What is it You Guys Actually Do?

Saturday 23rd March 2019
Author: Diarm Byrne

Questions Banner

What is EquiRatings?

EquiRatings is a sports data company for equestrian sports.

We are committed to advancing equestrian sports data usage by the media, professional teams, governing bodies and sport at large. Our passion for equestrian sport coupled with industry experience informs the work we do in this space, including the development of predictive models, analytical metrics and utilising multiple data sources.

What is sports data for Equestrian sports?

We use analysts and data scientists to study results and quantify risk and performance. We compile lots of metrics that are significant indicators towards risk and performance.

We also do fun stuff like notifying the world when a new record is broken or highlight important achievements.

Why is quantifying performance and risk useful?

We measure so that our clients can manage. The most efficient way to lose weight is to know how much you weigh today and how much you want to weigh in the future. Then you compile a programme that involves the consumption of calories (diet) and the burning of calories (exercise) for the duration of the specified period. This is a classic measure-manage approach that EquiRatings applies to risk and performance. Where are you performing today, where do you want to get to, and where are the most efficient places to make effective marginal gains that can help you to achieve that goal.

Can’t we just measure this stuff ourselves?

Most people measure outcomes which is dangerous when you are playing a game of chance. Performance and risk in sport are most definitely games of chance which are subject chance variation. This is why football managers get fired prematurely, tennis players change their swing when they shouldn’t and golfers change their clubs for no valid reason. Of course, changes can be made for the right reasons, but people normally tend to react emotionally rather than analytically when they are affected by chance. The EquiRatings team ensure that our clients don’t fall into this trap because we assess chance and variance rather than just the outcomes.

Explain the difference between an outcome and chance….

Take a coin. The probability, or chance, of flipping heads is 0.5 which is the same as 50% or 1 in 2. From a performance point of view, the 0.5 probability is the important number. Next, we flip the coins three times and see what outcomes we get. The probability distribution is as follows:

12.5% (or 1 in 8) chance of flipping three heads and no tails

37.5% (or 3 in 8) chance of flipping two heads and one tail

37.5% (or 3 in 8) chance of flipping one head and two tails

12.5% (or 1 in 8) chance of flipping no head and three tails

As you can see, the problem with assessing outcomes is their susceptibility to chance. Two coins with the same probability can perform very differently. This doesn’t mean that a coin is biased, broken, training badly, has a very poor technique or crumbles under pressure. It is just chance variation. In sport, if two horses start their careers, one with three clears from three and the other with no clears from three, we know the conclusions that people will reach. What the coin example above shows you is that if the underlying probability was 0.5 then that amount of variance could actually just be chance rather than underlying ability.

This is sport not coin-flipping, surely the best team wins?

The best teams, the best athletes and the best horses do not always win, which is part of what makes sport so exciting – not to mention making it worth over a hundred billion dollars as a betting industry. In show jumping, exceptional horses can have a fence down when doing very little wrong and average horses can rub two or three rails but still jump clear. The outcome doesn’t make the latter better than the former, or make it worth hundreds of thousands more, it is their underlying expected performance that matters going forward.

When a good team is beaten it’s because they performed badly…

This is not always the case. But it is the mentality of many top athletes and it is detrimental. Win, lose or draw there are likely to be areas of the performance that were good and areas that weren’t. Some will be easy to improve, others more difficult. Some will make big differences , others will be quite subtle. This requires good analytics. It is why we measure, so that we can manage. Athletes often struggle mentally and go ‘in and out of form’ is because they look for a reason where there is none. They get emotional rather than analytical. Teams win and think they are better than they are and teams lose and allow their confidence to drop when it shouldn’t. Analysis isn’t just a management tool. It is a strategic tool, a tactical tool – and a vital psychological tool.

So it’s OK to say you were unlucky?

