The Greatest Racehorses of All Time

Every racing conversation arrives at it eventually, usually late, usually with more warmth than rigour. Who was the greatest? The question survives because it cannot be settled, and it cannot be settled because nobody agrees what it means.

There are at least three questions hiding inside it. The first is about peak ability: on his best afternoon, how good was he? The second is about sustained achievement: how much did he win, against whom, over how long, carrying what? The third has nothing to do with either: how deeply did the horse lodge itself in the public imagination? A serious answer has to declare which question it is answering, because the three produce different winners.

And all three are haunted by the same statistical problem. Comparing horses across eras means comparing different populations, different training methods, different track surfaces, different timing technology and different breeding pools. Ratings help — Timeform has been publishing them since 1948 — but a rating is a measurement of a horse against its own contemporaries, not against the ghosts of another decade. The trap is familiar to anyone who has read about speed figures: numbers earned under one set of conditions do not travel cleanly to another.

With that caution stated and promptly ignored, as is traditional, here is the argument.

The flat pantheon

Frankel

The modern reference. Frankel ran fourteen times and won fourteen times, ten of them at Group 1 level, and he retired in 2012 with a Timeform rating of 147 — the highest the organisation has ever awarded a flat horse in more than seventy years of doing so. His official British rating, 140, remains the ceiling of that scale too.

The performance people reach for is the 2000 Guineas of 2011. Sent to the front by Tom Queally, he was clear of a good field by a distance that looked like an error of judgement, then simply kept galloping and won by six lengths — the widest margin in the race since 1947. He beat Canford Cliffs in the Sussex Stakes, won the same race again a year later, and signed off with the Champion Stakes on ground he did not want. There is no dark mark on the record because there is no defeat to explain.

Sea-Bird

Timeform's second-highest flat rating, 145, belongs to a French colt who raced in 1965 and beat the best horses in Europe with an air of not especially noticing them. He won the Derby, and then the Prix de l'Arc de Triomphe by six lengths against a field of Classic winners that is still cited as one of the strongest ever assembled. Those who saw him have never entirely accepted that Frankel was better.

The British greats

Brigadier Gerard won seventeen of his eighteen races between 1970 and 1972, beating a Derby winner and a 2000 Guineas field of exceptional quality; his single defeat came over a trip beyond his best. Mill Reef, his contemporary and in some minds his superior, took the Derby, the Eclipse, the King George and the Arc in one extraordinary year. Nijinsky, in 1970, became the last horse to win the English Triple Crown — the 2000 Guineas, the Derby and the St Leger — a feat now more than half a century old and unlikely ever to be repeated, because nobody campaigns a champion that way any more.

Dancing Brave's 1986 Arc, produced from an impossible position with a burst of speed that seemed to belong to a different race, is the most replayed finish in European flat racing. Sea The Stars had the most perfect season: in 2009 he won the 2000 Guineas, the Derby, the Eclipse, the Juddmonte International Stakes, the Irish Champion Stakes and the Arc — six Group 1 races in six consecutive months, each a different question, each answered.

Enable, the great modern mare, won two runnings of the Arc and eleven consecutive races, and lost the 2019 Arc, chasing an unprecedented third, only when soft ground took the finishing kick out of her legs and Waldgeist ran her down. Shergar won the 1981 Derby by ten lengths, the widest winning margin in the race's history. Two years later he was stolen from his stud in County Kildare and never recovered — an ending the sport has never quite absorbed, and which should not be allowed to overshadow what he did on a racecourse.

The jumps pantheon

Here the numbers stop being close.

Arkle

Timeform rated Arkle 212 over fences. The next horse on that list, his contemporary Flyingbolt, was rated 210; almost nothing else has approached either. Arkle won three consecutive Cheltenham Gold Cups in the mid-1960s, and he so distorted the handicapping of his era that races were framed with two sets of weights, one to be used if he ran and one if he did not. He gave enormous lumps of weight to top-class rivals and beat them anyway.

Two pounds beneath him sits Flyingbolt, trained in the same yard by Tom Dreaper, who would have been the horse of any other generation and had the misfortune to be foaled into this one.

