I have been meaning to write a some kind of description or review of Racing Dossier, the tool that I use as source for all of my data. But author of the tool Michael Wilding has beaten me and made a walk through video of how the tool works.
So, if you are not quite clear what this tool is about or if you have earlier seen the previsous Adobe Air version of the tool and I suggest that you take a look at the new browser based iteration of the tool.
We are having a discussion at Race Advisor forums about different speed figures provided in Racing Dossier and their relative strength in choosing contenders in All Weather races. I for one was telling my opinions just based on hunch and I wanted some data to confirm my own thoughts.
I looked at handicap All Weather races ran over 6 furlongs (rounded to full furlongs) with a field size of 8 to 10 runners. Data was collected from years 2012 to 2014.
I looked at three rankings based on four different speed figures which are
SHorPro – Horses projected speed rating in todays race
SpdFigLR – Speed figure earned last time out
ShorAvG – Average speed rating on todays going
ShorAvD – Average speed rating at todays race distance
I used ranking based on these ratings and not the raw numbers, so top rated would be ranked 1 on so on. To make the charts more readable I grouped the rankings to three groups, top 3, mid 4 and bottom 3.
First chart shows strike rates for those three groups combined from all of the tracks.
We can see that there are interesting variations between the tracks. But one ought to bear in mind that as data is only from three years there is relatively low number of data points (complete dataset was only a little over 3000 runs).
I use two approaches to my betting – for a while now I have been building an Bayesian betting system which calculates odds for each race on any given day. It is still work in progress but results so far are promising. And other avenue is betting systems where selections are made mechanically based on set of criterias.
In both cases I use same database to give me handy list of selections each day. So once the system is built and tested it is trivial to get the selections each day. That way it is easy to concentrate on the interesting parts and not spend lot of time in the menial stuff.
As a source of data I use excellent Racing Dossier by Michael Wilding of Race Advisor fame. Within the software there is handy export option to spit out csv-file of days races and set ratings that one has defined.
Once I have the data I upload it to my system and it will calculate the odds and prepare the list of selections. After that it is just matter of placing the bets, for them I use bunch of BOG bookies as well as Betfair. For figuring out which bookmaker to use I check Oddschecker for summary of odds available.
If you like to read more about my Bayesian system, I have written three part diary describing it for SmartSigger magazine. First part was published in July 2014 issue, second in August and third one will be in the yet to be released October issue.