Center Of Gravity¶
Core Idea¶
A center of gravity is the single source of power, cohesion, or operational continuity in an antagonistic system whose disruption disproportionately changes the whole contest. The concept names a structural fact about adversarial systems: not every node matters equally, and there is usually one — or a small set — whose collapse cascades into systemic collapse, while disrupting any other node is locally costly but globally absorbed. The pattern is recognisable by three structural traits: high betweenness (many critical functions route through the node), cohesion-bearing (it supplies a unifying force — a logistics base, a shared protocol, a command authority, a keystone organism — without which the surrounding structure decoheres), and substitution-resistant (there is no cheap fallback, and loss is not recoverable on the timescale of the contest).
What makes the prime do work — rather than reduce to "important node" — is the adversarial framing. The center of gravity is named relative to a contest in which someone is trying to disrupt it, and the operational move that ports across domains is the paired diagnostic question: what is their center of gravity, and what is mine? The pattern lives in the coupling topology of the system, not in the substance of the node, which is why it is recognisable across material, organisational, informational, and biological substrates with the same three traits.
The adversarial presupposition also distinguishes the prime from its non-adversarial neighbours and gives it a characteristic iteration. Because hardening the obvious center of gravity forces the adversary to find or build a new one, the analysis is not one-shot but recurs after each intervention — the locus migrates under selection pressure, much as an equilibrium shifts in a contested market. This combination of cohesion-topology and adversarial contest is the load-bearing structure, and it is what the grading records as framed: the skeleton is real, but the Clausewitzian origin and the contest presupposition lean toward the human-practice end of the spectrum.
How would you explain it like I'm…
The Key Block
The One That Topples All
Contested Critical Node
Structural Signature¶
the adversarial contest with someone trying to disrupt — the candidate node set in the coupling topology — the cohesion-bearing node of high betweenness — the substitution-resistance on the contest timescale — the disproportionate-impact disruption — the defender/attacker duality — the locus migration under hardening
A node is a center of gravity when each of the following holds:
- An adversarial contest. The system is contested — someone is trying to disrupt it — and the node is named relative to that contest; the adversarial framing is what keeps the prime from reducing to "important node."
- A candidate node set in a coupling topology. A set of components, roles, services, or populations whose interconnection structure, not their substance, carries the analysis.
- High betweenness. Many critical functions route through the node.
- Cohesion-bearing. The node supplies a unifying force — a logistics base, a shared protocol, a command authority, a keystone organism — without which the surrounding structure decoheres.
- Substitution resistance. There is no cheap fallback and the loss is not recoverable on the timescale of the contest.
- Disproportionate impact. Its disruption cascades into systemic reorganization, while disrupting any other node is locally costly but globally absorbed.
- Defender/attacker duality with migration. The same analysis runs twice — defend mine, deny theirs — and hardening the obvious node forces the adversary to find or build a new one, so the locus migrates under pressure and the analysis recurs.
The components compose a structural compression of a contested system: most of it reorganises around a few load-bearing nodes, so intervention concentrates at the cohesion-bearing, substitution-resistant node — the same object the reliability engineer calls a single point of failure, seen from the defender's side.
What It Is Not¶
- Not a single point of failure.
single_point_of_failureis the defender-side, non-adversarial twin: a node whose failure halts the system under random fault. Center of gravity adds the adversarial contest — an optimizing attacker choosing which node to deny — and the locus migration under hardening that a static SPOF analysis lacks. - Not competition.
competitionis the broad rivalry for a contested resource. Center of gravity is the specific cohesion-topology insight within a contest — the one substitution-resistant node whose disruption reorganizes the whole — not the rivalry itself. - Not a bottleneck.
bottleneckis a throughput-limiting stage that caps flow. A center of gravity is a cohesion-bearing node whose loss causes systemic decoherence, identified relative to an adversary; a bottleneck constrains capacity, a COG holds the structure together under attack. - Not a leverage point.
leverage_pointsare general high-impact intervention sites. Center of gravity is the adversarial specialization — a cohesion-bearing, substitution-resistant node analyzed twice (defend mine, deny theirs) with migration under pressure — not the generic small-intervention-large-effect notion. - Not merely an important node. Importance alone (value, size) is not COG status: a node can be valuable yet substitutable (not a COG), or worthless alone yet cohesion-bearing (a single norm, key, relationship — a COG).
- Common misclassification. Expecting a cascade from disrupting a node whose impact is actually linear in its size — bombing morale, decapitation that does not decapitate. If the effect scales proportionally to the node's size rather than cascading, it is important, not the center of gravity.
