A written verdict on every UK and Irish race, published at 8am before the first off. Nothing is ever deleted, including the ones that lose.
📊 Recent AI Results
Every AI selection stays public after the race — winners, losers and everything in between.
Return at 1pt level stakes
+24.5%
The dot is the measured return. The line is where the true figure plausibly sits.
95% range across 109 settled picks, resampled from the actual results rather than assumed to be bell-shaped. It includes zero, so this record is not yet distinguishable from breakeven.
New tips every day at 8am. Today's aren't up yet, so here's what we picked yesterday.
Published at 8am UK time, before the first race. A written read on each race, not investment advice. The ones we back settle into the record at the top of this page, win or lose; the rest are marked and scored separately below.
Board picks
Lingfield · Apprentice Handicap Stakes
You get
13/8Not bets · 14 more the AI read
The model named a horse in each of these but didn't back it. Either the price sat outside what we'll take, or the conviction wasn't there. None of them count in our record. Shown because the read is still worth having. Won/Placed/Lost badges below only describe how the horse ran.
Ffos Las · Happy 70th Birthday Brian Glanville Handicap Stakes
You get
5/2You get
4/6Catterick · Jacky Rylance 60th Birthday Celebration Handicap Stakes
You get
9/4Catterick · Hannah Blanchard 21st Birthday Handicap Stakes
You get
9/4Catterick · Racing Again 17th August Handicap Stakes
You get
5/2Ffos Las · Golwg Gwendraeth Handicap Stakes
You get
6/4Roscommon · Curragh To Curragh Charity Cycle on Saturday, August 29th Maiden Hurdle
You get
6/4You get
5/2Roscommon · James Lohan Engineering Excellence Maiden Hurdle
You get
7/4Lingfield · John Cunliffe Obf 80th Birthday Handicap Stakes
You get
10/3Lingfield · Trusted Betting Site Comparison gambling.com Handicap Stakes (gbbplus Race)
You get
evnRoscommon · Kepak Handicap Steeplechase
You get
11/4Lingfield · Best Betting Sites At gambling.com Handicap Stakes
You get
2/1Every race we read and passed, with the horse we would have named and why we left it alone. Kept on the page because a tipster who only shows you the bets is showing you half the record. Tap a horse for the full read.
| Time | Silks | Our call | Race | Why we passed | Result | Calls won | Price |
|---|---|---|---|---|---|---|---|
| 14:15 | National Pride | Catterick, Army Benevolent Fund/ebf Novice Stakes (gbb Race) | National Pride's sharp move from 5/1 to 9/4 already tells the story the market wants told, and backing a horse purely because it steamed in off a sixth-place run isn't a case, it's noise; Waakabb sits as the more secure form watch with a third-place finish and a rock-steady price, but at 7/4 there's nothing generous enough there to back with conviction either, and Palace Legacy's drift out to 9/2 takes it out of the picture. With the two market leaders essentially cancelling each other out and no isolated angle on either, this is one to watch rather than play. | Lost | 31% of 621 | 9/4 | |
| 15:00 | Perfect Nation | Ffos Las, Ebf Novice Stakes (gbb Race) | Perfect Nation heads the market on solid form and a going-friendly pedigree line, but the Boughey/Loughnane combination is a live negative angle this season and that's exactly the pairing running here, so the confidence you'd want in a 11/8 chance isn't there; Avionics has the second-best form figures and a strong course trainer stat but nothing pushes him past a solid but unspectacular contender, and the rest of the field (Midgham Man on a bookie-above-exchange steamer, George Cross with no form to judge) offer more questions than answers. | Won | 31% of 621 | 11/8 | |
| 15:30 | Cypriot Diaspora | Ffos Las, Ken Meredith Memorial Race Handicap Stakes (Div I) | Reruns named different horses | Lost | 10% of 67 | 3/1 | |
| 16:00 | Rosemary's Rose | Ffos Las, Ken Meredith Memorial Race Handicap Stakes (Div II) | Rosemary's Rose looks the one to beat on solid recent form (2-2-3 across her last three), a red-hot yard booking Oisin Murphy, but her price already reflects all of that and softening ground brings no fresh case for her or for market rivals like Kaleidoscope Eyes, whose sire profile actually prefers firmer ground. With three of the drifting outsiders (Manhattan Mirage, Venturing, Speculative) also carrying going concerns as the forecast turns damp, nothing in this Class 6 nine-runner scrap offers a clean, backable edge. | Lost | 31% of 621 | 15/8 | |
| 16:30 | Arbaawy | Ffos Las, Dandara Homes Maiden Stakes (GBB Race) | Three genuine contenders are bunched at the top of this maiden and none of them offers a clean case: Arbaawy's string of places (-43340) and course-suited pedigree are undercut by Billy Loughnane's rides running well behind market expectations lately, while unraced types Flagbearer and Wootz both arrive with strong yards but zero public form to actually assess against that money. Add in a track where this system has struggled (4/18) and a going call that could swing toward or away from the Wootton Bassett-sired runners depending on how much rain falls, and there's no runner here worth backing with conviction. | Lost | 31% of 621 | 7/4 | |
| 17:48 | Langdale Lass | Lingfield, Sky Sports Racing Sky 415 EBF Restricted Maiden Fillies' Stakes (For Horses In Band D) (EBF Restricted Race Qualifier) (GBB Race) | Langdale Lass sits well clear at odds-on off a solid debut second, but she's already priced like the best horse in the race and I can't find daylight between what she's worth and what she's asked to be backed at; behind her it's a muddle of one-run maidens and unraced types (Wild Orchid, West To Rest, Fortunate) where nothing stacks up strongly enough to take on the favourite. Ten runners, thin form lines, and a jockey-context 12/6 stat that doesn't move the needle enough to build a case on. | Lost | 31% of 621 | 4/6 | |
