Mnemonic Deluge

Nine mana is the whole argument. Where most graveyard recursion buys a single spell back, this one leaves the target in exile and instead stamps out three free copies of it on the spot. The math scales with the target, which is where the card graduates from value tool to payoff. Recurring a modest cantrip wastes the investment; the design wants the graveyard's largest instant or sorcery, the mass draw, the board wipe you would happily resolve three more times if the mana were free. Note the target trap this rewards: because the copies are cast without paying their mana costs, any X in the original spell is set to zero on the copies, so a Fireball you paid full price for the first time resolves as three do-nothings when tripled. The payload has to be a fixed effect worth casting again, not one whose value lived in the mana you sank into X. Exiling itself afterward is the concession that keeps the effect from becoming a repeatable engine: a single detonation, not a loop, so the setup has to earn its keep in one resolution. The sorcery-speed timing and the requirement that the target already sit in a graveyard push the same direction, asking you to sequence a big spell into the bin first and triple it later rather than assemble everything in one window. It sits at the maximalist end of a long line of copy-and-recast effects, sensible only where nine mana is reachable and the graveyard is stocked with things worth casting again.