# Revision two: review decisions

**Status: reviewed text proposal.** Four fresh deck reviewers and a fresh whole-set reviewer independently read the raw proposed cards, full lists, fixed commanders, and Oracle/rules evidence. Their reviews preserve initial findings and rechecks: [Zeus](zeus-cold.md), [Poseidon](poseidon-cold.md), [Hades](hades-cold.md), [Athena](athena-cold.md), and [whole set](set-cold.md).

The later **v2.1 tuning** reduces Metis to one draw, costs Lots at `{W}{U}{R}`, replaces Nisyros's Overrun with an 8/8 land-copy token and optional bounce, and reduces Riches to one card and one tapped Treasure per creature exiled. Its separate [focused cold review](tuning-cold.md) covers these four changes; the initial reports are historical, not assertions about superseded text. Complete deck membership is unchanged by this tuning.

## Corrections made during review

| Finding | Resolution |
| --- | --- |
| H03 combined a fixed group gaining hexproof/indestructible with a game-rule restriction that could reach later creatures. | Grant all three abilities to the same group: hexproof, indestructible, and “This creature can't block.” The Hades reviewer independently rechecked the printed repair under CR 611.2c. |
| Rhea's design note incorrectly suggested changing a one-target spell into a zero-target copy. | Corrected the YAML and play guidance. Copies retain target count: change to a second legal creature, or accept that phasing the shared target makes the original fail. Zero-target originals produce zero-target copies. Zeus reviewer independently rechecked; printed card text needed no change. |
| Poseidon's Oracle cache retained historical records but used an ambiguous count label. | Replaced it with `active_existing_card_count: 89` and an explicit historical/current provenance-cache description. The active list was already correct. |

No reviewer identified a remaining blocking rules defect, prohibited infinite, deterministic combo table kill, or repeatable lock in the proposed lists. This is not a proof that every possible interaction has been exhausted. The reviews explicitly preserve strong finite swings rather than reducing them merely for being more powerful than normal staples.

## Feedback coverage and accepted decisions

- **Zeus:** 34 instants, zero sorceries, two counters, six native flash creatures, and three broad flash enablers. Eight custom faces revised and fourteen existing slots replaced. Metis is an ongoing one-card draw engine; Nectar and Titanomachy are the signatures. Lots now costs three colored mana. Actual mana budgets, flash setup costs, and cast-versus-copy triggers are documented. Token armies are earned by the instant game and eventually attack.
- **Poseidon:** Six revised faces and six existing swaps. Nisyros creates an 8/8 trampling copy of a friendly land plus optional nonland bounce, providing development and a threat without an existing army. Prayer is unconditional modal interaction plus a cantrip. Veil draws two, Hippocamps rewards two land entries on the pilot's turn, and Triton costs a straightforward three mana. Ketos gains entry impact. Only Raft remains as a mass creature-displacement spell.
- **Hades:** Five revised faces and ten existing swaps. Helm attaches then phases; Look Back is targetless and explicitly forward-only; Judgment is four-mana destroy-only. Riches and Cerberus immediately mill and offer meaningful output. Fifteen overlapping fuel sources support a 29-creature list, while exiling fuel competes with resurrection. Empty-graveyard mills and tapped Treasure timing are treated honestly.
- **Athena:** Sthenelus counts its own death. Three existing swaps replace two self-defeating sweepers and one delayed attacker. The self-return/new-object interaction is intentional, paid for, and unsupported by a free creature-sacrifice loop in this set.

There are eight designated signatures, two per deck. Metis is no longer designated a signature after the requested reduction. Riches's payout is now at most three cards and three tapped Treasures per trigger; its fuel and immediate mill remain. These user-requested reductions are not offset by unrelated buffs. No signature promises a free or deterministic table kill.

## Remaining playtest priorities

1. Does Athena overwhelm the table now that mass removal/displacement falls from twelve cards to four, or do selective answers and combat finishes keep pace?
2. Can Zeus sustainably exploit all three opposing turns after paying for its six-mana god and flash engines? Track actual mana, retained interaction, resource bursts, and attacks rather than assuming all triggers are free.
3. How often does Riches consume one, two, or three creatures? Record its missed mills, depleted library, reserved resurrection targets, and the delay before upkeep-created tapped Treasures become available.
4. Do Poseidon's extra land permissions have adequate card supply? Track the utility of Nisyros's copied land, its creature-removal exposure, and the delay before it can attack or tap; count Provisioner Treasures after Sea as available mana.
5. Do retained setup-dependent threats justify their slots? Trench Behemoth and Athena's attack engines remain deliberate playtest candidates rather than automatically cutting every delayed payoff.

No actual multiplayer games were run. Existing images remain v1, and the artifact-frame correction is deferred to a future image task. The [text proposal](../proposal.md), [forty-card manifest](../cards.yaml), four complete lists, and [validation report](../validation.json) are the deliverables for this revision.