Hi All
I'm looking to collect data now so I can fine tune my methods...
Which data do I collect:
Number of runners
Money being matched on each horse
Type of races, Maidens, Novices, Handicaps
Price of horses and money being matched on individual horses, I was told this data you must collect... Like a strong favorite normally produces the same characteristics... is that right? Am I collecting the right data listed above? Does anyone recommend other data that helps push you forward...
What kind of data should I be collecting?
Betfair already provide a lot of this information for you in a daily excel spreadsheet at https://promo.betfair.com/betfairsp/prices/ so I would start there first. It is a great resource for the days/races you may not have already collected your own stats for.
-
ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
Really it's any and all. Collect everying available (obv within the limitations of whatever you're using to collect it). If you don't think you want something now then it's a dead cert that you'll want it in a year's time and regret not having it.
It's worth spending a while (even a couple of weeks) planning and organising what you'll collect, there's nothing more annoying than having a system that gets a little tweak every now and then and you end up with a dozen sets of data that are all slightly different or need to be reformatted into your latest incarnation.
My caviat for the first paragraph is that it's going to depend on your technical skills (or those you plan to learn) and the effort/expense you are prepared to invest. I see there being 3 basic sources of data, the first is everything that BA can pump into excel, the second would be to collect everything BF sends in the API stream (or subscribe to their data) and the third would be to supplement that with other less available data, perhaps bookie prices, the weather or the going, race results, prize money etc
Everyone's areas of interest will be different and the more niche you are the better your chances of finding an edge, perhaps you're interested in horse racing and situations where trainer x has an effect when he runs new horses >100 miles from home etc etc etc. Personally that stuff doesn't interest me, I prefer the maths and software side so I eschew all the stuff about the horses and going etc, (I don't even have racing TV or watch much sport generally ) so capture the whole api stream and treat it like an insane game of sudoku and try to find things that are inherent in the mechanics of the exchange probably more than the mechanics an individual event or even sport.
It's good you're doing some planning and a really interesting question to ask. You can tell from the length of my message that I think it's one that could be debated for days, (enough to fill a book on its own imo) and although it's so core to what we do it's something people will struggle to find much help with let alone what to do with the damn stuff once you've got it!
It's worth spending a while (even a couple of weeks) planning and organising what you'll collect, there's nothing more annoying than having a system that gets a little tweak every now and then and you end up with a dozen sets of data that are all slightly different or need to be reformatted into your latest incarnation.
My caviat for the first paragraph is that it's going to depend on your technical skills (or those you plan to learn) and the effort/expense you are prepared to invest. I see there being 3 basic sources of data, the first is everything that BA can pump into excel, the second would be to collect everything BF sends in the API stream (or subscribe to their data) and the third would be to supplement that with other less available data, perhaps bookie prices, the weather or the going, race results, prize money etc
Everyone's areas of interest will be different and the more niche you are the better your chances of finding an edge, perhaps you're interested in horse racing and situations where trainer x has an effect when he runs new horses >100 miles from home etc etc etc. Personally that stuff doesn't interest me, I prefer the maths and software side so I eschew all the stuff about the horses and going etc, (I don't even have racing TV or watch much sport generally ) so capture the whole api stream and treat it like an insane game of sudoku and try to find things that are inherent in the mechanics of the exchange probably more than the mechanics an individual event or even sport.
It's good you're doing some planning and a really interesting question to ask. You can tell from the length of my message that I think it's one that could be debated for days, (enough to fill a book on its own imo) and although it's so core to what we do it's something people will struggle to find much help with let alone what to do with the damn stuff once you've got it!