The horses the public chose

Red Rum was not rated anywhere near Arkle and it has never mattered to anyone. He won the Grand National three times — 1973, 1974 and 1977 — and finished second in the two intervening years, a record of consistency in the world's most chaotic horse race that beggars belief. He came back from a debilitating foot condition to do it, trained on a beach at Southport, and when he died the sport buried him at the Aintree winning post.

Desert Orchid, a front-running grey who jumped with visible enthusiasm, won four runnings of the King George VI Chase and a Cheltenham Gold Cup in 1989 on ground he loathed, in weather nobody should have been racing in. He was adopted by a public that was not otherwise watching jump racing. Kauto Star won five King Georges and two Gold Cups, and did something no horse had managed before: lost the Gold Cup, then won it back. Best Mate took three consecutive Gold Cups at the start of this century and looked, each time, as though he had not been asked a serious question.

The global asides

Britain and Ireland do not own this argument. Secretariat's 1973 Belmont, won by thirty-one lengths in a time still unbeaten, is probably the single most astonishing performance in the sport's history anywhere. Winx won thirty-three races in a row in Australia. Black Caviar retired unbeaten from twenty-five starts. Any of them would have a serious case in a fully global reckoning; each raced in a different tradition, over different trips, against different rivals, which is precisely the problem this article keeps running into.

What the numbers can and cannot say

Arkle's 212 towers over the sport, and the honest question is whether it towers over the horses or over the era. He raced when the best chasers met one another constantly and when handicappers could still ask a champion to concede two stone. A modern champion is campaigned more carefully, over fewer races, against fields assembled to be beaten. Arkle's rating reflects a horse who was repeatedly asked impossible questions and kept answering them; no modern horse is asked.

Are today's horses faster? Probably, marginally, in the way that all elite athletes are — better nutrition, better veterinary care, better surfaces. Are they tougher? Almost certainly not. Frankel ran fourteen times in three seasons; Brigadier Gerard ran eighteen times in three. Red Rum ran a hundred times over his career. Durability has been traded for brilliance, and ratings measure only one of those.

Ratings tell you how far a horse stood above the animals it actually met. What lay beyond that horizon, no number has ever known.

This is why cross-era comparison is a parlour game rather than a measurement — a genuinely enjoyable one, provided everybody admits what they are playing. The same caution applies to the far more tractable business of comparing this afternoon's runners, which is where models earn their keep and where our piece on whether data can predict a horse race takes up the argument.

A verdict, offered for demolition

Five horses, ranked, with one line of defence each.

• 1. Arkle. The largest gap between a horse and its contemporaries that racing has ever recorded, achieved while being asked to carry the sport on his back. Rated 212 when 200 is immortality.
• 2. Frankel. Fourteen from fourteen, the highest flat rating in Timeform's history, and not one afternoon on which anybody could construct an excuse for the horses behind him.
• 3. Sea The Stars. Six Group 1 wins in six months across every distance and every kind of race a miler-turned-middle-distance champion can be asked to contest. The most complete single season on the flat.
• 4. Red Rum. Greatness measured in the currency the public actually uses. Three Grand Nationals and two seconds, from a horse who was not supposed to race again at all.
• 5. Sea-Bird. Because 145 in 1965, and an Arc won by six lengths from a field of champions, deserves better than to be a footnote to Frankel.

The list is wrong, of course. Every such list is. Someone will want Kauto Star in it for the Gold Cup he regained, or Mill Reef, or Nijinsky, or Secretariat on the grounds that thirty-one lengths is thirty-one lengths. Someone else will observe — correctly — that ranking a steeplechaser against a miler is a category error dressed up as an opinion.

That is rather the purpose of the exercise. The horses are gone, the clocks have been recalibrated, the tracks resurfaced, and all that survives is the evidence: a rating, a margin, a grainy replay, and the testimony of people who were standing there. Argue with the list. Racing has been arguing about it for eighty years, and shows no sign of tiring.

Sportily is a statistical research platform for racing fans. It does not provide betting advice, and past results never guarantee future performance.