Race Selection 1

Whats behind the Race selection?

It wont have been too long since you last heard a pundit spout  "I'm going to overlook the favorite and look for some value"  - or something along those lines.

How do they get the job when they can't even understand the basick difference between Value and chasing a High Return?

There's a substantial amount of good research on the favorite Long Shot bias in markets.
I'ts not new at least 60 years of information. Plenty of info on the UK markets but;
about the strongest evidence I came across was the Snowberg/Wolfers * study of American racing where they claim;

The most discussed empirical regularity in racetrack gambling markets is the favorite-longshot bias.

- and they back up their research with 6.4 million horse race starts in the USA  1992 - 2001.

Put simply if you bet randomly on 100/1 shots or greater you're rate of return (deficit) is @ 61%.
If you bet randomly in the 4/1 - 9/1 range you're rate of return is improved @ 23%, while betting the favorites the rate of return deficit is down to 5.5%

But its in the DNA of punters (and most pundits / tipsters) to overestimate their skills and underestimate how they are simply riding their luck! They value favorities  i.e. the market too little, given how often the favorite wins.

Within the efficient market there is a built in bias. This is maintained by bookmakers as well as over confident punters, bookmakers build in most of their over round at the higher odds as a part of a risk averse business model.

In addition to the Long shot-favorite bias I also believe that the markets fail to discount fully the enormous variation in favorite returns between different types of races. Or more specifically Top of the market - Long shots.
Most punters acknowledge that favorites have less chance of success in handicap races but they don't have the slightest handle on the actual probabilities.

In a typical scenario the punter may investigate a bit of collateral form but then blindly plumps for a mid price horse hoping for a good return, probably flouting any factual information gained. The pundits do exactly the same because they don't want to be on the line for actually forecasting yes/no whether or not the short priced horse will win! If the favorite does win there's always an easy line in post race waffle - he ran right up to his mark, etc)


Which leads to the starting point for the race selection process for Dutching.
Here is a good sample size comparison between one of the race types included for Dutching (novice chases) compared with the lowest ratings found in turf handicap races (Handicaps - below in yellow) 

The fav. takes 43.6% of Nov chases (82.4% first 3 in market)
compared to turf handicap figures down to 27% for fav. (57% first 3 in the market)

The race selection format in the daily screen shots shows the 1 / 2 /3 figure + the Uplift factor I apply for 2nd & 3rd in market.
(Nov chase figures vary a little due to the applied filters on Class & field size)

To answer those pundits 'Looking for value' Value is a measure of a horses true chances against the odds on offer in the market. It can be (and more likely is) in the odds on favorite -  a probability forecast - just chasing a high return on a single hopeful selection is something quite different.

If there is Value and an edge to be found at the top of the market, because the top of the market has a high success rate in a particular type of race, then which horse is the Value priced into?

Well the answer on Betfair is usually found, with the benefit of hind-sight, when the In_Play odds are down to 1.01 and every bit of information is digested by the market! 
All sorts of micro anomalies have come into play prior to this - going, pace, tactics, jockey instructions, fitness peaking, etc, etc.

The Dutching format I follow accepts the anomalies, has the top of the market in the races selected covered -  holding to a formula always exceeding that benchmark 80% figure in terms of the Dutch book value that can be covered and takes an edge from the market bias.

Using this betting method as formatted here, delivers a strike rate at 90% + which means low and manageable volatility. 

The anomalies in both getting the horse to the start and during racing make the 'Value' equation pretty hard to forecast - but the cause and effect in the background is  - the Lower the Volatility - the easier it is to forecast the end result.


* 
Explaining the Favorite-Longshot Bias: Is it Risk-Love or Misperceptions?
Eric Snowberg / Justin Wolfers April 2010