Broad Use¶
- Military strategy (origin): the enemy's army, capital, alliance, or public will; WW2 air campaigns targeted industrial centers of gravity (ball bearings, transport, oil), and doctrine still codifies COG analysis as a step in operational design.
- Cybersecurity and resilience engineering: the single service — identity provider, certificate authority, key resolver — whose compromise cascades into systemic failure; threat modelling asks for the COG and defensive design duplicates or compartmentalises it.
- Organisational change: the unspoken cultural keystone — a founder, the one engineer who owns the deploy pipeline, an unwritten norm — whose departure redirects the whole organisation.
- Competitive business strategy: a rival's distribution network, key patent, talent monopoly, or platform lock-in as the disproportionate point of attack.
- Ecology and public health: keystone species (sea otters, wolves) whose removal triggers trophic cascades far beyond their biomass; superspreader nodes and bridge populations as transmission-network centers of gravity.
- Software systems and game strategy: the database, auth service, or build system whose outage halts the product (named "single point of failure" from the defender's side); the playmaking node whose neutralisation changes a contest more than matching any other element.
Clarity¶
Naming the center of gravity disciplines strategic reasoning to identify the one or two nodes whose disruption changes the contest, rather than dispersing effort across many merely important nodes. It separates important from load-bearing-in-the-contest: a thing can be valuable without being a COG (it has redundancy or substitutes), and a thing can be the COG without being valuable on its own (a single norm, a single key, a single relationship). This separation is what prevents the characteristic error of defending everything and therefore defending nothing.
The framing also exposes a common and costly error — misidentifying the COG. Many failed campaigns trace to a COG misread: bombing morale rather than the actual cohesion source, decapitation strikes that do not decapitate, layoffs that rationalise away the load-bearing engineer. The prime makes the diagnostic question explicit at every analysis: if I disrupted this node, would the contest change qualitatively? A node that fails this test is important but not the COG; a node that passes it concentrates the intervention. The clarifying force is to convert a diffuse sense of "what matters here" into a sharp, contest-relative judgement about which single disruption is disproportionate.
Manages Complexity¶
In any complex adversarial system the temptation is to model every node and edge. COG analysis is a structural compression that says most of the system will reorganise around the few load-bearing nodes, so identifying those simplifies the rest of the analysis. The intervention space collapses from "act on everything" to a small ordered list: defend my COG, deny theirs, and treat the remainder as second-order. The complexity-management work is priority-setting under adversarial constraint — without the prime, planners try to defend everything and so defend nothing; with it, defence and offence both concentrate at the points that actually move the contest.
The compression has a defined structure: locate the candidate node set, apply the cohesion function (which node's removal causes disproportionate reorganisation), apply the substitutability check (no cheap fallback on the contest timescale), and then concentrate the defender's family of moves {duplicate, distribute, harden, conceal} or the attacker's mirror {locate, isolate, deny, exploit}. Because the same machinery runs for both sides and the result is a short ordered list rather than a full system model, the pattern makes otherwise unbounded adversarial analysis tractable. The complexity it manages is the combinatorial complexity of a contested system with many nodes, reduced to the identification and protection-or-denial of a singular cohesion-bearing node.
Abstract Reasoning¶
The pattern licenses several portable inferences. The disproportionate-impact inference: if a node is the COG, expect its disruption to produce nonlinear, cascading effects rather than linear ones — and conversely, if such cascades follow a perturbation, a COG has probably been hit. The defense-offense duality: every COG analysis runs twice, asking what is mine (defend, harden, redundantise) and what is theirs (target, encircle, deny), using the same machinery for both.
Two further moves complete the toolkit. Hardening drives COG migration: defending the obvious COG forces the adversary to find or build a new one, so the analysis is iterative rather than one-shot, like equilibrium analysis under selection pressure. And the single-point-of-failure–center-of-gravity identity: the reliability engineer's SPOF and the strategist's COG are the same structural object viewed from defender and attacker respectively, so practices on one side transfer to the other. The reasoner asks, at every turn: which node has high betweenness and bears cohesion, is it substitution-resistant on the contest timescale, would its disruption change the contest qualitatively, and how will the locus migrate once I harden it?
