Gary, what an interesting topic... :D
Let me state right away that I DO believe that there are subtle indicators in a puppy's behavioral patterns and that these intrinsic patterns will project within a predetermined range into adulthood. The question as far as I am concerned is, are these indicators identifiable and are they reliable enough to allow for objective testing?
I have to say that I was wondering about the predictability of adult temperament for a long time. Several years ago, I started asking (human) parents, whether the personalities of their offspring were already "visible" at a time when their kids were still infants (less than 1 year old). Most of them told me that in retrospect, they could swear that certain features had always been there. However, they subsequently referred to very different aspects of their childrens personality; usually miniscule "signs" really that one would have never picked up right from the beginning. The key here is really that it was always in retrospect.
When any of you look at your adult dogs, you will probably quickly identify certain features about their grown dispositions, that have always been there, right from early puppy stage. I remember some of my dogs as being clumsy, smart, reserved, or very attached, just as they have been as puppies. In retrospect, it is easy to identify the traits that have always been there. In prospect however, it is a different story. There are certain traits that dissapear during development, some emerge at a later stage, others are simply modified by epigenetic factors. In science, this is called noisy information. I believe it would be extremely difficult to accurately identify those traits that would steadily persist into adulthood. Rather than dissecting a personality into discrete properties, it might be adventageous to evaluate a complex pattern such as "temperament" as an entity.
A clever "trick" would therefore be to emulate retrospective evaluation. This in fact is performed many times intuitively by breeders of dogs. They compare complex behavior patterns with previous dogs from their own breeding lines. Oftentimes, these assessments aren't really that tangible. People then refer to a puppy as having "it". Perhaps there is something strikingly captivating about a specific puppy that simply reminds the educated "eye" of some charismatic ancestor. This is what I consider the most subjective, but on occasion also the most reliable way to identify a good puppy.
Problems can arise, when a breeder is not experienced enough or simply doesn't have a good "feel" for his/her dogs. Moreover, it forces a potential buyer to heavily rely on the judgement of the seller, which understandably can make people feel uncomfortable at times. (It certainly would make me feel uncomfortable). An objective scientific test would empower the buyer to evaluate a puppy and furthermore render the necessity of a good "feel" for this particular line obsolete. At least in theory.
Well, in my personal opinion those objective attempts at testing a puppy's destiny only go so far. Simply said, these tests represent a dramatic abstraction of a highly complex matter such as a dog's adult personality. I still believe that patient observation of the puppies in a group and individually, as well as perceiving a certain "vibe" can prove superior to objective tests. However, it unfortunately requires talent - and experience.
This is what I believe old-style shepherds use (in addition), when they test puppies for their feasibility as lifestock guardians. The performed tests are often so "soft" in design that their outcome may even become a Self-Fulfilling Prophecy, yielding those puppies as superior, which the shepherd had already subconsciously assumed anyway. Don't get me wrong, lenient testing may not be such a bad idea afterall.
Stringent "scientific" methods may overly abstract the underlying problem and consequently result in unreliable outcome. Precise scoring doesn't leave too much room for "late bloomers", "mimicry" or simply that "gutt feeling".
While I personally have always been very diligent when it comes to picking a puppy, I do not employ any specific testing regime nor do I religiously follow any series of steps in the attempt to determine the one that has "it". At the end of the day, I tend to rely on my "gutt feeling". But that's just me.
As a Sarplaninac enthusiast though I have to add, it is ALL in the eyes..! 8)