Six mana double-black at sorcery speed, in a medium-speed format with one to two playable removal commons per color, is the math that decides where this sits. BG Food and WB Legends-attrition take it P1P3 to P1P5: not over the Turtle mythics or Stomped by the Foot, but ahead of most uncommons in those archetypes. UR Sneak and the faster Mardu builds let it wheel.
The setup tax is the gate. You need a creature in the bin worth six mana, and TMT's commons do not naturally fill the yard: Mill shows up on five cards total, and self-discard is thin. The targets that justify the engine are the legendary uncommons and Turtle mythics, which means the payoff is gated behind the same picks the rest of your deck already wants. Stripping legendary off the tokens does real structural work precisely because the best targets are the legends the format keeps pushing you toward.
What the format cannot do is grind the engine out once it lands. Chapter I triggers the moment the Saga enters, so you have a token the turn you cast it, and killing that token with Grounded for Life or Stomped by the Foot accomplishes nothing: chapters II and III copy the exiled card, not the token. The real answer is narrower. Enchantment removal in response to the chapter I trigger is brutal: the trigger still resolves and you keep one token, but the Saga itself is gone, so chapters II and III never arrive and a three-for-one collapses into a one-for-one. Counterspells are cleaner still. That single window, the turn you tap out, is the only place the engine breaks.
Maindeck wherever a graveyard plan supports it.
