I’m not going to go too overboard in this post mainly due to the fact that I have only recently started posting again on this blog after a year or so absence. Many of the blogs I have linked to in the past have since removed me from their own blogrolls (and rightly so) after a long period of inactivity so there’s only a few out there that still carry my link (and they too have been dormant for a good while also). Anyway the point I’m trying to make is that I don’t think that there will be many people who actually read this post but I’m writing it nevertheless.
I’m seriously considering a change in the way I’ve been doing things and going back to something that I still consider to be a potential good thing. There are a couple of reasons for this and they are that I have a fair bit of free time right now that needs filling plus my betting activities recently have been totally devoid of any kind of discipline/real focus/adherence to a set strategy. I have two choices – continue as I am or get off my lazy arse and do something about it. I’m choosing the latter.
Many years ago I completed an honours degree in Business Computing. My final year project was entitled Neural Networks For Horserace Prediction and without wishing to blow my own trumpet I have to say it turned out way better that I could ever have dreamed it would.
It will take some work but my plan is to resurrect this project and start again from scratch. The original floppy that contained all the training and testing information is long gone so that fact dictates that I must begin anew. The good news this time around is that the neural network shareware that I used for the project has had 15 years to evolve and get tweaked. That, plus the fact that the computer I have now has significantly more power than the Dell Dimension DX2 66MHz pc that I had used to complete the project the first time round. I am almost excited at the prospect of a long drawn out period of obtaining and feeding in the thousands of pieces of data that are necessary to retrain a new network to recognise the patterns that ultimately will single out future winners.
Anyone who has read this blog for more than a couple of posts will see that I rarely venture into backing horses, preferring to stick with football matches. There are reasons for that. I used to make my living from backing horses and I knew a fair bit about the game and I’m not naive enough to believe the claims by those who make them that racing is all above board. Bollocks. As a former ‘putter onner’ for a couple of (former) high profile jockeys I can tell you that isn’t the case (at least it never used to be). It’s quite tiresome though when you place money down in good faith on behalf of others and things don’t go the way you’ve been told they will and you soon learn that if you do your own thing then you’ve only yourself to blame when things go wrong. Having said that I’m not saying that every race is bent…if it was the sport would soon die a death if the racing public thought that was the case and so it follows that opportunities exist to accurately predict the winners of truly run races now and in the future.
I have toyed with the idea of creating something for football as well and I have a couple of ideas but rather than go to a lot of trouble and jump feet first into a football application I’m going to stick with racing as I already know exactly what data I need and where to find it. I just wish I was half decent at writing code as I could cut my input time exponentially by screen scraping the info and using it as my input rather than manually entering it in but my forte isn’t as a programmer so I have to go the long way round.
I’m also going to create a blog based site that I own rather than write about it on a free wordpress blog like this one. In fact I’m going to migrate some of the previous content that is on here to the new site as I feel there are a few really good older posts on here that it would be a shame to just kill them off.
Not that I expect a deluge of comments but if anyone has anything useful to add that might be in this ballpark (neural networks used in a betting application) I’d love to hear from you. For now though I have several things to attend to – I need to get hold of the latest version of the Neural Net simulator software, I need to start looking for and compiling a huge amount of past race data, I need to get a website up and running…there’s plenty ahead yet before this thing goes live but I promise you this – if this new network performs anything like the best performing network out of the 8 I created for my project then there’s going to be a lot of interest in my new site.
Watch this space for more news as things progress.