We are often asked the question....'if you were using these tests to buy your own horses, which ones would you buy?" Not a bad question to ask....sort of like, put your money where your mouth is. The answer is actually one that we have researched quite extensively.
In statistical terms it hovers around the measures of performance of sensitivity and specificity. We have talked about this before but sensitivity measures the proportion of actual positives which are correctly identified as such (e.g. the percentage of elite racehorses who are correctly identified as being elite) while specificity measures the proportion of negatives which are correctly identified (e.g. the percentage of non-elite racehorses who are correctly identified as not being non-elite). A perfect predictor would be described as 100% sensitivity (i.e. predict all elite racehorses from the elite group as elite) and 100% specificity (i.e. not predict anyone from the elite group as non-elite). Nothing in genetics, at least as it is currently measured has this type of sensitivity and specificity and for any test, there is usually a trade-off between the measures. For example: in an airport security setting in which one is testing for potential threats to safety, scanners may be set to trigger on low-risk items like belt buckles and keys (low specificity), in order to reduce the risk of missing objects that do pose a threat to the aircraft and those aboard (high sensitivity).
For us, sensitivity is the most important. We judge our performance based on the number of elite horses that we correctly identify as elite racehorses. For those that have taken a look at our genetic rating system, you will know that we score class from 0 to 100, with 0 being predicted as a slow horse and 100 the elite racehorse (regardless of distance) and the same 0 to 100 scale for distance with a 0 being an absolute sprinter and a horse scoring 100 wanting a mile and a half to be at its best.
Having now tested around 1000 horses we can say that within our scoring there are higher pockets of predictability. While we generally see horses that score a 70 or above on class becoming stakes winners at a much higher rate than those that score 30 or below, there is a smaller subset of scores that result in a much higher rate of stakes winners to horses tested - those that score a 95 or above on class.
We could easily supply you with a list of names of horses that we have tested that have scored a 95 or above, but for many their racing outcomes are known and for some the test we do was actually developed off their genetic profile so it wouldn't be a true representation of what our genetic test is about if we did. So, below is the list from the foals born in 2010 that are racing this year in the Northern Hemisphere that score a 95 or above on class that we tested last year and this year at the two-year-old in training sales.
There are 16 of them (I have left Double Zap on the list as the 17th, just to show where the next horse falls in the list) which represented about 7% of the horses that we tested, but as we tested a very commercial population (horses that had already been selected by an agent or client in what are select sales or horses off a commercial farm) it is unsurprising that they are 7% of the horses.
In the real world these 95+ horses are about 3% of the commercial population so in practical terms, if there are around 3000 horses at the Keeneland September sale this year about 90 of them will score 95 and above on our genetic test. Will all of those 90 become stakes winners? - no, certainly not. As I said no genetic test has 100% sensitivity and we have to deal with horses breaking down, getting a bad trip during running, etc, but based on our modeling about 30 of them would become stakes winners or show the potential to become one. So, of this list of 16 below, we are expecting that 5 of these should be stakes caliber horses.
Like everything there are some caveats. These scores represent our genetic test only. We haven't included any of the ratings for cardio and splenic capacity where we have done them. Some of them might have had bad x-rays (we know Gas Tank did for sure) or moderate throats that compromise their careers, but they are what they are - horses that score outstandingly well on our genetic model. It is fairly straight forward to read the scores. They are from 0 to 100 on class and the same on distance. Horses that score under 30 on distance are generally sprinter/milers, 30-75 are mile to a mile and a quarter horses and 75+ are true route horses.
Let's see how they go.
22nd June 2016 - So here is the update on the horses now that they are now six year olds.
Indian Style - unplaced in 6 starts
Illimitable - unraced
Chief Lion - Stakes Placed winner
Rose Junction - winner
Gas Tank - Stakes Placed winner
Lovely Daughter - Winner
GRAND TITO - G2 winner
JP girl - unraced
Forin Sea - Stakes Placed winner
Ghareeb - winner
Missed Count - winner
UNHEDGED - Listed Winner
Harborinthetempest - Winner
SHORE RUNNER - Listed Winner
HILLABY - G2 Winner
Comes the Dream - Stakes Placed winner
So, from 16 selections that rated 95 or better on DNA alone, we have ended up with 4 stakes winners (25%) and 4 stakes placegetters. When I did the post back in 2012 we were expecting 5 to be stakes caliber horses and it looks as though we exceeded that.