Originally Posted by
square_two
What I meant was that, perhaps there is a new legendary card that costs 2RR (think like, Pia and Kiran). To test whether it is as good as I think it is, I could see myself testing with a full playset just to see what happens when it hits the board. I don't think I'd ever run 4 of a legendary card at 4cmc. If it was a different deck and the card was Thalia (the old one) then that is a different story since it acts as an early lock piece instead of being just a threat, and it comes down early enough that you really want to see her in your opening hand most of the time.
I think forcing full playsets isn't going to be optimal in a lot of situations. Yes, having only playsets is technically the best way to see each individual card in a deck without filtering, but what if the best configuration for this deck is 9 or 10 3-drop threats? The true optimal configuration for a meta (which is impossible to evaluate, at least until we have super duper AI) might lean towards 5-9 of a type of card, or perhaps 2-3. At least that's what I think.