Skip to content
Chess Composition, Studies & ProblemsGuide

How to Record & Verify a Chess Composition

Use a reproducible composition record to preserve stipulation, authorship, diagram, solution, themes, soundness, software verification, provenance, originality research, rights and dated institutional rules.

A well-documented composition should be reproducible as both chess task and authored work: another specialist should be able to reconstruct the rules, solution, verification and provenance without guessing.

Composition Card

Composition

Authorship / Provenance

  • composer(s):
  • source/publication:
  • publication date:
  • original / correction / version:
  • award/distinction:
  • priority/predecessor notes:
  • rights/license:

Diagram

  • diagram:
  • FEN if orthodox and sufficient:
  • side to move:
  • legality status:
  • retro assumptions:
  • castling rights:
  • en-passant rights:

Stipulation

  • plain-language task:
  • shorthand:
  • number of solutions:
  • twins:
  • set play indicated?;
  • fairy pieces:
  • fairy conditions:
  • series/half-move convention:

Author's Solution

  • key / first move:
  • threat:
  • defenses:
  • variations:
  • tries/refutations:
  • set play:
  • twins/solutions:
  • endpoint:

Themes

For each:

  • specialist name:
  • formal definition:
  • pieces/defenses/functions involved:
  • phase structure:
  • mechanism:
  • minimal schematic:
  • anti-example:

Soundness

  • expected solutions:
  • found solutions:
  • cook?;
  • duals?;
  • short?;
  • no-solution?;
  • legality?;
  • unintended promotion?;
  • twin defect?;

Verification

  • software:
  • version:
  • options:
  • tablebase if study:
  • human reviewer:
  • verification date:

Artistic Note

  • economy:
  • unity:
  • paradox:
  • difficulty:
  • naturalness:
  • originality:
  • technical pieces:
  • artistic trade-offs:

Stipulation Reader

Before solving, restate:

  1. who moves first?;
  2. who resists/cooperates?;
  3. target?;
  4. move maximum/exact length?;
  5. solution count?;
  6. twins?;
  7. fairy rules?;
  8. special composition conventions?

If any answer is unknown:

  • do not calculate yet.

Soundness Test

Soundness Test

  • program/version:
  • stipulation encoding:
  • expected solution count:
  • solutions found:
  • cooks:
  • duals:
  • short solutions:
  • legality:
  • promotion alternatives:
  • castling:
  • en passant:
  • twins:
  • conditions:
  • unsupported-feature warning:
  • human review.

Codex Convention Record

Store:

  • Codex version;
  • adopted FIDE-law framework;
  • article/footnote if relevant;
  • historical convention if older work.

Current official anchor (checked August 2026):

  • WFCC Codex page: Vilnius 2019.

Never infer:

  • current OTB rule = current composition convention.

Phase Table

PhaseInitial move / premiseDefenseContinuation / mateFunction
Set...aA...
Try1.?aB / refutation...
Actual1.!aC...

Use the table to diagnose:

  • changed;
  • transferred;
  • reciprocal;
  • cyclic;
  • Zagoruiko;
  • Le Grand;
  • Dombrovskis;
  • correction.

Directmate Verification

For #2:

  • key;
  • threat/block status;
  • every defense;
  • every mate;
  • set/try play;
  • flights;
  • no cook.

For #3/moremover:

  • also record:
  • preparatory move;
  • second White move;
  • mainplan;
  • foreplan;
  • logical obstruction;
  • full-length branches.

Helpmate Record

  • Black begins?;
  • move count;
  • half-move?;
  • solutions:
  • twins:
  • final mating pictures:
  • move-order uniqueness:
  • model/echo relation:
  • dualization check.

Selfmate Record

  • desired black mating move/picture:
  • White compulsion:
  • Black resistance:
  • threat;
  • defenses;
  • forced checks/captures:
  • full solution.

Study Verification

  • FEN:
  • side to move:
  • win/draw:
  • key/unique first move:
  • best defense:
  • main line:
  • tolerated/serious dual note:
  • endpoint:
  • tablebase coverage:
  • WDL/DTZ if relevant:
  • tablebase source/date:
  • artistic idea:
  • tablebase truth ≠ artistic verdict.

Retro Record

  • missing material:
  • capture budget:
  • pawn origins:
  • promotion candidates:
  • last move candidates:
  • checks:
  • castling rights:
  • en-passant rights:
  • piece identities:
  • forced/possible history:
  • contradiction if illegal.

Proof Game Record

  • stipulated length:
  • exact vs shortest:
  • route constraints:
  • tempos:
  • captures:
  • promotions:
  • final diagram:
  • uniqueness:
  • solver/test software.

Fairy Rule Card

For every piece:

  • name:
  • notation:
  • move:
  • capture:
  • rider/leaper/hopper?;
  • hurdles/blockers:
  • promotion/royal status:

For every condition:

  • specialist name:
  • version/source:
  • what rule changes:
  • check/mate effects:
  • capture/rebirth effects:
  • ownership/color effects:
  • interaction with other conditions.

Never assume:

  • name alone is enough.

Theme Card

Composition Theme

  • name:
  • aliases:
  • genres:
  • formal relation:
  • mechanisms:
  • phase relation:
  • schematic:
  • reference example:
  • anti-example:
  • historical attribution/source:
  • ambiguity warning:

Use a named theme only after its formal specialist definition has been verified.