Knowledge Transfer¶
The center of gravity transfers because the structural move — find the singular cohesion-bearing node under adversarial pressure and concentrate intervention there — survives the strip-the-jargon test, even though the Clausewitzian vocabulary and the contest presupposition keep it toward the framed end of the spectrum. The role mapping is consistent: the candidate node set maps to components, organisational roles, informational services, or biological populations; the cohesion function maps to whatever supplies the unifying force in each substrate; the substitutability check maps to the absence of a cheap fallback on the contest timescale; and the paired defender and attacker postures map identically across domains.
The transfers are documented. STRIDE and attack-tree practitioners explicitly borrow the COG move — "what is the crown jewel, what one thing ends the game if compromised?" — with the same diagnostic posture and intervention family. The keystone-species concept in ecology is functionally a COG analysis of food webs (identify the node whose removal causes disproportionate reorganisation), and the language has crossed back into "keystone technologies" and "keystone employees." In organisational succession planning, the disciplined question "who is the COG of this team?" surfaces hidden dependencies that org-chart analysis misses and motivates explicit redundancy before the COG departs. And the intervention family travels intact: once a node is named a COG, the same small menu — duplicate, distribute, harden, conceal, insure, fall back to an alternate — applies, and these are precisely not the interventions for ordinary nodes, so the prime selects the right toolkit. The honest qualifications are two: the prime overlaps heavily with single-point-of-failure (which is the COG seen from the defender's side, without the attacker's optimisation problem), and its military-institutional origin and adversarial-contest presupposition mean the vocabulary carries a frame into each new domain. The unifying transfer move nonetheless holds: establish the contest, find the cohesion-bearing substitution-resistant node, concentrate defence on yours and denial on theirs, and re-run the analysis as the locus migrates under pressure.
Examples¶
Formal/abstract¶
The Allied strategic bombing campaign against the German war economy is the prime's founding instance and instantiates every structural trait — including the costly failure mode of misidentifying the node. The adversarial contest is explicit: two sides, one trying to disrupt the other's war-making capacity. The candidate node set is the German economy's productive topology — factories, transport, raw materials, labor — and the analysis is about its coupling, not the substance of any one plant. Allied planners hypothesized a center of gravity with all three traits: ball-bearing production. Ball bearings have high betweenness (nearly every engine, gearbox, and weapon routes through them), are cohesion-bearing (without them the whole machine-building base decoheres), and were believed substitution-resistant on the contest timescale. The disproportionate-impact logic was the prime's exactly: destroy this one concentrated node (the Schweinfurt plants) and the cascade would reorganize the entire war economy, far more than destroying any equivalent tonnage of ordinary factories. But the case is also the prime's clearest lesson in COG misidentification and locus migration: the node proved less substitution-resistant than assumed — the Germans dispersed production, drew down stockpiles, redesigned to use fewer bearings, and imported from neutrals — so hardening pressure forced the COG to migrate, and the campaign's planners had to re-run the analysis, eventually concluding that oil and transport were the truer centers of gravity (genuinely substitution-resistant on the contest timescale). The intervention the prime names is precisely the corrected one: concentrate force at the cohesion-bearing, substitution-resistant node, verify substitution-resistance before committing, and expect the locus to move once the adversary adapts.
Mapped back: The strategic-bombing case is center of gravity in its origin substrate — a contested productive topology, a high-betweenness cohesion-bearing candidate node (ball bearings), the disproportionate-cascade logic, and the locus migrating under hardening toward the genuinely substitution-resistant nodes (oil, transport) — confirming that the analysis recurs and that mis-reading substitution-resistance is the characteristic failure.
Applied/industry¶
Two domains far from the battlefield — cybersecurity threat modeling and ecological keystone management — run the same cohesion-bearing-node-under-pressure structure (with the prime's caveat that the Clausewitzian frame travels by translation). In cybersecurity, the contest is explicit (defenders versus attackers), and the candidate node set is the system's service topology. The center of gravity is the service with all three traits: an identity provider or certificate authority has high betweenness (authentication for nearly every other service routes through it), is cohesion-bearing (its compromise lets an attacker impersonate anyone and the trust fabric decoheres), and is substitution-resistant (there is no quick fallback mid-incident). The prime's defender/attacker duality is the threat-modeling discipline verbatim — "what is our crown jewel, and what one compromise ends the game?" — and its intervention menu is the security playbook: duplicate (redundant CAs), distribute (compartmentalize), harden, conceal. The prime's locus-migration insight is the security reality that hardening the obvious COG pushes attackers toward the next-most-cohesive node, so the analysis recurs. Ecological keystone management maps cleanly: the "contest" is reframed as a pressure on the ecosystem (harvest, removal, invasion), the candidate node set is the food web, and a keystone species (sea otters, wolves) is the COG — its removal triggers a trophic cascade (kelp forests collapse as urchins explode; deer overgraze without wolves) far beyond its biomass, the prime's disproportionate-impact trait, and it is substitution-resistant because no other species fills its role on the relevant timescale. The intervention the prime names is the conservation move: identify the keystone before it is lost and protect it preferentially rather than spreading effort uniformly across species — the ecological analogue of "defend the COG, don't defend everything." The prime's honest caveat applies: on the defender's side this overlaps with single-point-of-failure analysis, and the adversarial framing is imported rather than native to ecology.
