Longer games suck mostly when you lose.
The goal here is to make the game fun and worthwhile regardless of whether it was fast or slow.
There are many games out there where after a while you see players giving up after a few turns because they anticipate that they will either lose or it'll be a really long game.
I don't think Elo should be touched by this, for various reasons:
If you increase loss for longer games, people quit at the start as soon as they anticipate the game isn't going well.
If you reduce loss for longer games, losing players will just stretch things as much as possible.
Perhaps the problem is measuring it in time - as that will make players stretch turns to the max.
I think Turns or even Moves would be better.
Or something else entirely, like damage dealt, healing, faeria spent, etc.
Actually, I have an update on my suggestion for game mechanics:
1 - Allow us to pick 1 card when Building our Deck, to always start in our hand.
The aforementioned restrictions would still apply
2 - You draw 9 cards at the beginning of the game - which includes the 1 you can choose to always appear in your hand - then mulligan 6 to 9 of them, as you can only keep 3.
3 - Less important, but still worth giving it some thought, everytime you draw at the start of a turn, you'd draw 3 cards, keep 1, and place the other 2 back in the deck.
I don't really think it goes exactly as you said, though.
Every deck knows what they want to start their hand with, if they could choose.
The thing is, rush decks are all about overwhelming, while other decks have many different pieces to the puzzle.
Imagine getting buffs but not the creatures you want to buff.
Imagine having your "end game" cards without the setup.
I believe the answer is actually making all types of decks more reliable, as it allows you to truly balance rush decks.
Sure, proper deck building is all about making a deck that tries to beat the odds.
However, Luck is fickle, and:
1 - If a "perfect opening" becomes popular, a "perfect counter" will arise, so ultimately there will never be a perfect opening.
2 - The only matches that suck to play are the ones where you couldn't even play your strategy, because you simply got everything in the worst possible order.
3 - Between Faeria, Duelyst and Hearthstone, I consider Faeria to have a considerably higher potential for competitive play.
This game takes many steps in the right direction - pooling of Faeria across turns, lower RNG in cards, etc - but the more solid its various mechanics are, the worse RNG feels when you get on the wrong side of it.
Games shouldn't be ruled too much by what cards you got in the first 3 turns.
Plus, I think having starting hands range between "good and perfect" is a lot better than between "bad and perfect".