Marc -

For me, I define "talent" as an inborn, hardwired (aka genetic) propensity to find performing some action or operation more easily than other people.
I use the term "skill" to refer, regardless of any talent, to the ability of an individual to perform the actions and operations associated with that skill.
Now, these definitions have some implications that are both a) pretty obvious to anyone with two neurons to rub together, and b) anathema to the PC crowd, since they assume an innate, genetic difference between people.
I think this is an important observation and others have made this with respect to command. Interestingly there is a move afoot to see how some of this can be modeled. Could you for example take the 5 block model and make some qualifications about what education, experience and training constitute them? Could you then quantify this in a mathematical expression? Could you then use that to compare probability of success of one individual with a given combination over another dependent upon conditions and objectives?

It seems that until you can put these capabilities into a model and simulate it in a way which provides justification for making programmatic decisions, relevant anecdotal information by itself will not move the ball forward (much). *Note - this says allot about our bureaucratic culture that we are far more comfortable with a false reality of mean time between failure (MTBF) and probability of kill (PK) ratios (which are themselves based on observation and historical averages) then we are with trusting our senses and intuition (based on our own experiences) simply because we've found a way to mathematically express one and not the other. The equation of how you know a tank BN of X will destroy a tank BDE of Y does not match perfectly to the anecdote of sometimes you are the bug and sometimes you are the windshield.

Again, I'd refer anyone interested in the issue of how our perception of reality influences our judgment to Marc T's excellent website.

Best, Rob