Mapped back: Cybersecurity crown-jewel analysis and ecological keystone management both instantiate a cohesion-bearing, high-betweenness, substitution-resistant node whose disruption causes disproportionate reorganization (trust-fabric collapse; trophic cascade), with the prime's defender/attacker duality and locus migration, so its intervention menu — duplicate, distribute, harden, protect preferentially — transfers from military strategy to security and ecology, with the Clausewitzian frame translated rather than recognized natively.
Structural Tensions¶
T1 — Important versus Load-Bearing-in-the-Contest (scopal). A node can be valuable without being a center of gravity (it has redundancy or substitutes) and can be the COG without being valuable on its own (a single norm, key, or relationship). The failure mode is defending what is important rather than what is load-bearing, dispersing effort across many valuable-but-substitutable nodes and so defending nothing. Diagnostic: ask whether disrupting this node would change the contest qualitatively. If its loss is locally costly but globally absorbed, it is important but not the COG; only nodes whose disruption cascades concentrate the intervention.
T2 — Substitution-Resistance Assumed versus Actual (measurement). The COG must be substitution-resistant on the contest timescale; a node assumed irreplaceable but actually substitutable is a false COG, as the ball-bearing campaign learned. The failure mode is committing force to a node the adversary can route around — dispersing production, drawing down stock, redesigning to need less. Diagnostic: ask whether the adversary has a fallback reachable within the contest's timescale. If the node can be substituted before the disruption bites, its cohesion-bearing status is illusory; verify substitution-resistance before concentrating force, or the cascade will not come.
T3 — Hardening versus Locus Migration (temporal). Defending the obvious COG forces the adversary to find or build a new one, so the locus migrates under pressure and the analysis recurs rather than resolving once. The failure mode is treating COG analysis as one-shot — hardening the current node and assuming the contest is settled, while the cohesion point has already moved. Diagnostic: after hardening, ask where the next-most-cohesive node now sits. If the analysis is not re-run after each intervention, the defense protects a node the adversary has already abandoned; the locus moves like an equilibrium under selection pressure.
T4 — Defender's SPOF versus Attacker's Optimization (sign/direction). The same node is a single point of failure from the defender's side and a target from the attacker's; the two postures run the same machinery with opposite sign, and the attacker additionally solves an optimization (which COG is cheapest to deny). The failure mode is reasoning from only one side — hardening as if no adversary optimizes, or targeting as if the defender cannot relocate. Diagnostic: run the analysis twice, defend-mine and deny-theirs. If only the defender's SPOF view is taken, the attacker's choice among one's COGs is unmodeled, and the hardening may protect the wrong one.
T5 — Disproportionate Cascade versus Linear Effect (scalar). A true COG produces nonlinear, cascading effects on disruption; an ordinary node produces effects roughly proportional to its size. The failure mode is misreading a linear-impact node as a COG (expecting a cascade that never comes — bombing morale, decapitation that does not decapitate) or missing that a small node is a COG because its biomass or budget is small. Diagnostic: ask whether disruption would scale linearly or cascade. If the expected effect is proportional to the node's size, it is not the COG; if a small or cheap node would trigger systemic reorganization, it is, regardless of its standalone value.
T6 — Adversarial Frame versus Native Substrate (scopal/framed-boundary). The prime presupposes a contest with someone trying to disrupt, and its Clausewitzian vocabulary imports that frame — which fits military, security, and business but is translated, not native, in ecology or reliability (where the "adversary" is harvest pressure or random failure). The failure mode is forcing an adversarial reading onto a non-contested system, or conflating the COG with a plain single-point-of-failure that lacks an optimizing attacker. Diagnostic: ask whether a genuine adversary is optimizing against the node. If the pressure is non-strategic (random failure, environmental stress), the defender-side SPOF analysis applies but the attacker's optimization does not; the portable core is the cohesion-bearing-node insight, not the contest presupposition.
