The favourite–longshot bias is one of the oldest and most robust patterns in betting markets: on average, longshots are overbet and favourites are underbet. Punters pay a little too much for the dream of a big-priced winner and slightly too little for the boring short ones.
What the pattern looks like
If you group horses by their starting price and track how often each group actually wins, short-priced runners tend to win slightly more often than their odds imply, while very long-priced runners win less often than theirs. The market isn't wildly wrong — but the error is consistent, and consistency is what a model can exploit.
Why it persists
- The thrill of a longshot. A 50/1 winner is a story; a 5/4 winner isn't. People pay for the story.
- Risk-seeking on small stakes. Bettors treat a small stake on a big price like a lottery ticket.
- Bookmaker margins are often loaded into longer prices, deepening the effect.
What it means for value
The bias suggests value clusters nearer the front of the market more often than the crowd assumes — but "back all favourites" is not a strategy: margins and variance erase a naive approach. The point is subtler: a model that rates runners on form and conditions can identify which shorter-priced runners the market has still slightly underrated, and avoid the longshots it has overhyped.
How we treat it
Our model doesn't assume the bias — it learns the relationship between price, form and outcome directly from 38 years of results, then we measure value against the close using CLV. The favourite–longshot bias is one of many structural patterns that make a disciplined, data-driven read worthwhile.