‘Value’ National Hunt trainers
Crunching racing data varies from researcher to researcher, punter to punter. Gone are the days of collating data from hardback book copies of Timeform for example.
Computer programs, racing databases, racing websites etc. have made it so much easier to do this type of work, especially from a time perspective. Something that may have taken me 2 weeks to do back in the 90s I can now do in less than a minute.
For example, if I wanted to find out the performance of favourites in all flat races going back to 2014, that would not be difficult to do. Also, it would be super quick – essentially input the basic parameters you want to test, press a button and bingo – it is all there in black and white.
Then you have a smaller band of people who use more sophisticated methods such as Neutral Networks, Machine Learning, etc. I did some mathematical modelling when I was at university and hence, I can see its benefits and attraction. However, much of the coding needed for creating horse racing models is above my pay grade.
For me, in general, I like to keep methods and ideas simple where possible. Why complicate things? I am a believer in the saying ‘KEEP IT SIMPLE STUPID.’ This article is based on very simplistic ideas / methods.
My aim is to try and find National Hunt trainers that offer punters some value.
However, I won’t be looking at profit and loss, or indeed return on investment; I want to study two different ‘parameters,’ which I will explain as I go.
The data for this article is taken from 1st Jan 2017 to 12th Sept 2022 – UK National Hunt Racing only. This gives me a decent number of races to evaluate trainer performance.
1. Looking for Place Value
As someone who likes the placepot and the occasional each way double, I am always trying to find an edge in terms of place betting. So, I want to try and establish which National Hunt trainers offer value in the place market? And how would you find this out?
As stated earlier, my approach is going to be a very simplistic one, and I appreciate not without issues. However, with any research idea or method there are going to be flaws. Hence you need to be realistic and appreciate this.
Time to explain what I did. To begin with I looked at ALL trainer data for the period of study, looking specifically at market position. I wanted to gather placed data for all market positions from favourites to 7th in the betting, and then I combined all the data for runner’s 8th or bigger in the betting. When I use the word ‘placed’ this of course includes wins and placed runners combined.
Here are the findings for ALL runners:
Now I had the placed figures for each market position I then had an average figure or benchmark from which to work from. From here I simply looked at all trainers that had at least 50 runners within the relevant market position and compared their placed percentage with the overall average. I was looking for trainers that had higher figures in each market position compared with the average. I was not expecting any trainers to have higher figures across all eight, but you never know.
Essentially, I took each trainer’s placed percentage for that specific market position and subtracted the ALL runners average figure from it. So, for example if a trainer had a placed average of 65.34% for favourites the calculation would be 65.34 minus 60.84 to give that trainer a +4.5 ‘score’.
Obviously, any trainer with a percentage below the average would produce a negative ‘score.’ I have created a table below with a comparison for individual trainers across each market position.
Here are some of trainers that I would say had positive outcomes overall (figures in green are positive). Any empty cell means that the trainer in question did not have at least 50 runs within that market position.
These trainers I perceive offer us some place value. Good to see Milton Harris with a string of positives as he trains not far down the road to where I live. Also, as a fan of Emma Lavelle and Donald McCain so I was pleased to see their names there too.
Of course, as punters we cannot just be aware of ‘positive’ data.’ We need to be aware of all data. Hence it is worth sharing those trainers with more negative outcomes as well:
Gary Moore has clearly done well with favourites from a placed perspective but 6 of the 8 market groups are in the negative. Sad to see Rebecca Curtis in there as she is one of my favourite trainers! Maybe I need to rethink that.
I mentioned that there were potential flaws in this method. One such being that we are using market position rather than a very specific price band. Hence one trainer could have had a higher average price than another trainer in a specific market position. That of course would almost certainly have an impact on the placed percentages.
However, I did look at a handful of trainers and their average prices within each market position, and any differences were minimal. Another potential issue is that the number of runners in a race can make a difference.
If one trainer has favourites mainly in races of 7 or less, and another has favourites in races of 8 or more then this would not be a fair comparison. Once again, I cross-checked some trainers to check this wasn’t a huge problem and in general field sizes balanced out. This is probably because I used a decent sample size. Hence, I am happy that my very basic measure of value has valid results.
To read the rest of this article upgrade to a Gold or Platinum account now by clicking HERE
Already a Gold or Platinum member? Read the full article in Issue 96
David Renham
Featured Image: (CC BY 4.0) – Towcester-15 | Ryan Marsh | Flickr