Beats-Relation Matrix¶
Artifact — instantiates Cyclic Dominance Counterbalancing
A table or graph that records A-beats-B, B-beats-C, and C-beats-A relations, including confidence, context, and exceptions.
A Beats-Relation Matrix is the ledger that records who beats whom — one entry per ordered pair of options, each stating the direction of advantage, how confident you are in it, the context under which it holds, and the known exceptions. Its defining move is that it records direction, not magnitude: it does not say by how much A beats B, only that it does, and it insists on closing the loop so that when A beats B, B beats C, and C beats A, the cycle becomes visible on the page instead of being flattened into a leaderboard. It is the source-of-truth data structure the rest of the repertoire's governance reads from — the map, not yet the reading of it.
Example¶
A fighting-game community is arguing about which of three characters is "best." Every forum thread produces a different champion, because every player is arguing from a matchup they win. The Beats-Relation Matrix ends the argument by writing the relations down. The team lists the three characters as both rows and columns, then fills each off-diagonal cell from tournament replays: the grappler beats the zoner (closes distance and punishes), the zoner beats the rushdown (walls it out at range), and the rushdown beats the grappler (too fast to grab). Each cell carries a confidence tag ("high — 40+ recorded sets"), a context note ("on stages without walls"), and an exception ("reverses if the zoner has meter"). Rendered as a directed graph, the three arrows form a ring. The chart's payoff is not a winner but a shape: there is no best character, only a rock-paper-scissors loop, and anyone who bans the currently unpopular character is quietly removing the counter to whoever is winning this month.
How it works¶
- List the repertoire as nodes. The row/column set is the option repertoire; decoys and dominated options are excluded before any relation is drawn, so the graph has no phantom edges.
- Record each ordered pair as a directed edge. For every A→B, note the direction of advantage plus three annotations: a confidence tag (evidence quantity/quality), the activating context, and any exception that reverses the edge.
- Close and detect the loop. Because both directions of each pair are recorded, transitive chains and intransitive cycles both surface; a loop-detection pass flags any A→B→C→A ring.
- Keep it directional. Cells are booleans-with-caveats, not scores — the artifact deliberately refuses to encode "by how much," leaving magnitude to its numeric sibling.
Tuning parameters¶
- Edge resolution — binary "beats / loses" versus graded "usually / slightly." Finer grades capture near-ties but invite false precision and blur the clean loop the artifact exists to show.
- Confidence sourcing — how many observations, or whose judgment, an edge needs before it is drawn solid rather than dashed. Stricter sourcing suppresses phantom edges but leaves the map sparse early on.
- Context granularity — one global chart versus one chart per condition. Splitting reveals context reversals but multiplies maintenance and can fragment a real cycle into unreadable slivers.
- Exception handling — whether reversals live inline in the cell or in a separate register. Inline keeps them visible; a register keeps the main graph legible.
- Refresh cadence — how often edges are re-checked against fresh results, since a matchup chart goes stale the moment the population adapts.
When it helps, and when it misleads¶
Its strength is that it makes an intransitive structure undeniable: once the three arrows form a ring on the page, no one can honestly read a single champion off it, and the exceptions column stops any one edge from being over-generalized. It is the artifact that converts "everyone has an opinion" into "here is the relation, with its evidence."
Its failure mode is that a matrix will happily record a cycle that is not really there — impose a rock-paper-scissors story on what is actually a transitive ranking with noise, or a few small-sample flukes. This is the same trap the Condorcet paradox warns of: pairwise wins can look cyclic without a stable underlying loop.[n1] The classic misuse is the mirror image — reading a single "tier list" ranking off a chart whose arrows genuinely form a ring, eliminating the option that closes the loop. The guarding discipline is to treat the matrix as a claim to be tested, not trusted: draw edges only at a confidence threshold, keep exceptions visible, and hand the whole artifact to a pairwise audit before acting on it.
How it implements the components¶
option_repertoire— the row/column set is the curated repertoire; drawing the axes is the act of deciding which options are inside the cycle.counter_relation_matrix— the body of the artifact: every directed edge with its confidence, context, and exception annotations.dominance_cycle_map— recording both directions of each pair closes the loop, so the intransitive ring becomes a visible directed graph.
It records direction but not payoff magnitude, and it does not certify that the loop is real — the signed magnitudes belong to Cyclic Payoff Table, and the evidence verdict to Pairwise Dominance Audit.
Related¶
- Instantiates: Cyclic Dominance Counterbalancing — supplies the relational map every other mechanism reads from.
- Sibling mechanisms: Cyclic Payoff Table · Countermove Rotation Playbook · Portfolio Minimum-Viability Rule · Pairwise Dominance Audit · Adaptive Mix Review · Sentinel Option Trial · Nontransitive Scenario Simulation
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: A table or graph that records A-beats-B, B-beats-C, and C-beats-A relations, including confidence, context, and exceptions, making its operative form a non-executable information artifact that externalizes static or prospective structure.
Independent corroboration: The frozen evidence defines Beats-Relation Matrix as 'A table or graph that records A-beats-B, B-beats-C, and C-beats-A relations, including confidence, context, and exceptions', so its operative form is Representation, Specification & Plan.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Political Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Voting and social-choice theory established pairwise majority relations and the Condorcet cycle in which A beats B, B beats C, and C beats A.
Related originating lineages:
- Economics & Finance — Game theory and political economy analyze cyclic dominance and intransitive preferences.
- Mathematics — Directed graphs, binary relations, and adjacency matrices supply the representation of pairwise wins and cycles.
Review resolution: Authoritative social-choice references define Condorcet winners through pairwise contests and show that majority ordering can form A-beats-B-beats-C-beats-A cycles. That canonical structure is the page's central example and failure mode, placing political science first; economics and mathematics provide independently formative game-theoretic and graph representations.
Attribution caveat: The context-and-confidence ledger extends a voting relation into a generalized Encyclopedia artifact for any cyclic dominance system.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- Stanford Encyclopedia of Philosophy — Voting Methods
- Stanford Encyclopedia of Philosophy — Social Choice Theory
Notes¶
[n1] The Condorcet paradox — collective pairwise preferences can cycle (A beats B, B beats C, C beats A) even when each individual ranking is transitive, so no option is a pairwise winner. It is the canonical proof that a single ranking cannot always be read off pairwise comparisons, which is exactly why this artifact refuses to collapse the ring into a leaderboard. ↩