partly because many of those articles appear on the websites of fixed-odds bookmakers, who have a stake in representing favourite-longshot bias with, well, bias.

This now adds a second level of irony.  After all, this article appears on the website of a betting exchange. Make of that what you will.

The term favourite-longshot bias is used, interchangeably, to describe two different, although related, phenomena.

The first is the statistically observed tendency for punters to prefer betting on longshots as opposed to favourites.

The second is the tendency of fixed-odds bookmakers – recognising this punter bias – to price up events disproportionately, offering lower odds than is statistically fair on longer-priced outcomes.

For example, in a two-possible-outcomes tennis match, if one player has a 70% chance of winning and the other a 30% chance, their respective fair odds would be \$1.43 and \$3.33. (Divide 1 by any percentage to convert into odds: e.g. 1 ÷ 0.70 = \$1.43).

A fixed-odds bookmaker needs to make a profit, though, and so builds in a margin to each outcome, reducing the odds they offer to punters. In the theoretical example above, they might add a 2% margin to each outcome, resulting in odds of \$1.39 (72%) and \$3.13 (32%). This would mean that, provided punters bet on each outcome proportionally to its chance of occurring, the bookmaker would make a profit, no matter what the outcome.

Sometimes, though, recognising that punters are more likely to bet on bigger-priced outcomes, fixed-odds bookmakers will add a bigger margin to the less-likely of the two outcomes. In the example above, then, the odds become \$1.40 (71%) and \$3.00 (33%).

For the average recreational punter, whether this phenomenon exists is not especially important. They are likely losing money in the long-run anyway, don’t intend to change their betting habits to alter their profitability, and so the favourite-longshot bias means only that they will be giving a little more of their money to fixed-odds bookmakers, a little more quickly.

For the would-be profitable punter, though, the existence, or otherwise, of favourite-longshot bias, raises some questions.  Does it exist? If it does, how should you alter your betting behaviour to account for it? Is it an example of another cognitive bias that you need to be aware of more generally in your punting?

Does it exist?

There is a large body of peer-assessed academic work that establishes the existence of favourite-longshot bias in a variety of sports betting markets.

What is less clear is whether the existence of the phenomenon is explained by fixed-odds bookmakers simply reacting to market forces – shortening the odds on longer-priced outcomes as more people bet on them, whilst lengthening the odds on the less-popular shorter-priced outcomes – or whether they are consciously pricing-up markets knowing that the bias exists within their clients and wanted to take advantage of it.

There is some anecdotal evidence of the latter. Bookmakers’ odds for large-field horse races like the Caulfield Cup will often have many horses priced between \$51.00 and \$101.00 that, on examining their adjusted ratings, should be a significantly bigger price still.

What’s also interesting is whether the emergence of betting exchanges has reduced the bias.  Examining the academic evidence (there are many excellent summaries available, such as that provided by Snowberg and Walters (2010)), much of it is historical, with large bodies of data referenced dating prior to 2000.

Some Betfair data, meanwhile, suggests that favourite-longshot bias does not exist on the exchanges. In one small study, carried out in 2005, which looked at 65,000 results for British racing, horses seemed to win at a rate predicted by their odds. \$1.50-shots won two-thirds of their races, \$2.00-shots won half, and \$100.00 won one per cent of their races.

The data showed a miniscule tendency for the very shortest priced horses to win more regularly than expected, and the very longest to win less regularly than expected. But this is almost certainly explained by the minimum and maximum odds allowed on Betfair: some horses should be shorter than \$1.01 and some longer than \$1000.00, but the system doesn’t allow it.

This data, however, is from one sport, at one point of time, and cannot be claimed to be conclusive. It does suggest, though, that on betting exchanges, the mechanism of peer-to-peer competition to back and lay at the best available odds means that any favourite-longshot bias will be challenged.

Do you need to worry about it?

If you bet with fixed-odds bookmakers, yes. Ensure that you have a clear method of assessing the fair odds for any outcome and only bet when you are confident the odds are in your favour. There are numerous articles elsewhere on this site about value betting to help you do this.

If you only bet on Betfair, then there is, perhaps, less reason to worry. But it’s worth remembering that, although the mechanism of a betting exchange may eradicate the bias statistically, the cognitive tendency to accept poor odds on bigger-priced outcomes will still detriment the profitability of many individuals.

Much of the academic work on favourite-longshot bias has sought to explain why punters will accept relatively poor odds on less-likely outcomes.

A popular explanation is that punters, by nature, are risk-loving creatures, and so are more naturally drawn to outcomes that are less likely, but provide a bigger potential reward. Because of this, they are less price-sensitive: they see \$100.00 odds, they calculate massive returns, they don’t pause to consider that the fair odds would be more like \$300.00.

An explanation that is better supported by academic study is that humans have a cognitive bias that consistently misrepresents probability, especially when it comes to discerning the difference between things that are unlikely, and things that are very unlikely.

Cass Sunstein termed this – along with several other human failings to properly understand and utilise chance theory – probability neglect. And Jonathan Baron’s work on people’s inability to understand risk associated with medical conditions offers some of the most compelling evidence for its existence: with laypeople consistently unable to rationally decipher between risks such as 1-in-1,000 compared to 1-in-10,000.

We can think about similar scenarios in betting. If we are presented with two tennis players prior to the Australian Open, one at odds of \$1.50 and the other at odds of \$3.00, we would recognise that there is a significant difference between the chance of them winning. We are right to recognise this: the first player is twice as likely to be victorious.

If we are given two other possible outcomes, though, one at \$100.00 and one at \$1000.00, do we think of them differently, or do we consign both to the same category of “unlikely” and move on? Most punters – in common with humans in general – are naturally predisposed to view the unlikely and very unlikely as the same, even though, in our example, the \$100.00-shot is ten times more likely to win the Australian Open than their \$1000.00 competitor.

This cognitive bias (along with others that are explored in detail elsewhere on this site) can be guarded against by implementing a process that forces the punter to consider the value that any bet represents. Most successful punters use a spreadsheet programme, or something similar, to compare the odds they expect something to be with the odds they are getting, dispassionately investing on the outcomes that are probabilistically most in their favour, whether those outcomes are favourites, or longshots.

About The Author – Jack Houghton

As a passionate sports’ fan and punter, Jack has written about sports and betting for over a decade, winning the Martin Wills Award for racing journalism in 2002 and writing Winning on Betfair for Dummies, first published in 2006 and now in its second edition, having sold over 35,000 copies in two languages.