AI Horse Racing Picks An Opinion

The definition of AI is artificial intelligence.

My background is a freelance IT consultant for over 25 years on corporate Unix and relational databases, and I enjoyed it. Over the years I have always studied UK horse form manually and automatically using my own horse racing systems. AI is a relatively new area to myself, but I have dabbled with it and my own personal opinion it has its good and bad points. I do not like AI for content generation as I like to write natural content as it feels personal and I think it reads better.
With respect to AI horse racing picks in my opinion I think it is a two phased approach,
1.    Use AI to do the grunt work or as I say the heavy lifting.
2.    Manually take the input (horses) from 1 above and do the horse form.
Sounds boring but that is my approach at the moment, but I also use a bespoke relational database and bespoke SQL as a substitute for AI – Artificial Intelligence doing the grunt work. I have written many horse racing systems using this approach.
AI cannot identify the critical horse form gaps or markers that the experienced eye can see. This is gained over many years of studying horse form and watching horse races. Now do not get me wrong I think machine learning might achieve this with adequate data sampling. The approach could use 5 years of past horse form data over 1000’s of races.
Maybe I will start looking into this I might be pleasantly surprised!
I looked again at the AI for picking horses and started investigating the best programming languages to achieve this,
1.    The popular Python programming language.
2.    R Programming language.
Finally LISP programming language which has been around a long time, even at my polytechnic days and that was in the 1980’s.
I will leave it there as this is a large topic that needs to be addressed via a white paper.
How does AI factor in all the little snippets of horse form data that will be missed that can only be seen by the naked eye. There are fine degrees of margins in horse form data across horses in a race. I think a lot of horse form reading is instinct and gut feeling and past experience, how does AI factor this in?
In my opinion it cannot!

A double figured priced AI test?

I wonder how many doubled figured priced winning horses AI would pick in any given year?

I have not seen any metrics on this so I reserve judgement on this.

Hope you enjoyed this.