Well, considering your argument, you would turn your uncommons more unreliable because you add more uncommons to the pool. Since the rarity of the cards is calculated before the card pool itself is touched you have always the same probability for the raritys. adding more uncommons to that pool would reduce the probability for drawing the card that really fits into your strategy, turning your uncommon pool into an inconsistent pool either.
And by the way, if the common pool is build right, it features around the whole game the most consistent draws. That means that you really need to find out which commons you need or how to get rid of commons you don't need.
(For example, wish is a great common to have in many decks which don't rely on hand construction, like dragon or titan decks so, because it can refresh your hand to get the commons and uncommons for the tactic you need... On the other hand, having many keyword cards in a keyword based strategy is good, because then your uncommon and rare pool don't need to waste too many slots for keyword cards to fuel the keyword. A dragon/alliance hybrid deck of mine for example makes great use of it)
So, commons are the backbone of every deck, else the games would only turn into goodstuff-battles...
And, to add tot he discussion, i could bring up this artical (It is a MTG article, but the same can be said for arcomage, although you need to look at each rarity seperatly then): http://www.channelfireball.com/articles/frank-analysis-is-playing-more-than-60-cards-always-a-bad-idea/