TracksideRacing.AI began with a familiar Saturday-morning problem: too many tabs, too many spreadsheets and too much time spent assembling information before the real analysis could begin. A racecard lived in one place, course records in another, speed ratings somewhere else, and pedigree notes in a file that only made sense to the person who built it.
None of those sources was useless. The problem was the distance between them. Every switch of context made it easier to miss a relevant detail and harder to repeat the same process from one race to the next.
We did not want another page that simply names a selection. We wanted the research desk we kept trying to build for ourselves.
The purpose: make evidence easier to use
The platform is designed around a simple idea: racing analysis improves when the right evidence appears at the moment it becomes useful. Today's Card starts with the declarations for a single day's racing. From there, a user can inspect the runners, compare past performance, investigate the course and move into pedigree or speed-figure history without rebuilding the question each time.
That does not make racing predictable. It makes the analysis more organised. There is an important difference.
We want TracksideRacing.AI to help answer questions such as:
- Has this horse already produced a performance strong enough for today's level?
- Does its best form come under conditions that resemble today's race?
- Does the course favour a particular draw position or pace profile at this trip?
- What does the wider family suggest when the horse itself has limited evidence?
- Is a strong headline figure supported by recent, relevant performances?
The software should reduce the work needed to reach those questions. It should not decide the answers on a user's behalf.
Why another rating?
Trackside Rating—TSR—is our accessible 0–140 performance rating, where a higher number represents a stronger historical performance. It gives users a common reference point across a horse's form and makes it easier to see career peaks, recent direction and how previous runs compare with today's field.
A rating is useful because it compresses information. That is also its limitation. One number cannot fully describe pace, suitability, fitness, price or what might happen today. We therefore present TSR alongside the form that produced it, not as a sealed prediction that users are expected to trust.
Our more detailed speed-figure tools follow the same principle. A figure should be traceable back to a horse, a date, a course, a trip and a race. Context is part of the product, not an optional extra.
Built around the work, not the spectacle
Racing products can easily become noisy: more widgets, more alerts and more claims competing for attention. We are taking a more deliberate route. Features have to earn their place by supporting a recognisable part of the research process.
- Orientate: see today's meetings and decide which races deserve attention.
- Compare: review the runners on consistent measures while retaining the underlying form.
- Investigate: explore course patterns, family evidence and performance history where they add context.
- Decide: make your own assessment, including whether the available price reflects the uncertainty.
This is why some of our most distinctive work is not a tipping module. Family Tree, for example, turns a pedigree into an interactive research surface. Course Analytics makes historical patterns explorable rather than reducing a course to a single label. These tools help a user ask better questions.
Transparency matters
Racing analysis is full of hindsight. A method can look flawless if only its winners are shown. We want our public meeting reviews to take the opposite approach: state the measure used, restrict the calculation to information available before the race, show the full set of races and include the misses.
That standard will matter even more as we publish analysis around major meetings. A striking winner can illustrate what a rating surfaced, but it cannot prove that a method will keep finding winners. Useful evidence includes where a signal failed and what other factors may have mattered.
What we are building towards
Our aim is a platform that can serve the enthusiastic weekend form student and the experienced analyst without forcing either into somebody else's opinion. The free experience should be genuinely useful. More advanced tools should add depth, automation and flexibility—not hide the basic evidence needed to understand the product.
There will always be another feature we could add. The test is whether it makes the work clearer, more repeatable or more revealing. If it does not, it probably does not belong.
The short version
TracksideRacing.AI exists to bring serious racing research into one coherent workspace. It organises the evidence, keeps the context attached and leaves the decision with you.
TracksideRacing.AI provides data and analytical tools, not guarantees or personalised betting advice. Past performance is not a reliable guide to future results. Please gamble responsibly. 18+.