tbh - altho i've not really looked deeply into this yet, there's simply (imho) too few data points to make any assumptions from. the original premise was based on some sort of *wisdom of the yard*. however, i don't think this is actually the case. as i mentioned before, I'm of the opinion that some of these lower odds STEAMERS are probably over backed and may present a good opportunity to lay a select few. however, context is obviously key and I would be looking for a couple of close contenders in the market for this type of activity.
one other aspect that is worth delving into, is the amazing
black box of stats that sportinglife package up under the covers when loading their racecard pages. you can literally grab the OC bookies prices, stats on the recent runs for each selection etc, etc. in short, you can pull this data down pretty much 1-2 mins before each race and use your bot to determine if there's any value to be had (backing or laying). in the lower class races in particular, there are some very uncertain odds flying about. I'm sure you'll be able to work it out, but here's the
hidden api starting point for todays races:
https://www.sportinglife.com/api/horse-racing/racing/racecards/2020-12-28. from this, you can get the full cargo data for each race on the day (race_summary_reference.id) which is presented as fully formed json also. anyway, just another little additonal tool for the box...