The Favourite–Longshot Bias
What Is the Favourite–Longshot Bias?
The favourite–longshot bias (FLB) is a well-documented phenomenon in betting markets where:
- Favourites (low-odds horses) tend to be under-bet and offer better long-term returns.
- Longshots (high-odds horses) are over-bet, meaning they return significantly less than their fair probability suggests.
In simple terms, bettors consistently overestimate the chances of outsiders and underestimate favourites. As a result, the expected return from backing longshots is typically much worse.
For example:
- Betting on favourites might result in a ~5% loss.
- Betting on longshots can result in losses of up to 40% or more.
Historical Context in UK Racing
In the UK, the FLB has been recognised for decades. Classic studies have confirmed that:
- Bettors are often drawn to longshots hoping for a big win.
- Bookmakers may shorten longshot odds to limit liability.
- Bookmakers’ pricing strategies and bettor psychology both contribute to the bias.
A notable explanation is prospect theory, where bettors overweight small probabilities and thus favour outsiders disproportionately.
Another factor is bookmakers protecting against insiders—they price longshots cautiously in case some punters have better information.
Evidence from UK Horse Racing: Favourites vs Longshots
Data from UK and Irish racing circuits in the 2010s show the scale of this bias.
| Horse Rank by SP | Win % (Approx.) | ROI at SP |
|---|---|---|
| 1st Favourite | ~32% | –7% |
| 10th Favourite | ~2% | –44% |
- Favourites are far less risky and deliver relatively smaller losses.
- Longshots are rarely successful and suffer extreme negative ROI.
Recent Research
Recent academic studies concluded that:
-
- The FLB still exists, but it has weakened.
-
- A study by Smith and Vaughan Williams (2000s) showed the bias declining compared to the 1990s, likely due to:
-
- The rise of online betting.
-
- Increased market efficiency.
-
- Betting exchanges driving sharper odds.
-
- A study by Smith and Vaughan Williams (2000s) showed the bias declining compared to the 1990s, likely due to:
A 2024 study by Whelan found:
-
- Even in fixed-odds online markets, the bias remains.
-
- Causes include risk aversion by bookmakers and disagreement among bettors.
-
- These effects naturally recreate the FLB, even in efficient modern settings.
So, Does the Bias Still Exist?
I’ve always been fascinated by the FLB, as I’ve been convinced it’s a relic of days gone by. It’s one of those things, similar to “95% of traders lose”, that have become ingrained in people’s psyche, but are rarely updated.
Modern studies seem to be looking for it, rather than proving it still exists. But I’m sure it doesn’t. Some of that may come down to how to measure it.
As a baseline measure for more specific strategies, I lay every favourite, every day and have done for years. So far in 2025 you would have made a little money, before commission, laying the favourite blindly. This directly contradicts the FLB which says I should have lost money.
If you are more specific about the staking method, you will obtain varying results. Should you be backing with £100 each time to use proportional staking to back to a target profit? I suspect that many studies use level stakes, which can distort the results.
But it’s rare in my experience to find a longer term period where backing the favourite makes sense. People nearly always overbet the favourite on exchanges as it feels better to win frequently and be unprofitable, rather than win intermittantly and profit.
People also prefer to win £100 by staking £10, than win £10 by staking £100 by laying. It’s just human nature. That all contributes to a shorter priced favourite than you would suspect based on previous studies on the FLB.
Final Thoughts
The favourite–longshot bias continues to be a popular topic for research when seeking betting systems. It’s clear to see why, as it was a system that had proven academic support.
However, when looking in traditional betting markets its more likely to reflect the way a sportsbook has created a market and the thin odds they offer at larger prices to claw back margin.
On a betting exchange, I’m not convinced it has ever existed and that is reinforced by the fact all recent data points to laying the favourite as slightly perfereable to backing one.
However, people don’t like to accept the higher risk of laying, despite the fact that an efficient market means there should be no preference.
For that reason I suspect looking for overbacked favourites will continue to be my preference.