Every year, the running of the St Leger seems to provoke the same tired discussion about the decline of stamina in the thoroughbred. “Years of breeding for speed…” often starts some cleverly worded polemic, bemoaning the flimsy staying power of the breed when compared to its predecessor from the halcyon days of the sport.

This is a fallacy, yet, amazingly, it has survived for over 150 years. In 1856, the author ‘Stonehenge’ published a book in which he stated that “the modern racehorse has become more precocious and speedier, but in proportion has lost his qualities of endurance.” And the same notion has been rehashed many times since.

But there has never been any evidence to support this notion. In 1946, the great breeder Federico Tesio pointed this out in Breeding The Racehorse.

Tracing the development of the winning times for the Derby and Oaks over the previous 100 years, he wrote: “Clearly, the facts point to the exact opposite of what the ‘good old days’ school would have us believe. Modern horses take distance better than their ancestors because they cover it in faster time.”

With the benefit of another 65 years of evidence since Tesio did his analysis, let’s see if his axiom holds true. Starting with 1950, in which Galcador won the Derby and Asmena the Oaks, I have plotted the winning time of the two Classics over the next 65 renewals, including this year’s events won by Pour Moi and Dancing Rain respectively.

Rather than present these graphs, in which the times are scattered according to the pace of races and going conditions, it is more powerful to present the mathematical measure of correspondence between the winning time and the year – the correlation coefficient which describes the slope of the ‘best-fit line’. Here they are:
Derby -0.33
Oaks -0.24

These numbers – a perfect correlation is 1 – describe a trend that as the year number gets higher, the winning times of the Derby and Oaks have got lower. The inverse relationship is the reason for the negative sign.

In other words, winning times are tending to get faster over the years (the trend for the Oaks would be similar to that of the Derby, except for the recent tendency for the home turn to be dolled out, increasing the distance of the Oaks). They coincide with Tesio’s findings.

Yes, five of the six fastest times in the Derby since 1950 have been recorded since 1995. And this has happened despite artificial irrigation.

As Tesio himself wrote, stamina is sustained speed. Though the modem thoroughbred is no doubt aided by improvements in factors such as shoeing, feeding and medication in running faster times, it cannot be said that the breed is declining in stamina; the only objective measure of its staying power suggests the opposite is true.

Of course, the improvement of running times shown by winners of the Derby and Oaks is only part of a trend obvious across all distances.

Consider the correlation between year and winning time for some other Group 1 races since the Second World War:
1,000 Guineas -0.52
2,000 Guineas -0.50
July Cup -0.52

The stronger relationship (more negative correlation between year and winning time) governing races over shorter distances exists only because there is a higher proportion of renewals of the July Cup and the Guineas run at a true pace, so the winning time is a better reflection of the winner’s capabilities.

“Even without the best three-year-olds, the St Leger is seeing winners run quicker”

In selecting horses for stallion duties and evaluating the potential of their offspring, it is fair to suggest man has favoured speed and precocity over stamina for commercial reasons. But the breed has still proved resilient to a degree: consider that the best sire of the modern era, Galileo, and that of the previous generation, Sadler’s Wells, were both middle-distance horses.

Finally, what of the St Leger itself? First run in 1776, its cachet was once enough to attract the best horses of each generation. Nowadays, connections of the best three-year-olds run shy of sending their horse to Doncaster on the basis that merely taking part may encourage the perception they will sire offspring which is not commercially appealing.
However ridiculous the logic of this situation is, it must surely have resulted in a decline in the winning times of St Leger winners, right? Er, wrong.

Even St Leger winners are running faster times. Expressed statistically, the correlation coefficient between the winning time of the St Leger and the number of the year is -0.34.

As Tesio would have said, St Leger winners are now able to maintain a higher speed over the same distance, despite the best three-year-olds avoiding the race. It seems to me it is a good job we have been ‘breeding for speed’ because stayers are getting faster too.

Like Tesio established in 1946: modern horses handle distance better than their ancestors because they cover it in faster time.

2010 Workforce 2:31.33
1995 Lammtarra 2:32.31
2001 Galileo 2:33.27
2003 Kris Kin 2:33.35
1967 Royal Palace 2:33.36
2004 North Light 2:33.72
1936 Mahmoud 2:33.80*
1988 Kahyasi 2:33.84
1998 High-Rise 2:33.88
1987 Reference Point 2:33.90

1993 Intrepidity 2:34.19
1982 Time Charter 2:34.21
1989 Aliysa 2:34.22+
1980 Bireme 2:34.33
1927 Beam 2:34.60*
1988 Diminuendo 2:35.02
1934 Light Brocade 2:35.20*
1976 Pawneese 2:35.25
2009 Sariska 2:35.28
1952 Frieze 2:35.26

* hand time
+ disqualified for prohibited substance