Yearling Market Study

Talk about upcoming sales or auction results.

Moderators: Roguelet, WaveMaster, madelyn

Shammy Davis
Chef de Race: Classic
Posts: 4451
Joined: Sat Oct 02, 2004 8:23 am

Yearling Market Study

Postby Shammy Davis » Sat Dec 12, 2009 6:54 pm


User avatar
Joltman
Grade I Winner
Posts: 1743
Joined: Fri Jan 06, 2006 5:33 pm

Postby Joltman » Sat Dec 12, 2009 8:40 pm

Lots to think about. In particular the discount based on what the 'seller' is perceived to be doing, ie dumping his stock that is not race worthy. An inefficiency might arise in a situation like now, where the seller who is a 'Racer' HAS to unload because of finances. The buyer can get a better deal on a better horse in that situation. The emphasis on the sire is interesting too.

lots there

jm
Run the race - the one that's really worth winning.

LB
Eclipse Champion
Posts: 2388
Joined: Sat Jun 23, 2007 6:57 am
Location: Kentucky

Postby LB » Sun Dec 13, 2009 8:14 am

That article was clearly written by someone who is more familiar with a theoretical TB market than with the true TB market as it actually exists.

Shammy Davis
Chef de Race: Classic
Posts: 4451
Joined: Sat Oct 02, 2004 8:23 am

Postby Shammy Davis » Sun Dec 13, 2009 10:42 am

LB posted:
That article was clearly written by someone who is more familiar with a theoretical TB market than with the true TB market as it actually exists


I think you are right but can you elaborate? Lot's of what is written about TB sales is clearly written by spectators rather than true connections.

I remember a link that I found some years back that was published by a university in FL. The author was anonymous. It was a good read but I wondered about its worth in the real world.

JM posted:
. . . Lots to think about. In particular the discount based on what the 'seller' is perceived to be doing, ie dumping his stock that is not race worthy. An inefficiency might arise in a situation like now, where the seller who is a 'Racer' HAS to unload because of finances. The buyer can get a better deal on a better horse in that situation. The emphasis on the sire is interesting too.


It's almost like you need biographical data and Dunn & Bradstreet report on the sellers to figure out who and what you are dealing with. I'm beginning to think that TB sales market is a "little Wall Street."

JM, are you familiar w/Sam Rodgers? I don't know him personally, but I was very impressed w/his operation. I looked over a couple of his yearlings (FL breds) quite a number of years back at a F-T Mid-Atlantic sale. White Hall, I believe was the name of his farm. Both yearlings went for a hefty price the next day. Way out of my league.
Last edited by Shammy Davis on Sun Dec 13, 2009 5:45 pm, edited 1 time in total.

LB
Eclipse Champion
Posts: 2388
Joined: Sat Jun 23, 2007 6:57 am
Location: Kentucky

Postby LB » Sun Dec 13, 2009 11:31 am

To be honest, I only read the first couple of pages because I'd already found many eyebrow-raising "facts" and didn't care to continue. I tried to copy and paste a quote to illustrate but for some reason I'm unable to, so I'll just retype one sentence from the top of the second page:

Chezum and Wimmer hypothesize that the greater degree to which a seller races thoroughbreds, the larger price discount his yearlings receive.


This conclusion was determined by by an earlier conclusion that the Keeneland market for TB yearlings is exactly like the market for used cars and that all the same selling principles apply. :shock:

Anyone who can read and believe the above hypothesis is either someone who has never been to a TB sale or who never reached an understanding of anything that was happening while they were there.

Dave C
Restricted Stakes Winner
Posts: 812
Joined: Thu Sep 16, 2004 11:25 pm

Postby Dave C » Sun Dec 13, 2009 4:59 pm

Having read the article I would conclude that the author is very familiar with the TB industry. The statements that LB is taking exception to are the product of forcing the complex TB market into the mathematical constraints of economic theory. The paper was written to be read by people who are not familiar with TB's but are familiar with economic theory and the structure of the TB industry is communicated as being similar in functionality to the used car market: this is to ensure the reader draws upon the appropriate theoretical framework for comprehending the work in front of them, it is not intended for breeders to look at their horses like used cars.

The definitive conclusions drawn by the author are inappropriate given the low R-value achieved from their model. However, I also understand that it is necessary within the framework of academia to make definitive statements even when you think you might be wrong.

User avatar
Joltman
Grade I Winner
Posts: 1743
Joined: Fri Jan 06, 2006 5:33 pm

Postby Joltman » Sun Dec 13, 2009 8:09 pm

A statistical analysis of something as fluid as human behavior is going to, at best, give very general and broad conclusions with weak correlations.

Nevertheless, it would be interesting to see how the statistical analysis performed in another market year to see if it held up. These kinds of articles, if nothing else, prompts a 'common sense' analysis of the situation that might be helpful and certainly is going on at the sales. For example, if there are Breeders and Racers then Dispersers would also seem to have certain price characteristics. One would THINK that you would get a lower price at a dispersal but that that can certainly be a come-on as well with a reserve buyer in place.