Structural–Framed Character¶
Center of gravity sits on the framed side of the structural–framed spectrum, at the midpoint aggregate of 0.5. There is a genuine substrate-neutral skeleton underneath — a cohesion-bearing, high-betweenness, substitution-resistant node whose disruption reorganizes a contested system — but the prime is wrapped in a strong Clausewitzian-military frame and an adversarial-contest presupposition, and these are decisive enough to place it firmly framed, driven above all by a maxed-out institutional-origin score.
Walking the diagnostics: institutional origin is the heaviest, scored a full 1.0. The concept is Clausewitzian military doctrine through and through — "center of gravity," "decapitation," "lines of operation," codified as a step in operational design — and that institutional origin clings to every application; even ecology borrows the language back ("keystone" reframed as COG) rather than the term arising natively. Vocabulary travels with effort, scored 0.5: the strip-the-jargon test passes (the cohesion-bearing-node insight survives into cybersecurity crown-jewels and ecological keystones), but the military lexicon must be translated each time and the prime's own text says it "carries a frame into each new domain." Human-practice-boundness is partial, scored 0.5: the prime presupposes a contest with someone trying to disrupt, which is a human-strategic framing — yet the structure also applies where the "adversary" is non-strategic (harvest pressure, random failure), so it is contest-leaning but not strictly bound, and the prime explicitly notes the adversarial frame is translated rather than native in ecology and reliability. Import-versus-recognize sits at 0.5: invoking COG partly recognizes a real coupling-topology fact one can test (would disruption cascade or stay linear?) and partly imports the adversarial-strategic reading. The one diagnostic holding it back from deeper framed is evaluative weight, scored 0: a center of gravity is neither good nor bad — it is a value-neutral structural locus, equally the thing you defend and the thing you attack. That value-neutrality, plus the real cohesion-topology skeleton (which the prime notes overlaps single-point-of-failure on the defender's side), is what keeps the prime at the 0.5 midpoint rather than fully framed — faithful to the framed label and to the prime's own reading that the skeleton is real but the Clausewitzian origin and contest presupposition lean human-practice.
Substrate Independence¶
Center of Gravity is a strongly substrate-independent prime — composite 4 / 5 on the substrate-independence scale. Its underlying insight is about coupling topology — within an interdependent system there is a cohesion-bearing node or hub whose disruption disproportionately degrades the whole, and which therefore becomes the decisive point under adversarial pressure — and that topological insight ports cleanly across distinct domains, giving it maximal domain breadth (rated 5): cybersecurity (the critical asset or trust anchor), ecology (the keystone species whose removal collapses the web), and business strategy (the capability or relationship a competitor most depends on). What pulls the composite down to 4 is structural abstraction, scored 3: the vocabulary and the surrounding frame are strongly Clausewitzian-military — "center of gravity," decisive point, lines of operation — so applying the prime tends to import a strategic-warfare reading rather than presenting a clean medium-neutral statement of the hub-under-attack relation. The transfer evidence is 4: the keystone and critical-node mappings are concrete and recognizable, though they travel as analogical re-applications of a military concept rather than a shared formalism carried across substrates. Maximal breadth over a militarily-framed signature, held back by that inherited frame, places the composite at 4.
- Composite substrate independence — 4 / 5
- Domain breadth — 5 / 5
- Structural abstraction — 3 / 5
- Transfer evidence — 4 / 5
Relationships to Other Primes¶
Parents (1) — more general patterns this builds on
-
Center Of Gravity is a kind of, typical Leverage Points
Center of gravity is the ADVERSARIAL specialization of leverage_points — a cohesion-bearing, substitution-resistant node analyzed twice (defend mine / deny theirs) with locus migration under hardening. The file: COG is 'the adversarial specialization' of the general small-intervention-large-effect notion.
Children (1) — more specific cases that build on this
-
Single Point of Failure is a kind of, typical Center Of Gravity
*** single_point_of_failure is a CANDIDATE (CAND-R2-197-02), not canonical — recorded as a candidate-link, NOT a corpus reparent. *** The file: SPOF is the COG 'seen from the defender's side', the same structural object without the optimizing attacker + migration. COG adds the adversary; whether COG parents SPOF or they are dual views is the open question.
Path to root: Center Of Gravity → Leverage Points → Feedback
Neighborhood in Abstraction Space¶
Center Of Gravity sits in a sparse region of abstraction space (97th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely rather than landing on a neighbor.
