As an example, say we have X objects. Of X objects, there are Y properties. There are also N folders. We have two classes of subjects participating in the experiment. Subject class A is tasked with memorizing the Y properties of X objects. Subject class B is tasked with memorizing which folder of N the Y properties of X object is located.

If for example, X = 100 and Y = 5 then N = 100. An example result for A subjects might be an average of 5 objects where each of the 5 properties were memorized correctly. [That's 5 properties x 5 objects = 25 memory units]. The question I have, is if B subjects would remember an average of 25 objects stored in which folders correctly, or might they remember less/more correctly? If the Y properties vary in type (from color, to weight, to appearance, etc) do they require more memory power than a consistent X is in N folder memorization?

If it turns out you can memorize more about where information can be found, than you can about the actual information then it would seem that [for so long as the information can be looked up] it is more practical to catalog that information exists and where it can be found, than to engage in memorizing the information itself.

Then the idea of "being a jack of trades, master of none" has a slightly different caveat. In that one could be more-so a jack of all trades than one could master a single trade. Ie, the sum of knowledge held by the jack is greater than that held by the master.