| 18:18 | Jazzy Blue | Lingfield, Sky Sports Racing Sky 415 Ebf Restricted Maiden Fillies' Stakes (For Horses In Band D) (Ebf Restricted Race Qualifier) (GBB Race) (Div II) | Jazzy Blue and Minnal effectively share market leadership on fair prices, with Lushill's improving placed form (532) also in the mix, and nothing in the form book or today's conditions cleanly separates the top two; Jazzy Blue's串 6252 profile shows a filly who keeps knocking on the door without winning, while Minnal's 64 is thin claims-wise for a market leader, leaving too many plausible outcomes to back one with conviction at these prices. | Lost | 31% of 621 | 9/4 | |
| 18:30 | Sound Judgement | Roscommon, DNG Ivan Connaughton Handicap Hurdle | Fifteen runners and at least four have live claims: Sound Judgement backed up her recent win for the powerful Elliott/Kennedy team, Knocknagoran Lady and Mr Dibbs both arrive on the back of victories too, and She'll Be Kept has won last time out as well — too many overlapping in-form types in a big handicap field to single one out with confidence. Add in four fitness question marks returning from long absences and first-time headgear on two others, and the picture is muddier still. | Lost | 31% of 621 | 7/2 | |
| 19:00 | Hashtag Pretender | Roscommon, J.F. Hanley Groundscare Beginners Steeplechase | Hashtag Pretender heads the market on the strength of a placed record and a red-hot yard, but a fall last time out in a beginners' chase is exactly the kind of jumping question that can resurface, and at 11/8 there's no room to absorb that risk; Impero has the profile on paper — best rating in the field, a pedigree that likes this softer ground, and first-time cheekpieces suggesting intent — but he's yet to win and the headgear switch is as much an admission of a problem as a fix. With three of the seven carrying recent jumping errors (Hashtag Pretender, Killeemore Lad both fell last time) this is a beginners' chase where the form book can't be trusted, and I won't back a horse into that uncertainty at a short enough price to leave no margin for error. | Won | 31% of 621 | 11/8 | |
| 20:00 | Il Silenzio | Roscommon, Campbell's Bar Ballinlough Handicap Steeplechase | Sixteen runners over three miles on ground that's just gone from good to yielding to soft turns this into a stamina lottery, with three first-time headgear horses, two fitness-unknown returners in Decimation and John Storey, and a form book too scattered to trust; Il Silenzio looks the classiest on recent placings but a 7lb claimer and a stack of unexposed types behind him make it too murky to back with conviction. | Won | 31% of 621 | 3/1 | |
| 20:18 | D Day Major Winter | Lingfield, Download The At The Races App Handicap Stakes | Six runners in a shabby class six with the market split three ways: D Day Major Winter looks the one they fancy on quick ground pedigree and a strong record for James Owen's yard at this track, but his fair price already reflects that, Offiah's Boy is fresh off a quick turnaround and a clear second last time out yet trained by a yard firing blanks, and Gert Lush drifted a full point overnight despite decent recent form figures. None of the three separate cleanly enough on the numbers to risk backing at the price on offer, and a fourth runner in Monks Mead can't be dismissed either given the yard's recent handicap hit rate. | Lost | 31% of 621 | 13/8 | |
| 20:30 | Mountain Sky | Roscommon, Brian Keenan On-Course Bookmaker Flat Race | This Roscommon bumper is a genuine puzzle: Iron Sword sits favourite off a single run and a plain 5th, Mountain Sky's 5-2 form is progressive but unproven at this trip, and Lola O C is unraced for a top yard at a short price — three live contenders with nothing at the prices that separates them cleanly, on ground that's just been downgraded to Yielding To Soft with several runners' surface form unknown. Add a returning-from-273-days runner and half the field with no form figures at all, and there's no honest angle to isolate one horse here. | Lost | 31% of 621 | 6/1 |
“Calls won” is the record of our most-likely-winner calls on passed races, scored against those horses’ own market prices. It is a separate record from the bets, and a weaker one: across every passed race the call wins at roughly the rate its price already implied. How we measure.
We read every race more than once. When those reads name the same horse, the verdict has been right far more often than when they disagree. That gap is wide enough that it belongs next to the verdict itself rather than in a statistics page nobody opens.
We only stake the top tier. The rest are published anyway, because a record that shows you only its confident calls is not a record.
Every rerun agreedStaked
The model read the race again and named the same horse each time. These are the only ones we stake.
211 settled picks
36%
Reruns split
Some reruns named a different horse. Published, never staked.
143 settled picks
24%
No agreement at allNo longer published
No two reads named the same horse. We ran the test before deciding: across 67 such races the calls we make on them won 10% against a 22% market expectation, so we concluded they are not worth backing. We no longer publish these as picks: the race becomes a pass instead.
Several figures appear across this site and they measure different things. Here they are together, each one put through the same calculation at 1pt level stakes.
Roughly, each row is wider than the one below it in what it demanded: we publish a read on every race, some clear the bar to be worth betting, and we stake a subset of those.
Board picks
The picks we actually put money on. Fewer than qualified — a daily cap until 3 July, and since then prices that moved or fell outside the band we bet.
Every published tip
Every pick we published, including the ones that never qualified. Most of these were never bettable.
Attributed to a course
The total across all 90 course pages. A tip counts once per course page it appears on, so this is deliberately not a count of distinct tips.
Closing line value sits outside this table because it is not a return: +1.1% mean against the closing exchange price across 85 priced bets, beating the close on 45.9% of them. Measured against Betfair SP rather than bookmaker SP, which carries the overround and flatters this figure. How we measure.
