Every autumn a handful of yearlings sell for millions. Some become champions. Many never win a race — and a surprising number never run at all. Here are fifteen of the biggest misses from Tattersalls' Book 1 and the old Houghton sale, the reasons buyers keep paying, and what that means for anyone buying a yearling.
About £40.9 million at the hammer, before fees, training and keep.
Four more ran without winning. Only three won a race — five wins between them, and not one of them in a black-type race.
Which is the first clue to why the money was spent — several went on to be valuable broodmares.
All bought as yearlings at Tattersalls (Book 1, or the Houghton sale before it), ordered by price. "(f)" filly, "(c)" colt. Look at "The selling point": nearly every one was a close relation of a champion.
| Horse | Sold | Guineas | Sire × dam | The selling point | What happened |
|---|---|---|---|---|---|
| Al Naamah | 2013 | 5,000,000 | Galileo × Alluring Park (f) | Full sister to Oaks winner Was | 1 win from 7 starts; Group 3 placed |
| Poker | 2024 | 4,300,000 | Wootton Bassett × Park Bloom (c) | Dam a full sister to Was and Al Naamah | Winless after 6 starts (career continuing) |
| Maria Danilova | 2017 | 4,000,000 | Galileo × Dank (f) | — | Never ran; dam of a winner |
| Darain | 2018 | 3,500,000 | Dubawi × Dar Re Mi (c) | Full brother to Too Darn Hot | 3 wins (two novices, a US allowance race) |
| Skylark | 2020 | 3,400,000 | Galileo × Shastye (f) | Full sister to Japan and Mogul | Unraced (as last reported) |
| Liffey Dancer | 2007 | 2,500,000 | Sadler's Wells × Brigid (f) | Full sister to Listen and Sequoyah | Never ran; her family later produced Treasure Trove |
| Hydrogen | 2012 | 2,500,000 | Galileo × Funsie (c) | Closely related to Derby winner Authorized | 2 starts, unplaced |
| King Power | 2017 | 2,500,000 | Frankel × Prowess (f) | From the family of Oaks winner Talent | Winless in 7 starts; Listed placed |
| Dubai Sun | 2001 | 2,300,000 | Desert Prince × Park Express (c) | Half-brother to New Approach | Remained a maiden |
| Veruschka | 2015 | 2,100,000 | Dubawi × Loveisallyouneed (f) | Dam a sister to Yesterday and Quarter Moon | Never ran; resold as a broodmare |
| Nazym | 2011 | 1,700,000 | Galileo × Brigid (f) | Half-sister to Listen and Sequoyah | Never ran; dam of Musidora winner Nausha |
| Al Jassasiyah | 2012 | 1,500,000 | Galileo × Alluring Park (f) | Full sister to Was | Never ran; dam of Listed winner Mashael |
| Secret Gaze | 2016 | 1,350,000 | Galileo × Shastye (f) | Full sister to Japan and Mogul | Never ran; kept as a broodmare |
| Time Control | 2006 | 1,200,000 | Sadler's Wells × Time Away (f) | Dam won the Musidora | 1 minor win from 6 starts; dam of Group 1 winner Cursory Glance |
| Manayer | 2002 | 1,100,000 | Sadler's Wells × Dazzling Park (f) | Dam a daughter of Park Express | Never ran; produced winners |
Records as last reported; Poker is still racing, so his story may change. A disappointing racing career is not the whole story for a filly — see below.
Three of the fifteen were sold at sales whose catalogues are in the HorseBytes database, so we ran them through the HorseBytes System as yearlings, exactly as a buyer would have seen them on the day.
| Horse | Price | HorseBytes Index | Rank in the sale |
|---|---|---|---|
| Skylark (2020) | 3,400,000 gns | 100 | 1st of 547 |
| Al Jassasiyah (2012) | 1,500,000 gns | 100 | 6th of 577 |
| Hydrogen (2012) | 2,500,000 gns | 94 | 99th of 577 |
Two of them were at the very top of their sale on pedigree — any honest system would have ranked them there. And still nothing. That is the most important lesson on this page: a great pedigree raises the odds, it does not make success likely. Even in the top fifth of the HorseBytes Index, only about one young horse in sixteen goes on to win black type. Fifteen out of sixteen don't — however well bred.
