Watching Day stream Hearthstone (and in particular how he plays his Subtlepaw deck) leaves me with the impression that he both doesn't have, and doesn't consider that perhaps he should have, any metric for measuring how good a play is. While I think some other Hearthstone streamers exaggerate the importance of "value," designing and piloting a deck with success requires beating your opponent in some aspect of the game.
As I see it, there are 4 major resources your opponent has that you can try to negate, and therefore 4 major paths to victory:
1. Life total. A deck that only cares about having a way to bring the opponent to zero life ASAP is usually a combo deck in Hearthstone. All the other decks can resort to this when in desperation, but even aggro decks generally can't just ignore board position.
2. Mana. This is the usual way for an aggro deck to win. If you can make better use of your mana each turn than your opponent can (e.g. by negating a 3 mana card with a method that only spends 2 mana), and you manage to spend all your mana each turn, then you will have threats on the board that he hasn't dealt with yet, and these will eventually win the game for you.
3. Cards in hand. If you can force your opponent to play all his cards and enter topdeck mode, and you still have cards left in your hand, then you can generally both answer and play threats, while the opponent can only do one or the other.
4. Cards in deck. The opponent only has so many threats total in his deck; once you've negated every last one of them, using the remaining cards in your deck to secure a victory should be a breeze.
While it may seem like 3. and 4. are similar (they'd both get a "control" label in M:tG), they are completely different. Day's Subtlepaw falls flat precisely because of an inability to distinguish between the two goals.
Consider Naturalize. A deck trying to starve the opponent's hand would generally chalk up resorting to this card as a critical failure - the card is, by design, giving your opponent a "one for negative-one", for a net loss of two cards. But to a stalling deck, this is perfectly fine - you killed that card dead, and you had planned to eventually kill those other cards you just handed him anyway.
Compare this to card draw (say, the card draw form of Nourish that Day did a turn 2 Coin + Innervate to rush out). Provided the card efficiency deck can afford the 5 mana's worth of time to cast the card, it's everything they want - more cards, out of nowhere! However, to a deck whose game plan is to have more total threats and answers in the 30 cards alloted, every card that doesn't directly affect the board is at a huge premium. You're going to eventually draw all the cards in your deck anyway.
So, Subtlepaw isn't going to win in the early game because it doesn't care about mana efficiency, it's not going to win in the mid game because it doesn't play enough threats often enough / answer threats efficiently enough to outpace the opponent's cards in hand, and it won't win the late game because too much of its deck space is spent on cards that either don't do anything at all on their own (like Innervate and Nourish) or pay some power to draw cards (like Novice Inventor).
Winning a game of Hearthstone will come off the back of maximizing one of your resources compared to the opponent. Even if you have to change which resource you're measuring your success by depending on the matchup (in aggro vs aggro, one of the decks has to resort to card efficiency) or even the board state (say, switching to yolo style life total smashing if reestablishing board control appears impossible), it is important to have a sense of how you're progressing to victory, and be careful not to throw away a possible win in one category while scoring points in another.