--- SD can't say I had known much of Rogers, although he had some success with some stakes horses I believe. Any reason for your special interest there?

jm
Run the race - the one that's really worth winning.

Dave C
Restricted Stakes Winner
Posts: 812
Joined: Thu Sep 16, 2004 11:25 pm

Postby Dave C » Mon Dec 14, 2009 9:05 am

The human behavior is the easy part to model because people are not buying the TB yearlings for abstract reasons: they are trying to buy a champion. That means they will act relatively rationally which economic theory is very good at modeling.

The difficult part is modelling the actual yearlings. The catalogue page itself is difficult to reformat into a statistical model (I tried about 10 years ago as part of a marketing course). But so much of how a yearling sells is presentation. If a yearling is an absolutely stunning individual it well sell well above its' catalogue page and if it is a wreck it will sell well below. Where is the record of how a yearling looked at the sale? How do you quantify it?

When I was doing econometrics the rule of thumb was that unless your model had an R-value of at least .8 it was of no value. The study in question has an R-value of .4 which means that the conclusion that a racer can expect a 35% discount (with a margin of error of 60%) is not reliable as a predictive indicator.

User avatar
Joltman
Grade I Winner
Posts: 1743
Joined: Fri Jan 06, 2006 5:33 pm

Postby Joltman » Mon Dec 14, 2009 9:46 am

I don't think the behavior thing is so easy to predict because a certain component, in many purchases, is non-rational. So, the Green Monkey draws 16 mil in the ring as people bid on a 'future champion' based on emotions.

That scenario is probably played out over and over again as the desire (human passion) for a champion clouds/biases the process of decisionmaking and the consistency of a scientific model is compromised.

jm
Run the race - the one that's really worth winning.

Dave C
Restricted Stakes Winner
Posts: 812
Joined: Thu Sep 16, 2004 11:25 pm

Postby Dave C » Mon Dec 14, 2009 4:58 pm

TGM going to 16M is perfectly understandable within the English auction model which is why it is so popular for selling things. Both the English and Dutch auction models have been extensively studied and statistical models developed which account for the human components.

The rationality comes into play because people are buying for a specific purpose. People do not buy yearling TB's because they're 'cute'.

If you were to talk to the economists working for your credit card provider they have an extremely complex model which will basically tell them everything about your shopping habits, including things you hadn't even realized and they are working on finding ways to sell this information to marketers so that you are inundated with 'special offers' exactly at the time they think you are going to buy that type of product. Humans like to pretend they are unique and make their own decisions. Reality is that we are just biological computers and once you figure out the programming are pretty easy to understand and manipulate.

Shammy Davis
Chef de Race: Classic
Posts: 4451
Joined: Sat Oct 02, 2004 8:23 am

Postby Shammy Davis » Wed Dec 16, 2009 6:37 pm

JM posted:
--- SD can't say I had known much of Rogers, although he had some success with some stakes horses I believe. Any reason for your special interest there?


Not really. The couple of experiences that I had looking over his horses made a positive impression. Maybe, I was surprised that his operation was based in VA.

tbrace
Restricted Stakes Winner
Posts: 960
Joined: Fri Nov 05, 2004 7:52 pm

s

Postby tbrace » Fri Dec 25, 2009 10:45 am

The r value of .4 means that independent variable, in this case whether a seller races or not, has less than a 50-50 chance of affecting the model.

In other words, it is like flipping a coin.

The benefit of the study, though, is to begin to quantify auction outcomes.

On the whole, the premise of the study is accurate. It is very true that breeders who race say that they sell best and keep the rest, when in fact, breeders who are committed racers, like Frank Stronach, Ken Ramsey, etc, always keep their best prospects.

The catch comes in that even the best experts cannot tell which yearlings are really going to run and which ones won't. Look at the list of highest selling yearlings of all time vs. their earnings. Most are failures at the track.

Dave C
Restricted Stakes Winner
Posts: 812
Joined: Thu Sep 16, 2004 11:25 pm

Postby Dave C » Sat Dec 26, 2009 5:24 pm

I'm not sure that the premise is correct. As is mentioned in the paper, there is an assumption of perfect knowledge. That is that all buyers know who is a racer and who is a breeder. At a sale like Keeneland, there are so many sellers and buyers, it is more reasonable to assume that for every horse in the ring, there are at least 2 buyers who do not know if the seller is a racer or strictly a breeder. Since it only takes 2 bidders to drive the price up, even if everyone else knows that the seller is a racer, it won't be reflected in the price. That the study shows a strong correlation may be just that variability is being attributed to the dependent variable that is actually due to some other factor.

tbrace
Restricted Stakes Winner
Posts: 960
Joined: Fri Nov 05, 2004 7:52 pm

e

Postby tbrace » Sun Dec 27, 2009 11:22 pm

Dave,
You make a good point. The dependent variable, - in the study being price of yearlings, may be due to a different independent variable, one other than knowledge of racing or selling by the buyers. As you know this is a common mistake in correlational studies.

It would be interesting to do a multi-variate study (although very difficult and costly) and then run multiple regressions to see what kinds of variances in price could be accounted for in a sample of buyers and sellers, over a number of sales.

In the mean time, let's buy the best looking one!