Inside the model
What 18 months of racing data taught us about reading a race
Most of us spend the morning going through the card and still choose on a feeling. We spent eighteen months building something that does the heavy lifting instead.
The hours that lead nowhere
If you follow the racing, you know the routine. You open the card. You read through the runners. You weigh recent form, the trip, the going, the trainer, the draw. An hour later you have a shortlist and a nagging sense you have missed something. Most days you choose on instinct, because there is simply too much to hold in your head at once.
How much there actually is to weigh
A single runner carries dozens of signals. Recent finishing positions. How the going suited it last time out. Whether the trainer is in form right now. How it shapes up against today's field on the numbers. The trainer and rider combination. How long it has been off the track. Hold all of that across a twelve-runner race and you are tracking hundreds of moving parts. No one reads all of it by eye, every race, every day.
So we stopped trying to
Over eighteen months we gathered the racing data, runner by runner, race by race, and built a model that reads every one of those signals at once. It looks at [196,633] runners across [19,627] races, and it scores every horse in every race before the off. Not a shortlist someone felt good about. A calibrated probability for the whole field, produced the same way every single day.
What the model weighs
It considers six categories on every runner:
- Form patterns. How a horse has actually been running.
- Conditions. How the going and the trip suit it.
- Connections. The trainer and rider signals.
- Breeding. What the pedigree suggests.
- Race context. How it stacks up against today's field.
- Time since last run. Freshness and fitness.
It does this for every horse in the race, then surfaces the ones it scores most strongly. The same process, applied the same way, whether it is a Tuesday at Wolverhampton or the feature on a Saturday card.
We publish the record
We show the full record on the homepage. Every scored selection, settled and shown, with the strike rate updated daily for anyone to check before they sign up. The chart goes down on the bad days as well as up on the good ones, because that is what an honest record looks like. We would rather show you the real one than make a claim we cannot stand behind.
See every runner scored, the same way, every day
You do not have to take the morning apart by hand anymore. We do the reading. You see the scores.
Keep exploring
Hand-picked articles and pages to take you deeper.
How to Read a Horse Racing Form Guide Without Spending Three Hours
A structured reading order for horse racing form guides — going, distance, class, connections, recency, and race context — so the analysis takes minutes rather than hours.
StrategyHow to Use Horse Racing Data to Build a Consistent Betting Approach
Which racing data points actually carry weight, why manual application drifts over time, and how algorithmic scoring fits into a consistent betting routine.
ExplainerHorse Racing Tips vs Algorithmic Selections: What Is the Difference?
How traditional horse racing tipsters and algorithmic selection services differ in methodology, transparency, and what you can actually verify before subscribing.