Originality / Priority Card

  • first-publication date:
  • corrected/version date:
  • databases searched:
  • search terms/position:
  • predecessors:
  • full anticipation?;
  • partial anticipation?;
  • matrix similarity only?;
  • judge/director determination:
  • source links.

Do not convert:

  • database found nothing

into:

  • first ever.

Competitive Solving Record

Current WCSC/ECSC rules snapshot:

  • accepted: 1 Nov 2025;
  • Round 1: 3 #2 — 20 min;
  • Round 2: 3 #3 — 60 min;
  • Round 3: 3 studies — 100 min;
  • Round 4: 3 h# — 50 min;
  • Round 5: 3 moremovers — 80 min;
  • Round 6: 3 s# — 50 min.

Also record:

  • exact solution-writing requirement;
  • points;
  • elapsed time;
  • no electronic solving aids;
  • event edition/date.

Recheck against the rules for the event or edition you are documenting.

FIDE Album Record

Current snapshot:

  • three-year periods;
  • 26 Albums reported by WFCC;
  • latest published: 2019–2021, July 2025;
  • eight sections;
  • three judges per section/subsection;
  • 0–4 points/judge;
  • at least 8 total points for selection.

Composer-title current snapshot:

  • problem: 1 point;
  • study: 1.67 points;
  • joint work divided among composers;
  • GM: 70;
  • IM: 25;
  • FM: 12.

Recheck against the current Album instructions/title regulations.

WCCI Versioning Record

For 2025–27:

  • 68th WCCC decision, Magdeburg, Aug 2026;
  • 4–6 entries per section retained;
  • judging panel reduced 5 → 3;
  • same judges for corresponding FIDE Album cycle.

If the older generic WCCI rules page still says five:

  • cite the later decision;
  • flag consolidation pending.

For every reference example:

  • composer;
  • first publication;
  • source;
  • award;
  • reproduction right/license;
  • diagram-right considerations if applicable;
  • whether solution/commentary is newly authored.

Do not build the future corpus from unattributed scraped diagrams.

Reproducible Publication Checklist

  • [ ] Stipulation stated in words?
  • [ ] Specialist shorthand defined?
  • [ ] Codex/version appropriate?
  • [ ] Diagram legal under correct convention?
  • [ ] Cook checked?
  • [ ] Dual classified, not mislabeled?
  • [ ] Short solution checked?
  • [ ] Multiple solutions/twins intended?
  • [ ] Set/try/actual phases separated?
  • [ ] Theme formally defined?
  • [ ] Fairy rules fully declared?
  • [ ] Study verified exactly where possible?
  • [ ] Computer solver/version recorded?
  • [ ] Human artistic review separate?
  • [ ] Originality/predecessor search performed?
  • [ ] Composer/provenance preserved?
  • [ ] Rights/license resolved?
  • [ ] Current WFCC competition facts date-stamped?
  • [ ] OTB titles/rules not conflated?
Key idea

A composition is not fully documented until another specialist can reproduce the task, solution, rule conventions, verification and authorship without guessing.

Verification Has Four Independent Layers

A reproducible composition record should let another specialist answer four questions without guessing:

  1. What exactly is the task? — diagram, stipulation, conventions, solutions and twins.
  2. Is it technically correct? — legality, cooks, duals, short solutions and solver output.
  3. What does it express? — phases, themes, mechanism and artistic note.
  4. What is the authored object? — composer, publication, version, award, predecessor research and rights.

Do not let a successful computer test stand in for the other three layers.

Date Every Moving Institutional Claim

The WFCC Codex currently presented on the official site is the Vilnius 2019 version. The WCSC/ECSC rules page states that its current version was accepted on 1 November 2025. WFCC reports 26 FIDE Albums, with the 2019–2021 Album published in July 2025. In August 2026 the 68th WCCC approved a three-judge WCCI format for the 2025–27 cycle. These facts are useful precisely because they are versioned; future updates should replace the dated snapshot, not silently coexist with it.

Verification Is Reproducibility, Not Authority

Software output becomes useful evidence only when the program, version, options, stipulation encoding and expected solution count are known. A database search becomes useful originality evidence only when the database and query strategy are recorded. A tablebase result becomes useful study evidence only when coverage and interpretation are clear.

The goal is not to make every reader rerun every check. It is to make rerunning the check possible.

What to Save When the Position Is Not Fully Expressible by FEN

FEN is convenient for orthodox positions, but it cannot encode every composition convention and is especially inadequate for many fairy, retro, series or multi-solution tasks. Preserve the diagram as published, the stipulation in plain language, the side-to-move convention and any special rights or historical assumptions. Treat FEN as an additional machine-readable representation when sufficient, not as the authoritative object in every genre.

Separate Solver Verification From Human Review

A strong record should contain two conclusions:

  • technical verification — what the software found under a specific encoding;
  • human review — whether the encoding matched the intended stipulation and whether the resulting play has the claimed thematic/artistic interpretation.

This separation catches an important failure mode: a solver can prove that the wrong task is perfectly sound if the task was encoded incorrectly.

Version Corrections Rather Than Overwriting Them

When a composition is corrected after publication, preserve both records. A correction may change material, key, solution or thematic content, and historical priority can depend on which version existed at which date. Use explicit labels such as original, corrected version and later reconstruction rather than silently replacing the published work.

Rights and Attribution Belong in the Technical Record

A chess position can be mechanically represented as coordinates, but a composed work also has authorship and publication history. A public corpus should not infer that a diagram is free to reproduce merely because its moves can be encoded. Preserve the source and rights basis with the same seriousness as the solution and soundness record.