Debunking the Curse of Apollo?

By Nicole Sauer

Not since Apollo in 1882 has a horse that was unraced at age 2 gone on to win the Kentucky Derby (gr. I). The famed "Curse of Apollo" will be put to the test once again this year, this time by Rebel Stakes (gr. II) winner Hoppertunity. Many argue that the right horse just hasn't come along yet, and it's only a matter of time until the curse is broken. But others are adamant that the maturity gained during a 2-year-old campaign is essential for a Derby victory.

The charts below compare horses that raced at 2 with those that did not and their subsequent performance in 3-year-old graded stakes from 1973-2013. During this period, 73% of graded stakes starters raced at age 2, while 27% were unraced as 2-year-olds. If "having a 2-year-old foundation" is important for graded stakes performance at 3, then we should expect a higher proportion of 3-year-old graded stakes winners to have raced at 2. This is the case, but only by a 2.2% margin: 75% of 3-year-old graded stakes winners raced at 2 compared to 25% who didn't.

Click for larger graph

Click for larger graph

Isolating only the Kentucky Derby starters, horses that raced at 2 also have a slight advantage. Though 8.4% of all Kentucky Derby starters since 1973 were unraced at 2, only 6.5% of in-the-money finishers were unraced at 2. This 1.9% differential again suggests a slight disadvantage for 3-year-olds lacking a 2-year-old foundation.

Click for larger graph

Click for larger graph

While it seems that whether a racehorse begins its career at 2 or at 3 has only a small impact on 3-year-old graded stakes performance, the fact remains that it has been 132 years since a horse that was unraced at 2 went on to win the Kentucky Derby.

The most likely reason for Apollo's curse still holding strong is simply because there have been so few to challenge it. In the past 31 years, only 55 horses that were unraced at 2 even competed in the Derby. Since 2000, horses without a 2-year-old start have finished second (Bodemeister, 2012), third (Curlin, 2007), fourth (Atswhatimtalknbout, 2003), and fifth (Wheelaway, 2000) in the classic. It likely is only a matter of time before Apollo's curse is broken.


Leave a Comment:


Hoppertunity has a chance to break the ole dreaded curse, too bad he comes in the gate with California Chrome. Another good one was lost due to injury, Constitution, and two on points,  Social Inclusion and Bayern. Seems to be more trying it, and even if the Apollo curse falls soon, the curse will just be worded different, " only one horse since 1882 " has won the Derby without racing at age 2.

24 Apr 2014 6:58 PM

That's an interesting slant you offer, Nicole. Your graphs seem to illustrate how difficult it is to ready an unraced 2 yr. old for the relatively early, long distance Derby (8.4% Derby starters vs 27% graded starters during their entire 3 yr. old year). I did some quick research, and found that about 43% of all "current 2 yr. olds" start at 2 (this is a % significantly greater than what is calculated by comparing 2 yr. old starters to registered foals). I also estimated that approx. 70% + "named foals" race by the end of their 3 yr. old years. So it does appear that there are a somewhat greater number of 2 yr. old starters that race at three than 3 yr. olds who were unraced at 2. This difference, though, appears to be far less than the disparity noted by your graphs re-graded stakes starters, where the differential is approx. 3-1. No doubt there are many variables in play to account for this; perhaps the chief being racing quality and conditioning, where the factor of conditioning exerts itself more (but, likely, still the lesser factor) relative to Derby entrants...My quick bit of research also indicated that during the past 13 years there has been essentially no change in the % of 2 yr. olds that have raced at 2. They have, though, competed in less races, time forward. This snapshot is temporally limited, and one wonders if the "Apollo Curse" may be short lived due to a lessening differential in the conditioning factor (raced vs unraced 2 yr. olds). For one, we do know that 2 yr. olds ran more often in years past; did also a greater % of the crop race at 2? Did more recent unraced 2 yr. olds perform better in the Derby than in years past?..I still believe that the 3-1 difference in graded stakes entries has more to do with the quality factor, than anything else. I hope, however, that none of this should suggest to some that the better path to graded stakes entry/success comes from racing at 2. 2 yr. old racing likely imparts no quality onto the horse, rather those with more inherent quality tend to race at 2. Let's face it, the goal for most owners is to race at 2, so in most cases it is a negative occurrence (sometimes a lack of demonstrated ability) that prevents it.  

25 Apr 2014 12:25 PM
Old Old Cat

What is the percentage of 2 year old starters NOMINATED for the Derby vs actual starters, and the percentage of non-2yo starters Nominated vs entries.  That is: Is the lesser amount of non-2yo starters an indication of subsequent 3yo training showing their deficiencies compared to 2yo starters.

