I've been parsing the Equibase Scratches and Changes XML since 2009.
In 2013 or 2014 I started saving scratches and changes parsed from the XML to a table in an Access database - and from there started generating a daily changes report.
• My primary objective for starting this project was the ability to generate a report showing new scratches and changes - so I could quickly do number crunching for races that have undergone a change.
Link to yesterday's changes report so you can see what that looks like:
http://www.jcapper.com/messageboard/..._07082021.html
I've also attached an Excel Spreadsheet that contains yesterday's scratches that I exported from the Access database so you can see what that looks like.
• One of my secondary objectives for doing this was the ability to test a couple of hypotheses:
1. Is there an roi-boost to be had when scratched horses come back to race again?
2. When scratched horses come back to race again does the reason for the scratch matter? For example, vet scratch vs. scratched for various other reasons? (Another player had assured me there was an roi-boost to be had especially for vet scratches.)
After analyzing data for several thousand vet scratched horses in their subsequent starts - the data suggested there wasn't much there that wasn't already reflected in the odds.
Afterwards I had a conversation with the player who had assured me there was an roi-boost to be had especially for vet scratches.
He told me I was painting with too broad a brush - and that every horse has a story.
He told me a lot of vet scratches aren't really vet scratches at all - that many arise not because the horse had a cough on race day or might still be a little sore from a fast recent workout --
But rather because the connections saw something they didn't like on race day.
Maybe it rained and the track came up sloppy... or they drew an inside post and they suddenly realized the speed of the race was drawn to their immediate outside... stuff like that.
So they decide to scratch and find an easier spot. And to avoid any resistance from the racing office they find a vet willing to sign off on it.
He told me those are the vet scratches where you can occasionally make a nice score.
But to do that, Imo, it's easier if you have connections with a horseman or two at your local track. If you're there every day, after they get to know you - sometimes you get to know a horse's story.
• I'm currently clearing the Access table at the end of each race day because I'm no longer looking at the scratches data as a basis for potential plays when horses come back to race again.
But I am archiving the daily scratches and changes xml files downloaded from Equibase in case I ever decide to give the idea a second look.
•
And like Tom mentioned -- the scratches data exists in the HDW Charts Data - columns CS and CT if you import into Excel (which by the way is easy to do.)
And like you mentioned CBYRacer - to make it work, because horses often scratch at Track-A and show up a few days later at Track-B - you'd need to become a monthly downloader.
-jp
.