Favorite/Longshot Bias
 

I found this really fascinating from Barry Meadow's new book "The Skeptical Handicapper"

it refers to the favorite/longshot bias which Tom Ainslie identified in his classic book from the 1960s

I've heard many say that the favorite/longshot bias no longer exists

obviously it still does exist and is quite strong

these are the results from data from about 863,000 races - IV means impact value which I cannot now give a good definition of but it's not really necessary to understand the implications of this data:

sorry but I don't know how to make the image the proper size

https://i.imgur.com/pVznpkZ.png


for horses going off greater than 15/1 the ROI is worse and continues to get much worse as the odds go up

The fav / longshot bias does still exist but there are a few of things to be aware of.

1. In a single race there can only be one true favourite (or horse in the odds range of 0.10 to 0.99) whilst there can be many horses with odds of 9.00 to xxx.xx so the figures can be slightly misleading.

2. The Impact Value is not the best metric to use here. You are far better off using the A/E stat.

3. In the UK betting exchanges have affected the fav / longshot bias. Bookmakers have tended to supress the odds on outsiders whilst free market exchanges have helped to lengthen them and thus reduce the bias.

4. The bias will be stronger at the bigger festival meetings where bookmakers have to balance books.

It’s not just the UK, US exchange tracks so a similar phenomenon.

Long shot odds are suppressed on the tote.

Horses that are 10 to 20-1 range on the tote go off way more on the exchange.

I will post today at Monmouth.

Allan