25 Apr 2014 1:26 PM
Ian Tapp


I also believe that the historical 2YO vs. graded stakes figures have a significant correlation with quality. Good athletes tend to be precocious, so it makes sense that the better 3YO's had enough ability to race at 2.

That's why the Curse of Apollo is somewhat of a ridiculous concept. Like you explained, it's really about opportunity. If every horse in the field was unraced at 2, then the curse gets broken. If the Derby attracted 20 AZ-bred geldings one year, the superfecta would be four AZ-bred geldings and who knows how many curses that would break.

25 Apr 2014 2:15 PM
Pedigree Ann

Did you test your differences like 2.2% for significance? Just saying it looks small doesn't cut it; you need a rigorous statistical test to determine if it actually is smaller or larger than would be expected by random chance. Go back to your Stats 121 and review testing of hypotheses.

25 Apr 2014 2:16 PM
Ian Tapp


The results for 3-year-old GSW's having raced at 2 are significant to a p-value of 0.003.

The Kentucky Derby results are not significant, however, having only a p-value of 0.197. This is because there have been only 654 Derby starters and 123 top-3 finishers since 1973.

25 Apr 2014 2:25 PM
Pedigree Ann

Thank you for the clarification. That information, that one figure was statistically significant while the other wasn't, should have been part of the article, don't you think?

And significance is not a function of the sample size, but derives from the nature of the data. 654 or 123 data points are plenty for most statistical tests - did you do a chi-squared test?

25 Apr 2014 3:50 PM
Ian Tapp

Thanks for the suggestion, Ann. We could definitely improve our blogs with more statistical detail. I understand you are a former Statistics instructor at the University of Kentucky, so I'm sure you know your stuff. Thanks again.

25 Apr 2014 4:20 PM

Hi Ian:

Here's something that may be worth ruminating over a bit- You stated "good athletes tend to be precocious". I'd agree, but question whether we formed that belief from observations/results of the more distant past (re-racehorses). I haven't yet worked it out to my satisfaction, but here's a few thoughts: Alan had likened much of the present American thoroughbred population to a "sub-breed" of the world's thoroughbred population-a sub-breed of the precocious thoroughbred. This sub-breed no doubt came about through our attempts to selectively breed for precocity-we, here, value highly the trait of precocity. The rest of the world seems to be going in this direction as well (perhaps due to the weight of economic factors-hefty costs of upkeep, higher % breed for sale, etc.). Well, while it may be accurate to see a correlation between precocity and athleticism-when all else was relatively equal (similar to what was in years past); it may not follow that the selectively bred precocious-type is, in reality, the better athlete. The waters also become a bit muddied if this "selectivity" piggybacks speed with precocity. While these traits appear often to go hand and hand, there is yet (to my knowledge) no proof that they are not (genetically) independent from one another. I think in the old days there were more great equine athletes that managed to excel at an early age due mostly to their inherent athleticism ALONE, rather than for their genetic predisposition for precocity. Two who readily come to mind are Buckpasser and Graustark. Future history seems to suggest-via their records at stud- that neither had a genetic predisposition of precocity. These mere two examples are far from proof of this thesis, but offered to illustrate the point. Also, few would argue that today's less precocious European elite are not superior "athletes" to many of our (U.S.) elite. I could delve into this more, but this taste may be enough to stimulate some thought on your part.      

25 Apr 2014 4:46 PM

Whether or not there is a difference of a few % points in stakes winners (per the graphs) is basically irrelevant in this case. Why the focus-Ann?

25 Apr 2014 5:03 PM
Pedigree Ann

Why the focus? Because the author is trying to make a point with data and a valid statistical analysis can be a strong argument for or against that point, much better than a mere 'impression' or opinion about the numbers. A small difference to the eye can still be a genuine and important difference.

Yes, Ian, I spent many a semester as a TA, teaching the basic statistics courses to (mostly) business majors at the University of Kentucky.(Most of it went in one ear and out the other with that lot - they just wanted to complete the requirement.) After I finished my Ph.D., I was sporadically a PTI=Part Time Instructor, the wage-slave of the academic world, but found it didn't pay enough for the time it took. And with my husband's tax bracket where it is, we only got to keep half of it anyway. So I became a stay-at-home mum with an obsessive interest in TB bloodlines.

