ShaunWhite wrote: ↑Sat Aug 26, 2023 2:22 pm
Automation is far from easy because although you have scale and speed you have to do without subjectivity.
As far as collecting data is concerned then if you've got a software engineer in your family then you might want to consider connecting directly to the Betfair api using c# or python (or vba) and storing it in a database (I use SQL). Betfair supply all the libraries you need. You'll end up with millisecond data as per the data they sell.
https://historicdata.betfair.com/#/home. You can run that 247 on a vps (such as AWS or the BA vps) for whatever sports you want. Slightly better than watching and using a notepad
You could then analyse that data, or, if you're serious and want to take it further, you could build/use a simulator to backtest against that data. Again there's open source software to help with that (Google 'flumine' or Betfairlightweight')
What you're doing now is fine, but that gives you some idea about how far you can take it. Don't expect any easy money though. Margins are small and the competition is fierce. There's only so much profit that can be shared around.
Thanks for the advice on connecting to the Betfair API I think we did have a look at it in the past but at the time we had other priorities. Perhaps an oversight, nice one for nudging me in that direction.
We have dabbled in some data-centric projects around football and late goals but decided to shelve that (for now) to focus on the racing as it was more enjoyable.
I did Odin Project a couple of years back so have some basic programming skills, but I admit, it just isn’t my bag, obviously I see the value in programming though, I’m not a melt. A.I though, big changes, I do feel we are slightly fkd but that’s another thing haha.
I am so so so lucky to have a programmer in the family, not only that but if I give her Sherry, she is happy to watch Midsommer Murders and code away ( I make her sound about 90 lol), I just need to guide her on the project. But yeh, my focus isn’t on that currently. Getting married in 5 weeks so all go! haha.
Right now, we're pulling lots of racing data from online sources, but it's sitting idle due to time constraints, I always think, it’s probably non more valuable to me in that sheer data form than how I currently approach things as an organic thinking machine with a back tested plan. I will be doing some machine learning courses though in the future, think Mike Pound has a basic course I could get into.
Will have a look into open-source (flumine/betfairlightweight) stuff cheers!
I see your comments about the forum, and you are always super helpful to everyone.
Brilliant stuff!