Most stallion rankings fail to capture the one question rankings really should answer: how good is a sire right now?
It’s no good merely aggregating statistics because the resulting rankings will be biased towards older sires who have had more crops.
And the problems tend not to end there. If the metric chosen for the rankings is earnings, for instance, the true proportionality between races is lost because their value follows what is called a power law. In other words, the scale from the lower-valued races to higher-valued increases exponentially, like the force created by earthquakes. And, while the Richter scale correctly allows for this by using a logarithmic transformation (each point is a factor of ten), your friendly neighbourhood stallion rankings often use, well, merely a plain average.
Then there is the problem of allowing for the rarity of winning a Group or Graded race. All other things being equal, a stallion who gets ten Grade 1 winners in the U.S. is likely less potent than one who produces ten Group 1 winners in the UK. This isn’t difficult to understand: considering the relative populations of horses with a Racing Post Rating of 115+, there are far fewer races to go at in the latter country.
And, so the list goes on….
Most stallion rankings are really designed for advertising purposes. But TRC Global Rankings are different: our focus is to predict winners of the world’s greatest races right here and now, so we seek to evaluate the present value of a stallion’s impact: how good is he now?
And that’s why, if you look down the rankings of stallions we produce, you will find names who sometimes won’t feature in other lists for a good few years. And, the irony is, by that time you won’t need the numbers.
All this brings us to Frankel. In this week’s rankings, he sits at #13, up six places after gaining four points for the G3 Rose of Lancaster win of his son Frankuus. He’s ahead of stallions like Street Cry, Fastnet Rock, Shamardal and Kitten’s Joy, who are upper-echelon producers with many more representatives running round.
But, get this: Frankel’s TRC Global Rankings points do not include those he will earn for this week’s victories at Deauville of Eminent in the Group 2 Prix Guillaume d’Ornano and Lady Frankel in the Group 3 Prix de Lieurey.
TRC Rankings learns from what are known as ranking violations. These occur when any lower-ranked jockey, owner, trainer or sire defeats a higher-ranked one. The algorithm tries to determine why it happened: is racing in a certain part of the world undervalued? Is a particular type of race stronger or weaker than officially described?
Had Frankel’s standing being merely been a function of his aggregated wins as a sire, the rankings algorithm would have made a big mistake prior to those Deauville wins. Because the son of Galileo was ranked ahead of more established sires like Dansili, Invincible Spirit, Pivotal and Holy Roman Emperor – all of whom have sired more Group and Graded winners – Frankel was our #1 ranked sire in both Eminent’s race and Lady Frankel’s.
There were no ranking violations, in other words.
Frankel is heading for the top. He has already passed Uncle Mo on our numbers, and he is firmly on course for the top ten soon. He’s carried on the great start that first suggested he could live up to the considerable hype of his all-time-great racing career.
You can follow the progress every week right here of all the world’s best sires – evaluated mathematically against each other using everything we can to make a probabilistic judgment of who is best now – including the group of sires Chris Smith called the ‘Golden Generation’.
Click here for a list of last week’s biggest TRC Global Rankings points gainers.
Click here for a list of the week’s Group and Graded winners.