26 Apr 2014 8:44 AM

I rarely make reference to the alleged Apollo Curse. I just do not believe in its relevance. The reasons  Apollo's feat has not been repeated in such a long time has nothing to do with Apollo. Yes! I think Apollo has been getting a bum rap.

The number of horses in each Derby field that are not subjected to overcoming the alleged ‘Apollo Curse’ is significantly greater than those trying to do same. Base on available records, of the 131 Derbies since Apollo achieved the feat, 39 contained 58 horses who tried to overcome the alleged curse. Verazzano was the last great hope that failed to overcome the alleged curse. The annual disparity in the numbers does not favor those in the minority.

The reason Apollo's feat has not been repeated over such long period can be viewed from a simplistic angle. There is roster of prestigious 2YO races with lucrative purses as appealing options for the connections of fast developing 2YOs. Talented 2YOs viewed as being capable of winning races like the Hopeful, Champagne and BCJ, will encourage their connections to capitalize on early opportunities to secure returns on their investments.

Why wait to make 3YO debuts when the late developers are in abundance and competition is enhanced? Why not allow the animal to start contributing to its keep and care as soon as possible?

There have not been enough horses impacted by the allege curse in annual Derby fields to provide a reasonable chance of it being repeated. Those in recent times with the talent i.e., Verrazano, Bodemeister and Curlin just found either one or a few better on the day or didn’t say the trip.

26 Apr 2014 12:29 PM

Pedigree Ann:

Statistics isn't my field, so please correct me if I'm mistaken, but-

it appears your initial concern was that the 2.2% may not be statistically "significant". You profess to be an avid, long-time follower of thoroughbred racing. You would then have a fairly good idea (ball-park) of the numbers of 3 yr. old stakes starters, and winners over the forty year period, 1973-2013. Would not that knowledge, and your statistical background enable you to know, off-hand,that those numbers are large enough without having to question the significance of the reported 2.2%. Really, just asking, but recall that your initial post highlighted the 2.2%.  

26 Apr 2014 1:44 PM
Pedigree Ann

Sceptre - NO! Human perceptions are colored by so many intangibles that my impression or guess at whether a figure means something or not is not reliable. We tend to remember some things more than others, unique to ourselves. Statistical analysis tests the data without our perceptions mucking up the works.

27 Apr 2014 10:22 AM

One way to think about the unraced-at-2 in-the-money stats is to compare observed versus expected.  We would have expected 10 in-the-money finishes, and there were 8.  Also, in defense of the author, statistical significance *is* partly a function of the sample size.  It is perfectly plausible for a 2% difference to be significant in a larger sample and not in a smaller sample, even if everything else were the same.

27 Apr 2014 3:07 PM

Not directing these comments to the study at hand, but it should be said that even though a study may receive the "statistical seal of approval" does not, it itself, grant its CONCLUSIONS to be valid.  For example, the study's hypothesis may fail to embrace certain pertinent variables.

28 Apr 2014 10:24 AM

There are at least a couple of studies highlighting the fact that running as a 2yo heightens the possibility of remaining sound - I note this as I used to not race at 2 being misinformed!  If that's true, then that factor would militate against those unraced at 2 - especially trying to go from unraced to a mile and a quarter in 5 months with the intense training needed and what is probably less of the foundation that builds soundness. Rather than the 2yos being rushed (save for the 2yo sales maybe) those brought along reasonably, at speed at an ever-increasing distance, would be in position to get the mile and a quarter, and remain sound enough to even give it a try in the Derby.

30 Apr 2014 8:00 AM


I've read those studies. As I recall, they don't really distinguish training at 2, from racing at 2. The better approach (not speaking economically) is likely to have them in training at 2, but to postpone the racing until later.

30 Apr 2014 1:51 PM

The question pursued needs all of its parts. It is "All other things being equal, does x vary as y?"

All other things are sooooo not equal between horses racing at two and horses not racing at two. The first group does not contain ANY two-year-olds who failed to stand training well enough to get to the starting gate. The horses racing at two are doing so, as a group, because they are sounder at two. So what did these studies really prove?

02 May 2014 3:19 AM

Cassandra - From my (limited)_recollection, I think the issue is that those the started at 2 vs. those that didn't. They actually would have been at a disadvantage in part, needing to stay sound thru training at 2 and beyond.  The ones that didn't start training til three were not subject to those early stresses, yet still did not outperform those that ran at 2.  The hard/tough training of old time trainers running a dozen or more times at two may have been beyond the need, but the bone remodeling at early years is generally deemed to be beneficial, to a point.

02 May 2014 8:37 PM

Recent Posts

More Blogs