Luck doesn’t exist, it is simply an extreme point on a probability distribution. Outcomes are unlikely but never unlucky. Once you use the word luck you stop trying to understand that outcome or accept responsibility for it. If you don’t understand it or measure its likelihood, you will never effectively manage that area of performance or risk. One football team can create 25 scoring chances and none find the net, while the other team can have one shot and it gets a deflection and goes in. The result of a game like that is extremely unlikely but it isn’t unlucky. It was always a possible outcome.

A win is a win, it doesn’t matter if it was unlikely or not

This is true. However, EquiRatings is about performance management.That is all about the next result rather than what just happened. An unlikely win may well increase your chance of success going forward, but it doesn’t mean that you are likely to win again, or that you will have a good chance of success if you step up a level. We see riders have a cross-country round full of mistakes and near misses but by chance they avoid penalties. Another rider can have a smooth round and then one minor slip can result in a penalty. The fact is that one rider got a qualification and the other didn’t, but that doesn’t mean that one rider will do better at the next level than the other ride or that the first rider is a better rider.

Can EquiRatings help me to see a stride?

Once a competitor enters the arena in eventing, it is down to the rider, the horse and chance – there is nothing left for EquiRatings to do that can impact that result.

The objective of EquiRatings is to help your training and your performance to become more purposeful. Focusing on winning isn’t always helpful, because winning involves beating other competitors and their performance is out of your control. You can perform brilliantly and not win and you can perform quite poorly and win, therefore our aim is that you focus on the metrics that matter.

What are the metrics that matter?

The two metrics that should guide an eventing competitor are expected finishing score and cross-country jumping reliability. They are to eventing what calories consumed and calories burned are to weight loss. Reduce your expected score and increase your reliability and, over time, chance will likely put some good results on your CV.

The big strategic metrics which contribute to the expected total are expected dressage penalties, expected show jumping penalties and expected cross-country time penalties. It is important to gauge how effectively and efficiently you can control these metrics whilst still increasing your cross-country jumping reliability.

How do EquiRatings calculate these metrics?

By the time horses and riders are competing at the top level internationally they have a data footprint that allows our team to calculate these key metrics. We build and improve predictive models all the time. As we know from the sports betting markets, no model will predict the future so it is about managing the standard error. We account for factors such as competition type, courses difficulty and the standard of competition because not all events are as challenging depending on conditions. Our models essentially summarise a large amount of data and give a scientific estimation of the true level of performance.

When are EquiRatings metrics most useful?

For either managers or athletes, the road to improvement starts today. The EquiRatings metrics are like giving athletes a weighing scale. Some are disappointed when they see the information while others are pleasantly surprised, but the key is that the driven and determined athletes have something to aim at. Most humans need a target when it comes to maximising performance, focus and motivation. Yes, the shoulder-ins and gridwork will remain the tactical training tools, but the EquiRatings metrics drive purposeful practice and ensure that athletes are focusing in the right areas.

Know where you want to go. The metrics are the map.

What do clients have to provide EquiRatings with?

A predictive model is as good as the data you put in to it. We get all our management clients and athletes off to a good start with their key metrics from their past competitions, but now some of our clients are really understanding how to maximise the system.

We also capture pre-event data from managers, coaches and athletes that can be analysed too. Data-driven performance management is no different to many of the best things in life, you get out what you put in.

What’s new in the EquiRatings portfolio?

Increasing the efficiency and productivity of a training system will inevitably improve performance so we have had our sights on this space for a while.

Our eventing high performance clients will have the chance to start trialling some of the tools quite soon. We are looking at some of the wearable technology that is emerging. We would strongly suggest that potential clients get started on our current systems for now as everyone will find their own preferred ways of using the data. Once you get started then it will evolve and progress.

Is it only open to teams?

We started working with the Olympic teams but we are moving closer to being able to work directly with individual clients.

What do you want to achieve? A top ten in a five star follows a different path to winning a CCI2*-S. Know where you want to go and then we help plot the map.

We are discussing launching a simulation tool for riders which allows you to enter your target scores and then see where these scores would leave you against the averages at each level of national or international competitions. Busy but exciting.

Don’t forget to share…