I believe that it does pertain to the subject. It has already been discussed that some Rares are useful for almost any deck, whereas many don't work well with other planning at all. If the option is given to take only those few that would certainly be useful, then you are inflating the chance of getting that particular card, giving it a greater overall effect as it would influence more games more often. If I had the option, and I hope that I never do, to take only OotWL and PoT as my only Rares, and each then had a 3% chance of appearing in a vacant slot in hand, I'd certainly go for it, and the game would be less fun.
My opinion is that if a player is granted the option of taking fewer Rare cards, the chance to draw a Rare should be adjusted to grant the same overall chance that it had before. If the current chance to draw a specific Rare is (6/100) * (1/15) = (1/250), then a player with only 9 Rares in his deck should have a (1/250)/(1/9) = (9/250) = 3.6% chance, adjusted again for the 2.4% loss overall. (100/97.6) * 3.6 = 3.69% chance to draw a Rare if there are only nine Rare cards in a deck. (The percentages for drawing a Common or Uncommon would also have to be modified for the 100/97.6 difference, and any other possible total number of Rares would yield again different numbers.)
This ignores that cards that have the effect of drawing a Rare in your deck as well as one not in your deck would have their odds of relevancy altered as well.