Let it never be said that I pass up an opportunity to geek-out. If you doubt me, exhibit one is that I used a five-hour delay for my flight to Hong Kong last week, not in the bar, but in looking extensively into some of the striding figures provided by Total Performance Data elsewhere on this site.
It was probably less a matter of “get a flight” and more one of “get a life”. The reader may well wonder, with justification, why such study failed to produce something better than the three well-beaten selections that came in last week’s blog. Sorry.
What I confirmed was that peak cadence (stride turnover) does indeed have a strong correlation with stamina, and that stride length is affected by ability and circumstances – such as track conformation and state of the ground – but varies little by distance. A reminder that a horse’s speed is a product of the two: cadence and stride length.
The bad news is that peak cadence is good at identifying sprinters but less good at discriminating at longer distances, where striding categories overlap more. The following is a summary.
What this means is that if you have a horse capable of turning over its stride at more than 2.52 times per second it is probably a 5f performer (or Winx), but that if you have one whose peak frequency is 2.35 strides/second it could plausibly be anything from a 7f/8f horse (more likely the latter) to a 14f or further (more likely the former) horse. Hmmm.
Still, the point about analysis in this area – and many other areas – is to tip the probabilities in your favour, not to aim for utter but unachievable certainty. I am confident that will become more feasible as time goes on.
As a matter of interest, I have Cracksman hitting a peak cadence in the home straight of 2.33 and showing a massive stride of 26.9 feet (24.4 is average) in winning the Prix Ganay last weekend, while Gosden’s younger hopefuls Sevenna Star and Lah Ti Dar managed 2.32 cadence/23.7 feet and 2.26 cadence/24.6 feet respectively on softer ground in their recent wins.
That cadence for Lah Ti Dar, a daughter of Dubawi and Dar Re Mi, is encouraging, in that it suggests she may be able to extend the ability she showed at Newbury to trips significantly longer than the 10f she ran over that day. We shall see.
The next piece of analysis was prompted by reading some nonsense about “lucky” and “unlucky” stalls for horses in this coming weekend’s Kentucky Derby. This sort of thing helps to create a “narrative” but is a source of annoyance to anyone who believes in looking at data in a remotely intelligent way.
As has been pointed out many times before, wins alone are a particularly crude way of measuring whether an effect exists, unless your sample size is large (in which case the relevance to the here and now may well be questionable).
Besides anything else, wins alone fail to factor in field sizes which may vary substantially.
One option is to measure performance by % of rivals beaten, so that a winner beats 100% of its rivals, the last-placed horse beats 0% of its rivals, and every horse in between is on a scale of 0 to 100 depending on its finishing position and the size of the field. This is not perfect, but it is a massive improvement on many conventional methods.
Using %RB gives the following figures for the Kentucky Derby since 2000 (figures are an average of the stall in question and the stall on either side, for a draw is a discrete entity but its effect should not be).
The first thing to say is that there is not a huge amount of difference there. Figures between approximately 46% and 54% can usually be regarded as neutral. Only five stalls fall outside those parameters, and then not by very much.
Nonetheless, I would rather be drawn between stalls 3 and 15 inclusive than on either flank, all other things being equal.
I am pretty sure that MENDELSSOHN is good enough to win an average Kentucky Derby, but this looks to be a superior one, and the race will have a very different dynamic to the one he bossed so impressively – and with particularly fast early sectionals – in the UAE Derby. His draw in stall 14 is wider than connections might have hoped for, but the evidence is that it should not count against him, all other things being equal.
Speaking of the draw, a stall of 9 in a field of 9 would not normally be regarded as optimal at 7f at Lingfield, but there are a number of reasons why I think that CRACK ON CRACK ON may be able to overcome that in the 4:00 at Lingfield on Thursday.
Chief among them is that he promises to be a useful performer and gets in here off an attractive mark of 81, but also the prospect of a strong pace means he should be able to slot in and come late.
At two years, he finished fifth in a strong Newmarket minor race – turning in some sharp sectionals – before running Qaysar (now rated 98 by the BHA) to a length in a similar contest at Kempton, and an end-to-end gallop may further help his free-going style.