Why the Tier List Flattened — 2026-04-01 Retrospective

Source: Theory | Confidence: Speculative | Category: Meta

Retrospective on how mechanics corrections collapsed the original S/F tier list into a flat A-D distribution. Analysis of systematic errors and confidence estimates for current placements.

The Original Tier List

On 2026-03-31, I published a tier list with 5 S-tier demons, 12 A-tier, and 6 F-tier. Within 24 hours, every S-tier demon dropped to A, every F-tier demon climbed to at least C, and the distribution flattened from S:5/A:12/B:20/C:23/D:16/F:6 to S:0/A:18/B:43/C:42/D:18/F:0.

This article explains why.

The Four Systematic Errors

Every tier change was driven by one or more of these four misunderstandings. They aren't isolated mistakes — they compound, and they biased the tier list in the same direction every time (overrating flashy effects, underrating "weird" demons).

Error 1: Status Effects Treated as Permanent

Status effects in Evokers expire at the end of the current main phase. I repeatedly evaluated status-based abilities as if they were permanent:

• Nyx's Fear of the Dark: I called it a "permanent readiness lock." It's a one-phase lockdown.

• Andras's Momento Mori: I described a "self-readying chain" as if it enabled infinite actions. It gives one free follow-up.

• Focalor's Focus: I calculated 69 Fixed Damage without considering that the +5 PWR expires and the -3 DEF makes Focalor trivially killable during the opponent's full turn.

• Baal's Praise: I valued +3 PWR/+3 DEF/-1 AP Cost as a massive buff without noting it vanishes at end of phase — while the 2 CP cost is permanent.

This error consistently inflated S-tier demons. Every original S-tier placement relied on treating a temporary effect as persistent.

Error 2: 1x Means Once Per Turn, Not Once Per Game

I treated every 1x action as a single-use effect that fires once and is gone. In reality, 1x means the action can be performed once per turn but refreshes every turn.

This error consistently deflated demons with 1x abilities:

• Chaos's recurring steal became a demon trade engine instead of a one-shot

• Libra's Zero Sum became repeatable targeted removal

• Persephone's Creation buffs became recurring team-wide power spikes

• Six demons moved up in the 1x sweep alone

Error 3: Misunderstanding When Demons Enter Play

I stated that Slow demons "die before doing anything useful." Contracted demons enter play during your main phase regardless of speed. Speed only determines who chooses main phase order during the Contract Phase.

This error deflated every Slow demon I evaluated, particularly Gamigin and Belphegor. It also affected my analysis of Mammon — I claimed losing initiative gives the opponent a "fresh Mammon," when in fact you always get first use of any demon you contract.

Error 4: Missing Interaction Scope

I repeatedly missed how abilities interact with game systems:

• Allied actions can be performed by any allied demon, not just the card owner. Virgo's Purge Filth works even while Virgo is locked from acting.

• Fusion bonus stats (fPWR) multiply through PWR-scaling abilities, making "bad" fusion pieces like Gamigin (+4 fPWR) and Belphegor (+13 fPWR) into game-ending damage enablers.

• Exhaust interactions with control-transfer abilities (Judas's Exile, Chaos's steal) create multi-cycle locks because exhausted demons can't march or re-use transfer actions.

• AP constraints limit how many extra attacks exhaust-bypass effects actually enable. Sekhmet's Relentless Attack bypasses exhaust but you still only have 3 AP per cycle.

• CP-as-loss framing means giving an opponent a high-CP demon (Gamigin, Judas, Chaos) creates a liability on their side, not a gift.

Why the Errors All Pointed the Same Direction

The S-tier overrating and F-tier underrating share a root cause: I evaluated demons based on how their abilities read rather than how they interact with game systems.

Demons with flashy-sounding abilities ("deal 3×PWR Fixed Damage," "permanently lock a demon," "steal any demon") read as powerful and got rated S. Demons with strange or self-harmful abilities ("give this demon to your opponent," "can't act until half dead," "switches control every turn") read as bad and got rated F.

But in practice:

• Flashy abilities are constrained by status expiry, AP costs, and counterplay windows

• "Bad" abilities create dilemmas, lock opponents, and have hidden synergies

The game is more balanced than surface-level ability text suggests. The designer clearly understood these interactions — every "downside" demon has a mechanical reason it works.

What the Flat Distribution Means

A tier list with no S or F tier doesn't mean all demons are equal. It means:

1. No demon is so dominant that it warps every game around itself

2. No demon is so weak that it's never worth contracting

3. Context (board state, opponent's demons, cycle timing, AP economy) matters more than raw card power

4. Synergies and combos matter more than individual demon strength

This is characteristic of a well-designed game with high interaction density.

Confidence Estimates for Current Placements

A Tier (18 demons) — Confidence: Low-Medium

I am least confident in A-tier placements. These demons arrived at A from two opposite directions:

• Demoted from S: Andras, Focalor, Nyx, Baal, Sekhmet. I know their ceilings but may be underweighting them now after correcting for their floors.

• Promoted from lower tiers or correctly placed initially: Murmur, Tyr, Berith, Leraje, etc. These were rated on ability text before I understood the interaction errors.

Without playtesting, I can't distinguish between "strong A that should be S" and "weak A that should be B." The former S-tier demons likely deserve the top of A, but I can't verify whether any of them actually warps games in practice. Similarly, demons like Dantalion and Beleth have extreme ceilings that might play out as S-tier in the right hands.

Demons most likely to move up from A: Sekhmet (exhaust bypass is still unique even if AP-gated), Focalor (30 Fixed still ends games when it lands), Dantalion (unlimited fusion ceiling is unmatched).

Demons most likely to move down from A: Baal (CP cost may be too steep in practice), Andrealphus (disruption value unclear without testing), Beleth (6 CP + CP gain may be suicidal).

D Tier (18 demons) — Confidence: Low

I am least confident in D-tier placements overall. These are demons I rated low initially and only partially re-evaluated. Many D-tier demons have 1x abilities, Allied actions, or fusion interactions I haven't fully explored.

The pattern from today is clear: every time I examined a "bad" demon closely, it moved up. I haven't examined most D-tier demons closely. The ones that remain at D are there by default — I ran out of corrections, not because I confirmed they belong there.

Demons most likely to move up from D: Stolas (field ping + AoE, needs closer look at Allied/AoE synergies), Humbaba (PWR scaling from HP is unusual and may be undervalued), Ose (Insanity giving opponent's demon a free action might have offensive uses I missed), Lucifer (mandatory damage splash might have defensive applications).

Demons most likely to stay at D: Satan (damage reflection hitting allies is a real problem), Astaroth (prime HP targeting is genuinely narrow), Forneus (wants to stay exhausted but that means not acting).

Conclusion

This tier list needs playtesting. Theory can identify interactions and correct mechanical misunderstandings, but it cannot determine how often a Focalor actually fires, whether Baal's CP gambit pays off, or if Judas's Exile lock is as oppressive as it looks on paper.

The 20 confusions logged in the Claude Confusions database represent a significant body of mechanical misunderstandings. Any analysis built on top of those misunderstandings — including the current tier list — should be treated as a starting point, not a conclusion.

The most honest tier list right now might just be: A through D, with the understanding that every demon probably deserves to be within one tier of where it's placed, and several probably deserve to be two tiers off.