A useful chess-variant catalog should do more than collect names. It should let a reader answer four questions quickly:
- What changed in the rules?
- What new experience does that change create?
- What strategic consequences follow?
- Which source or implementation defines the game being described?
A catalog built around those questions can grow to hundreds of games without becoming an arbitrary list.
Use Depth Tiers, Not Quality Rankings
Not every variant needs the same editorial depth. A practical catalog can use tiers such as:
- Core — major variants that deserve complete rules, strategy, notation/data and implementation coverage;
- Significant — established games that merit a substantial rules and strategy profile;
- Reference — narrower, historical or experimental games documented accurately but more compactly.
Depth should reflect reader value, source availability and ecosystem significance—not whether the game is “better.” A culturally important regional game or historically important predecessor should never be demoted merely because it is less common on a particular platform.
Start With the Rule Delta
A rule delta is the smallest change that explains why the game behaves differently from a chosen reference ruleset.
Examples:
- Chess960: randomized legal back rank + generalized castling;
- Crazyhouse: captured pieces enter the capturer's pocket and may return as drops;
- Three-Check: the third delivered check is an additional win condition;
- Antichess profile: capture compulsory + non-orthodox king/objective rules;
- Atomic: captures transform nearby material through an explosion rule.
The delta gives you a research hypothesis. It does not replace the full ruleset.
Derive Strategy From Mechanics
For every rule change, ask what familiar resource changed value.
Objective Change
If the victory condition changes, evaluation changes first. In Antichess, material can become a liability. In King of the Hill, central king mobility becomes a direct race resource. In Three-Check, a legal check can have persistent score-like value even when it wins no material.
Capture Change
Mandatory captures, explosions and capture-to-hand mechanics reshape tactics. A geometrically defended piece can be tactically unsafe in Atomic because the relevant object is the blast zone. In Crazyhouse, a sacrifice may create both an immediate attack and a future pocket resource.
Geometry Change
Larger boards, wrap-around files or hexagonal lines change distance, center definitions, piece reach and pawn routes. Orthodox square-count heuristics should be retested rather than translated literally.
Information Change
Hidden information creates beliefs and information-gathering actions, not just uncertain calculation. Blindfold chess does not do this: the state is fully knowable even when the player cannot see a board.
Player/Team Change
With more than two players, turn order and incentives can matter as much as piece activity. In coupled-board team games, a tactically bad capture on one board can be strategically correct because it transfers a required piece to a partner.
Balance Is Not the Same as Symmetry
A symmetric starting position can still give a first-move advantage. An asymmetric game can be competitively balanced by objective, material, turn or compensation rules. When evaluating balance, define the claim:
- equal theoretical outcome with perfect play?
- equal practical winning chances among matched players?
- balanced scoring across sides in a match format?
- stable metagame without a dominant strategy?
Those are different questions.
Randomness Needs Two Labels
Separate pre-game randomness from in-game randomness.
Chess960 randomizes the initial position before play; once the position is set, the game is deterministic. A game that uses random events during play creates a different decision model because future states are probabilistic.
Record the random source, generation constraints and seed/position number if reproducibility matters.
Provenance: Inventor, Earliest Source and Community Tradition
Modern designed variants may have a clear inventor and publication. Older community games may not. Historical chess-family games require evidence-aware wording. Use categories such as:
- documented inventor/publication;
- earliest currently located source;
- community attribution;
- uncertain or disputed origin.
“Widely credited to” and “first documented in” are safer than inventing a clean origin story when the evidence is messy.
Brand and Ruleset Must Stay Separate
An event brand can promote a standard ruleset. Current FIDE material, for example, identifies Freestyle Chess with Chess960/FIDE Fischer Random Chess, while Freestyle Chess also functions as the name of a commercial event ecosystem. The event brand is not a new castling rule.
The same principle applies to platform labels. A site can call a mode Fog of War, Giveaway or Four Player Chess; the product name does not automatically define the universal ontology.
Implementation Differences Are First-Class Data
Two platforms can use the same variant name but differ on:
- stalemate result;
- promotion options;
- king royal status;
- draw claims;
- atomic blast edge cases;
- bughouse termination;
- communication;
- pocket identity;
- notation/export conventions.
If a difference can change legal moves or results, it is not cosmetic. Record it as a versioned implementation profile.
Detect Degenerate or Dominant Strategies
When assessing a new variant, ask whether one mechanic overwhelms the rest of the game. Useful tests include:
- does one opening or policy dominate regardless of opponent response?
- can a player force an uninteresting loop or exploit?
- does the first move produce overwhelming structural advantage?
- does a scoring rule encourage collusion or non-play?
- does a newly introduced piece make others strategically irrelevant?
A variant can be fun despite imbalance, but the catalog should describe the actual experience rather than imply competitive balance without evidence.
A Durable Catalog Entry
A mature entry should include:
- canonical name and aliases;
- classification: ruleset/mode/historical relative;
- provenance and evidence level;
- rule authority + verification date;
- concise rule delta;
- complete Ruleset Profile link;
- decisive edge cases;
- strategic signature;
- notation/data signature;
- platforms/engines with version caveats;
- community/history notes;
- implementation differences.
The catalog then becomes a map to reproducible knowledge rather than a popularity list.
A Complete Variant-Analysis Workflow
For a new or unfamiliar game:
- classify what kind of chess activity it is;
- identify the rules source and implementation;
- write the rule delta;
- build the full Ruleset Profile;
- test edge cases that can change legality/results;
- derive strategy from the changed mechanics;
- identify orthodox heuristics that break;
- encode the full state and round-trip it;
- verify compatible engines/tools;
- record provenance, brand and version caveats;
- compare it with neighboring variants using defining conditions rather than names.
That workflow is sufficient to expand a variant encyclopedia indefinitely without losing conceptual discipline.
Test Balance With the Claim You Actually Need
A casual game does not need the same evidence as a competitive title event. Choose the relevant balance question first, then collect appropriate evidence:
- engine/self-play for theoretical tendencies;
- large platform data for practical side results;
- elite match data for high-skill balance;
- opening diversity for dominant-strategy signals;
- player feedback for experience and clarity.
Correct for selection effects where possible. Popular positions and experienced players are not random samples.
Catalog Change History
Rulesets evolve. Preserve a changelog when an implementation modifies starting setup, draw result, piece movement or scoring. Old games should remain interpretable under the rules that produced them rather than silently being re-labeled with the newest profile.
A mature catalog therefore contains both identity and version history.