Racing Alpha v1.0: A Split Decision at 46 Races, But Profit Still Flows

A Day of Contrasts: Profit in the Numbers, Lessons in the Losers

Yesterday brought 46 races across Hamilton, Wolverhampton, Listowel, Ffos Las, Fairyhouse, and Leicester. For Racing Alpha's v1.0 AI card, it was a day of split fortunes — but the bottom line stayed black.

Of the six settled races on our card, two won, four lost, and the P&L nudged to +2.50 points. That's not headline-grabbing, but it's another day of positive return in a sport where consistency is the real prize.

The Winners

Aljoon (3.60) obliged comfortably, justifying the AI's confidence. More eye-catching was The Hare Rail (4.90), who made a mockery of market expectations. Both wins were exactly the kind the model is designed to find — value prices on horses the ratings fancied.

The Losers

We won't sugarcoat it. Midnight Call (15.00), Volto di Medusa (9.00), and Chianti (3.60) all failed to fire. Chianti, in particular, stung — the short price reflected strong model approval, but racing doesn't always follow the script. Maury (8.20) also came up short. Transparency is our brand, and these losses are part of the ledger.

You can track every result — wins, losses, and the running P&L — on our public performance page. No cherry-picking, no spin.

What This Tells Us

Two from six is a 33% strike rate, but with a positive flat-profit return. That's the edge — the model doesn't need to land every winner; it needs the winners to outweigh the losers at the prices. Yesterday, it did.

The broader card of 46 races also threw up plenty of non-selected winners, which tells us the model remains selective. It's not trying to predict every race — just the ones where the numbers give a clear edge.

Today's Racing

We're already looking ahead. Today's card is loaded with opportunities across the UK and Ireland, and the v1.0 AI is crunching the numbers now. If you want to see where the ratings point next, head over to today's racing for the full card.

As always, we publish every rated race, not just a single tip. Yesterday showed again that short-term variance is real, but the process holds. See you at the track.