Welcome to the Eventing Prediction Centre – the home to the new eventing statistical model from EquiRatings built on the latest predictive analysis tools available from SAP.
The model was trained on over two million rows of data. The Eventing Prediction Centre allows eventing fans to assess for the first time who the favourite is, who the outsiders are and how each phase impacts on those likelihoods.
We have a section which allows you to track the movers after each phase, and a section where users can vote on their predicted podium, and see how their predictions line up with the greater audience, and with our predictive model. For the fans who like to dig deeper, you can read more about how the Eventing Prediction Centre generates these probabilities here.
Our SAP Eventing Insights blog is here to keep you updated throughout each competition, looking solely at what the model tells us and explaining it as we go along.
We start with the Event Rider Masters opener at Chatsworth CCI4*-S. Oliver's decision to run Cillnabradden Evo at Badminton means that we now have a very clear favourite in Laura Collett and London 52. Before a single horse walks down the centre line, the model has her well ahead of the pack. A winning probability of 20% is two and a half times that of the next highest rated competitors. She is the clear favourite here and is likely to be one of the stronger favourites we see all year. Winning probabilities at this 20% mark are among the highest the Eventing Prediction Centre ever gets to for fields as strong as Chatsworth ERM.
ERM stalwarts Bill Levett and Gemma Tattersall arrive here with Shannondale Titan and Quicklook V. While the model has them well off the favourite - they still make up a large share of the total win probabilities. Previous ERM leg winners Treason (Sarah Cohen) and Zagreb (Alex Bragg) follow those. Add in Julia on her young mare Amande De B’Neville and we have our market leaders picked.
Below these six, the win percentages start to flatten out, with the next eight competitors having between a 3-5% chance of coming in first. However, these should not be written off. When running our model over all the top level competitions of the last two seasons, combinations with a 5% probability did indeed win 5% of the time. And if you look at their podium odds, all of these competitors have a podium chance of at least 10% with Piggy French-Brookfield Innocent having a 16% chance of a podium finish.
Hopefully the Eventing Prediction Centre will roughly agree with the readers intuition of the sport. Occasionally, it will throw out something a bit unexpected. This might be a a result of a quirk in the model, or perhaps it has uncovered something we might not have considered at first glance. In Chatsworth the second most likely combination to place in the top 10, according to the Eventing Prediction Centre simulations, is Julia Krajewski with Amande De B’Neville. We predict them as more likely than not to finish in the Top 10. That isn’t something I would have guessed from the outset, and is definitely worth keeping an eye on.
We are really excited to see how our Eventing Prediction Centre performs throughout Chatsworth. It is the first time the sport has the ability to create favourites based on statistical analysis and bring it to the fans. Probabilities will be updated at the end of each phase, and SAP Eventing Insights articles updates will tell you what the big stories of the day are. Keep an eye on this space to follow along, and let us know on any of our social media channels, do you agree with our predictions or who your selections are.
See you in Chatsworth.
Share this story: