Horse Racing's Biggest Problem Isn't Finding Information. It's Knowing What to Trust.

Horse racing has never had more data, opinions, or analysis. The problem is knowing what you're actually looking at.

Most racing fans don't have a problem finding information. There are ratings, tipsters, form guides, statistics, databases, communities, experts, and countless websites offering their view on today's races. And that's supposed to make analysing a race easier.

Sometimes, it does the opposite.

Take something as simple as the draw. You look at one reputable racing source and it tells you that a low draw has historically been an advantage at a particular course. You check another reputable source and it says the opposite: higher-drawn runners have the advantage.

So which one is right?

The uncomfortable answer is: they both might be.

One source could be looking at every race at the course over the last ten years. The other might be looking specifically at today's distance, going, and field size over a 5-year window. They're not necessarily disagreeing about the data—they're looking at different slices of it.

Unless you can see exactly what is being measured, over what period, under what conditions, and why, you're left with two conflicting answers and no real way to judge them.

That's the problem with modern racing analysis. It's not a lack of data. It's a lack of clarity around the data.

Racing Has a Trust Problem

The racing industry is built around people who spend years learning the intricacies of horses, trainers, jockeys, tracks, conditions, and form. But wherever information can potentially make someone money, there is also an incentive to sell certainty.

Tipsters show their winners. Ratings highlight impressive numbers. Data services hand you hundreds of statistics. But the core question remains: what exactly am I looking at, and how was this conclusion reached?

When information becomes fragmented across multiple sites, you end up with genuine information that isn't clear information. And without clarity, trust is impossible.

That's Why We Built RaceChat AI

We didn't want to build another black-box rating that simply tells you what to back. We wanted to build something you could question — and get a clear, conversational explanation from, rather than a wall of data you're left to interpret yourself.

Ask about a race. Ask why it likes a horse. Ask why it doesn't like the favourite. Ask about the draw, the going, the pace, the trainer, or the distance. Ask a follow-up.

Transparency creates clarity. Clarity creates trust.

We won't tell you RaceChat will make you rich, and we won't guarantee winners. What we can do is show you what we're analysing, explain how the model thinks, publish every prediction, and let you judge the results for yourself — wins and losses both. No cherry-picking. No quietly removing losing runs.

What Transparency Actually Looks Like

Instead of opening a race card, a form guide, and three different odds comparison sites to build a picture of a race, you just ask.

User: "What's the best value bet on today's card?" RaceChat: Identifies a specific runner in a specific race and explains the gap between its estimated winning probability and the current market price.

User: "Why not the favourite?" RaceChat: Walks you through why the favourite's short price doesn't match what the model sees in the recent form and pace shape.

That is the shape of the product: a conversation, not a scoreboard. You can follow up, push back, or explore another angle—just like you would with a knowledgeable friend who has actually done the homework.

What RaceChat AI Delivers

• Live UK & Irish Racing Data: Cards, form, results, and odds, refreshed daily—not a static database.
• Elite v14 AI Model: Ranks every runner in the field, trained and calibrated on historical results rather than tipster intuition.
• Personalised Daily Brief: A tailored view of the trainers, jockeys, courses, and horses you actually follow.
• Bet Tracker: Records your own picks against results so you can monitor your personal performance over time.
• Plain-English System Builder: Describe a strategy (e.g., "Horses ranked in our top 2 between 3/1 and 8/1") and backtest it against historical data before risking a penny.

Performance & Results: The Live Record

We don't ask you to trust a description. Here is the unfiltered record from our live model:

The +11.93% ROI isn't a hand-picked month; it includes every qualifying selection over the full quarter. When a newer version underperforms in testing, the older model stays live. That isn't a marketing decision—it's the only way a performance number remains meaningful.

Put It to the Test

You don't need to take this page on faith. RaceChat is live right now. Ask it about today's card, challenge an answer, and compare what it says with what actually happens on the track.

Prefer to check the numbers first? See every prediction settled, including the methodology →

One more thing

We're building RaceChat because we think racing analysis should be something you can question and verify — not something you're expected to just believe. We haven't solved everything. We're building this in public: publishing the losses alongside the wins, and improving the model as we go.

That's what the Founding 1,000 actually is. Not just early access — the first people who get a say in what this becomes.