Family — Unclustered & Miscellaneous (91 primes)
Nearest neighbors
- Sacrifice Periphery To Defend Core — 0.67
- Adversarial Boundary Navigation — 0.66
- Interior Lines — 0.66
- Competition — 0.65
- Overton Window — 0.65
Computed from structural-signature embeddings · 2026-06-14
Not to Be Confused With¶
Center of gravity's most consequential confusion is with single_point_of_failure, because the two are genuinely the same structural object viewed from opposite sides, and the prime itself flags the overlap. A single point of failure is a node whose loss halts the system, identified from the defender's, non-adversarial standpoint: it asks "what one failure, by accident or random fault, brings everything down?" Center of gravity adds two things SPOF lacks. First, the adversarial contest: a COG is named relative to an optimizing attacker who is actively choosing which node to deny and who solves the additional problem of which of the defender's vulnerable nodes is cheapest to attack — a strategic optimization absent when the threat is random failure. Second, the locus migration: because hardening the obvious node forces an intelligent adversary to find or build a new cohesion point, COG analysis is iterative, re-run after each intervention, whereas a SPOF analysis against random fault can be one-shot. The practical consequence is that on the defender's side, where the threat is non-strategic (a server that fails randomly), SPOF analysis suffices; but where a genuine adversary is optimizing, the COG frame is required, because it models the attacker's choice and the moving target that SPOF does not. Conflating them leads to hardening a static SPOF as if the contest were settled, while an adaptive adversary has already shifted to the next cohesion node.
Center of gravity must also be distinguished from bottleneck, with which it shares the feature of a single critical node but differs in what kind of criticality and whether an adversary is present. A bottleneck is a throughput-limiting stage: the narrowest point in a flow that caps the rate of the whole system, and relieving it increases capacity. A center of gravity is a cohesion-bearing node: the point whose disruption causes the surrounding structure to decohere, identified relative to a contest. The two criticalities are different — a bottleneck constrains how much can flow, a COG holds the system together under attack. A node can be a bottleneck without being a COG (a slow stage everyone routes through, but whose removal would not collapse the system's cohesion, only its speed), and a COG without being a bottleneck (a command authority or shared protocol that imposes no throughput limit but whose loss decoheres everything). Conflating them sends the analyst to relieve a capacity constraint (the bottleneck fix: add parallel capacity) when the actual concern is a cohesion node an adversary will target (the COG fix: duplicate, distribute, harden, conceal). The bottleneck lens optimizes flow; the COG lens defends or denies cohesion under contest.
A third genuine confusion is with leverage_points, because a center of gravity is a paradigm case of a place where a small intervention produces large effect. The distinction is between the general concept and an adversarial specialization. Leverage points are any sites in a system where a well-aimed intervention yields disproportionate change — a broad notion spanning feedback gains, rule changes, and paradigm shifts, and applicable to non-contested systems being improved. Center of gravity is the specific case under adversarial pressure: a cohesion-bearing, substitution-resistant node, analyzed through the paired defender/attacker postures (defend mine, deny theirs), with the locus migrating as the contest adapts. Not every leverage point is a COG — improving a system's goals or information flows is high-leverage but involves no adversary, no cohesion-decoherence, no migrating target — and the COG carries commitments (substitution-resistance on the contest timescale, the attacker's optimization, the iterative re-analysis) that the generic leverage notion does not. Treating a COG as merely "a leverage point" loses its adversarial machinery: the dual analysis, the substitution-resistance check that the ball-bearing campaign learned the hard way, and the expectation that the locus moves once hardened.
These distinctions matter because each isolates what the center of gravity adds: a single point of failure is the defender-side, random-fault twin (where COG adds the optimizing attacker and migration), a bottleneck is a throughput constraint (where COG is cohesion-decoherence under contest), and a leverage point is the general high-impact site (where COG is its adversarial specialization). A practitioner who conflates them hardens a static SPOF against an adaptive foe, relieves capacity when cohesion is the concern, or loses the dual-posture machinery in a generic leverage frame. Holding center of gravity as the specific cohesion-bearing / high-betweenness / substitution-resistant node analyzed under an adversarial contest with locus migration keeps the analyst asking its real questions — what is theirs and what is mine, is the node truly substitution-resistant on the contest timescale, and where will the locus migrate once I harden it?
Solution Archetypes¶
No catalogued solution archetypes reference this prime yet.