This week’s TRC Global Rankings for Horses reflect the results of the 2020 Breeders’ Cup races at Keeneland. Let’s examine how the changes came about and what they are implying.
TRC GLOBAL HORSE RANKINGS: THE TOP 15
By our reckoning, the Breeders’ Cup Classic was the contest with the most strength in depth. Before the race, we had #6 Maximum Security with a Performance Index (points) of 1265, with #9 Tiz The Law on 1249, #17 Improbable on 1216, #34 Tom’s d’Etat on 1170 and #36 Authentic on 1164.
Our system had Tiz The Law above Authentic even though they had made up the exacta in reverse in the Kentucky Derby because Tiz The Law’s earlier form – his Belmont win and, especially, his Travers romp – seen as a portfolio represented more compelling credentials.
Moreover, Authentic was turned over in the Preakness Stakes subsequently by the filly Swiss Skydiver, who was rated #12 going into the Breeders’ Cup Distaff and is now at #25 after her heavy defeat (in fairness, she missed the break).
Authentic rises no higher than #9 for his victory because the mathematics behind our rankings sees no justification for ranking him higher. TRC Global Rankings uses well-established maths to arrange the world’s best horses in an order consistent with the order of merit established by past results – but favouring relationships that are the most likely to be sustained in future races. Using no more than past results, this is the fairest way to construct a robust hierarchy.
Authentic is duly our #1 rated U.S. horse, but the system can only be so confident about this assessment.
He sits on 1252 (+88) now, to Tiz The Law’s 1247 (-2), with Maximum Security on 1243 (-22) and Improbable 1226 (+10). It’s easy to see how the last-named runner gains points – he finished second, which was better than we thought.
To refine its computations, TRC Global Rankings uses machine learning to tweak various parameters related to a horse’s form and its profile. Common to all so-called supervised learning techniques in this domain is a loss function – a way of keeping score.
Our loss function is the sum of all rankingviolations in the next week’s results. A ranking violation occurs when a lower-ranked runner defeats a higher one, and, when this inevitably happens, the loss function counts the difference between Performance Indices of the two horses as the magnitude of the error.
And this is why Authentic isn’t ranked higher. The system sees the order of the top U.S. dirt horses as a little fuzzy. Yes, Authentic has won the two best-known races in the Kentucky Derby and the Classic. But, the reputation of races is no bar to form being turned round in future races. And, before Churchill Downs, Authentic had not seemed anything like the same horse he does now.
Of course, what is needed is further confirmation of his talent. Sadly, he has been retired to stud attended by that common refrain “he had nothing left to prove”. While connections are, of course, free to do whatever they want with the horses they own, this claim is, shall we say, fake news.
For better of worse, the reason that Contrail is our world #1 for now, and rated 77 points above Authentic, is because his superiority over a large number of well-bred Japanese horses seems a lot more clear-cut. He is one of three Triple Crown winners in Japan this century, the others being his sire Deep Impact and former world #3 Orfevre.
Japanese horses have travelled to Australia and Hong Kong in the last few years and proved utterly dominant, so the level of Japanese form in the TRC Global Rankings database is relatively high. Those parameters referred to above have slid up since Orfevre because our system underrated Japanese runners somewhat for the races they won and, without overreacting, have moved the ranking of Japanese runners up in the attempt to reduce the loss function caused by future encounters.
Look at how many U.S. dirt horses are rated around Authentic. Japanese runners other than Contrail are ranked #5 (eight-times G1 winner Almond Eye), #23 (the unbeaten fillies’ Triple Crown winner), #27 (top-notch miler Gran Alegria, who has won her last two G1 races) and #36 (Chrono Genesis, a dual G1 winner with some of the strongest form in Japan). In short, there is no way we are systematically overrating Japanese form.
Contrail rates the best of these because his influence runs deep into the Japanese form book. While he hasn’t beaten older horses himself, he defeated old-rival Salios (a potential brilliant colt himself ranked no higher than #87 by us) In the Japanese 2000 Guineas by three lengths. That horse made mincemeat of some talented older rivals in the G2 Mainichi Okan at Tokyo last month.
There’s some concern th(at one of the metrics we use to rank horses, Impact Value (IV), tends to flatter Japanese horses because the field-sizes they face are bigger than elsewhere. We reject this notion completely, for two reasons: first, we downrate the influence of finishers further down the field (this helps Irish trainer Aidan O’Brien, incidentally, because we don’t penalise him heavily for running pacemakers) and, second, Salios himself has an IV of 8.64, which is higher than 22 of the top 25 horses but still sees him ranked no better than #87.
Regardless of how we might rationalise Contrail’s #1 ranking though, the maths is the maths. Rest assured, if lofty IVs were causing rankings violations, our system would tune them down. But they don’t. And, anyway, Contrail must face the trial of the track in the Japan Cup, where a showdown with Almond Eye and Daring Tact (and some other very talented older horses) awaits him.
If he is defeated there, he will take a hefty blow in the rankings (the change in rankings is proportionate to the so-called ‘statistical shock’ of future results).
After American Pharoah completed the U.S. Triple Crown in the 2015 Belmont Stakes, these were the next week’s rankings of June 7:
American Pharoah wasn’t #1 like Contrail, but we actually rated him an awful lot higher – his Performance Index was 1403pts to Contrail’s 1339. And, we can assure you, the level of our rankings is just the same. It’s just that, at that point, the best U.S.-trained horses were more dominant then then they are now: Shared Belief was rattling off 130 ratings for fun from our system.
Tarnawa could be the tops
While Authentic is heading to the breeding sheds, it would be a shame if Breeders’ Cup Turf winner Tarnawa follows suit. If her owner, the Aga Khan, bucks the form book and keeps her in training, she would immediately rate as the best horse in Europe next season, on our figures anyway. Surely she deserves at least one crack at the Prix de l’Arc de Triomphe.
Tarnawa was terrific in the Turf. She closed like a mustang off a slow pace, running the last quarter of the mile-and-a-half contest in 22.93sec.
Traditional handicapping techniques, which form ratings from the distance between horses at the finish, cannot possible evaluate her superiority properly; by contrast, TRC Global Rankings uses a nonlinear measure and is more concerned with the order of finish rather than the margin of victory.
Tarnawa now sits at #4 and, with Contrail needing to win the Japan Cup to hold his place at #1, and #2 Enable and #3 Ghaiyyath already retirees, she is a possible world #1 even before next year begins. That, of course, is only of academic interest, and what racing fans really need to see is more of a super talented filly like this on the racecourse.
Already confirmed as staying in training – despite an abbreviated career to this point – is the top-class Monomoy Girl.
Subsequent to her G1 Breeders’ Cup Distaff, she went through the ring for $9.5m but will remain in training with Brad Cox, presumably with another crack at the same race in mind.
Because our system takes the uncertainty a horse will reproduce its best form each time into account, a record with absences from the track counts against a horse. Otherwise, she would rank even higher than #10.