I've been promising this for a while but as always, I struggle to get the time to pull out the data and write a blog post.
I can't recall doing this before on the blog but I'm not going to write about the results so far this season and the tables below. I said at the start of the season that I didn't want to draw any conclusions about the value ratings until we saw the results this season and that hasn't changed. The data is all there below and I'll share an update of this in the Summer with all subscribers but really, what we want to be doing is combining the value ratings with the systems and seeing what the results look like. However, this is a major project and one that needs to be undertaken in the Summer.
As you will see below, the sample sizes for the algorithms may be OK but once you start splitting the games between the value bands, the data gets very thin in places and even drawing conclusions from the historical live results isn't easy.
The other issue is that so far this season, the results are nowhere near as good as the last 3 seasons and therefore, drawing conclusions isn't easy due to this.
Hence, it sort of feels like a bit of a teaser post this one but I think the real work still needs to be done on the value ratings and how we use them going forward.
Let's park this until the Summer and we can pick it up again then. Anyone reading this can hopefully see some clear areas below where we can increase profits using the value bands and the systems but let's not jump to conclusions too quickly.
No comments:
Post a Comment