Eleven of the fifteen were fillies. To a major breeding operation, a full sister to a Classic winner is worth a fortune as a mare to send to its own stallions, whether or not she ever wins. Nazym bred a Musidora winner, Time Control a Group 1 winner, Al Jassasiyah a Listed winner; Veruschka was resold as a broodmare. On the racecourse they missed; on the stud book many did not.
Almost every one was a full or half sibling to a champion — Was, Japan, Too Darn Hot, Listen, New Approach, Authorized. Buyers go back to the same mare hoping for lightning twice: the Alluring Park family alone took 10.8 million guineas across Al Naamah, Al Jassasiyah and Poker. But racing ability is only partly inherited, so a champion's full sibling is expected to be good, not great. The best one in a family is usually a one-off.
Ten of the fifteen are by Galileo or his father Sadler's Wells, the dominant sires of their day. When every major buyer wants the same pedigree, the price reflects the competition for it, not the chance of success.
A colt who wins a Group 1 can be worth many times his price as a stallion. Buyers of colts like Poker, Darain and Hydrogen were paying for that possibility — a lottery ticket with a very big prize and long odds.
Records happen when two of the world's richest operations want the same yearling. At that point the price is set by them, not by the pedigree or the horse.
Size, looks and walk are judged by the best eyes in the business, but none of it shows whether a yearling will stay sound, grow on, or want to race. Eight of these fifteen never reached the racecourse at all — often because of injury or simply because they showed nothing at home.
The most famous misses in the sport were American record-breakers of the 1980s and 2000s, and the story repeats wherever the money is.
| Horse | Sold | Price | What happened |
|---|---|---|---|
| Snaafi Dancer | 1983, Keeneland July | $10.2 million | First yearling to sell for over $10 million. Never ran — his trainer later called him "no bloody good" — and sired only four foals because of fertility problems. |
| Seattle Dancer | 1985, Keeneland July | $13.1 million | A world-record yearling price. Won two Group races in Ireland and was second in the Grand Prix de Paris: useful, but nothing like the price. |
| Meydan City | 2006, Keeneland September | $11.7 million | Won two of his first three starts, then nothing in five more. |
| Jalil | 2005, Keeneland September | $9.7 million | Won a maiden and a Group 2 in Dubai, then nothing in nine further starts. |
| The Green Monkey | 2006, two-year-old sale | $16 million | A record for a two-year-old at auction. Three starts, no wins. He had sold for $425,000 as a yearling a year earlier. |
| Noble Dynasty | 2019, Tattersalls Book 1 | 3,600,000 gns | Half-brother to Barney Roy. Five wins, but just over £58,000 in prize money. |
Galileo filly who won the Oaks at 20-1. Her full sisters Al Naamah and Al Jassasiyah cost more and did far less — the lottery runs both ways inside one family.
Danehill colt who won the 2,000 Guineas, the Queen Elizabeth II Stakes and two more Group 1s. Sometimes the expensive yearling is exactly as good as he looked.
A Group 1 winner bought for a tiny fraction of the prices on this page. Champions turn up across the price range — which is where evidence, not fashion, earns its keep.
This page deliberately collects misses. Plenty of million-guinea yearlings win Group races; the point is how often they don't, and how little the price tells you on its own.
"Sister to a champion" sells a yearling, but the close family's whole record — how many of the dam's foals have won, how the 2nd dam's family is producing now — tells you more than one famous relative.
A champion's sibling is likely to be good, rarely as good. Value the family's average, not its best.
The same evidence that pushes fashionable lots out of reach often points to sound, well-bred yearlings from unfashionable sires or young, unproven mares. That's where value lives.
If you want a racehorse, a filly's broodmare value is a bonus, not the point. If you're breeding, a moderate racemare from a top family can be exactly right.
Even the very best pedigrees succeed only now and then. Several sensibly bought yearlings give you far more chances than one record-breaker.
HorseBytes reports rank every lot in a sale on the HorseBytes System, flag the dark horses whose pedigree is stronger than their catalogue page looks, and show how the dam's other foals are actually doing on the track now — including the unraced youngsters. They can't judge the horse in front of you, but they help you pay for evidence rather than fashion.
Spotted something we've got wrong, or want a topic explained? We'd like to hear from you.