A big part of the optimal QI is the number of turns you expect to play.
In these tests the opponent is defenseless, against a real opponent you should expect shields/healing/CC most of the time, thus the game will last longer.
When a game lasts longer, the pillars will produce more quanta.
The deck formed of pillars and golems kills a defenseless 100hp opponent in 7/8/9 turns. On average, each played pillar will only produce a total of about 5 quanta. In this case 18 pillars and 12 golems is the optimal.
However, if we set the opponent's hp to 200 to make it more close to an opponent with defenses, the games will last 11/12/13 turns, the pillars will produce more and so the optimal QI is now higher. In this case something more like 14 pillars and 16 golems would be better.
In addition, the Abomination Deck has a lower TTW than the Emerald Dragon Deck, but, theoretically, middle-ranged attackers are better in rush decks.
The reason middle hitters are generally better is due to single-target CC wich is quite common.
If this game were only about pillars and creatures without abilities, the dragons would be the top choice.
Of course the decks are hypothetical, you cannot put 20 frogs in a deck. :/
