Finding Value at the Horses in Training Sales

There’s a part in Moneyball where Peter Brandt talks about building a championship team out of a island of misfit toys.

It’s this approach that resonates with me when we’re looking at buying from the Horses in Training sales. Here we have horses that are mostly unwanted by their current owners, largely due to under performance or physical issues that have prevented them from realising their full potential.

The 2021 October HIT sale may present somewhat of a different setup due to the large scale sell off from Shadwell, but ostensibly we are looking for under-performing (& under-valued) horses that we can somehow fix or improve.

We believe that by taking an extensive analytical approach; combining data analytics, pedigree study, stride & sectional data and race video analysis, that there is an opportunity for buyers to identify value at these sales. Purchasing horses that can not only win on their subsequent starts, but can also potentially improve their ability and sales value.

There are a number of angles that we try to exploit in order to identify value at the horses in training sales.

1. Ability Potential

The fundamental aspect here is that we need to identify horses that have ability. You can’t make fundamental poor horses good. We need to identify horses that have a level of potential ability, but have as yet not showcased this on the track.

What we as purchasers then need to unravel is the question of why the horse has not been able to realise this ability potential so far. It can be through a number of factors;

Injury — limiting racing opportunities on the track or aspects like poor wind (breathing) requiring resolution

Bad Luck— poor draw on their race starts or tactical setup of races not suiting their run style

Keenness — attributes like being too keen in races limiting their chance to show their true ability,

Lack of optimal conditions — horses being run over the wrong trip or group by their current trainer. Time and time again we see horses running over the incorrect distance based on pedigree analysis and stride data. Finding these horses can give us a good buying opportunity.

2. Trainer Uplift

Trainer ability is one of the most significant aspects that influences a horses performance. Good trainers can make average horses good and bad trainers can make very good horses perform at a poor level. We need to handicap trainers themselves in order to identify those which could present us with value purchasing opportunities. Moving a horse from a poorly performing yard to a stronger operation can hopefully contribute to us to start to realise their potential.

In order to properly quantify this impact I use an overall trainer ability metric in all my models and also a rolling 1 year trainer ability metric to identify yards that may be out of form.

We’ve also gone back through all the Horses in Training sales since 2012 and analysed the performance of subsequent starters sold out of yards across the UK & IRE (subsequent starts only in UK & IRE, as I don’t have any foreign racing data). Here we can gain a view of which yards present us with good purchasing opportunities, we can also limit risk by staying away from the yards that have seen runners perform poorly post sale. Here’s a extract of the type of analysis we are doing, with the trainer names removed:

Winners to Runners Ratio from October HIT since 2018

One UK trainer has seen 11 of his horses sold at the HIT sales since 2018, with 9 going on to be subsequent winners in the UK & IRE and one improving their official rating by 50lbs.

I mentioned risk avoidance, as this is a huge aspect we are also looking at in our data profiling. We want to avoid fatal mistakes such as buying out of operations where horses may have been over-worked or where the operations themselves are very shrewd judges of ability. If we can handicap the yards that we do and do not want to buy out of, it can start to give us an edge.

Trainers not to buy from

Finding Value

We bring all of these aspects together via 3 analytical models which assess the factors outlined below.

1. Potential Ability for each horse — defined by a Predicted OR level and % chance that a horse has the potential ability of the OR band.

2. % chance that each lot can be a future winner. This is based on past HIT data.

3. % chance that we can improve the form of the horse. This is also based on past HIT data.

We use the outputs of the 3 models above, all brought together into a cloud based analytical platform to work through our analysis. In the everlasting search for horses with a high level of potential ability, that are just waiting for the right opportunity to demonstrate it.

Now the search begins…

If you’re interested in being part of one of our syndicates then you can drop me an email on tomwilson41986@gmail.com