Exposure Creep¶

Prime #: 850
Origin domain: Environmental Science Sustainability
Subdomain: hazard and risk management → Environmental Science Sustainability

Core Idea¶

Exposure creep is the pattern in which, between rare-but-severe shocks, valuable assets — people, capital, infrastructure, organisms, information stores — gradually accumulate inside the impact zone of a known but low-frequency hazard, because the absence of the shock during the accumulation period is read as evidence that the location is safe, and because each unit of accumulation locally lowers the marginal cost of the next. By the time the shock arrives, the total stake exposed is many times what it was the last time the hazard materialised, and the loss is correspondingly larger even though the hazard itself has not intensified.

The arrangement carries five roles. There is a hazard with a long mean inter-arrival time and severe impact when it arrives. There is an impact zone in which its damage concentrates. There is a protective measure — physical, regulatory, financial, or cognitive — that has lowered the short-horizon frequency or apparent severity, leaving short-horizon observers with a flat damage history. There is a value gradient that draws assets into the impact zone: good farmland in floodplains, cheap land near fault lines, yield-chasing assets in opaque markets. And there is a recency-weighted decision rule among the asset-placers that weights the recent flat history more heavily than the long-period hazard distribution. The result is monotonic accumulation between shocks, so that when the shock arrives the apparent "increase in losses" is real but mechanically driven by the stock placed in harm's way, not by any change in the hazard process.

The pattern is distinct from behaviour change in response to a safety measure and from gradual deterioration of an existing stock. It is the gradual placement of additional stock into a known impact zone — not a behavioural adjustment, and not the slow damaging of stock already present.

How would you explain it like I'm…

Crowding The Danger Spot

Imagine a spot by a river that floods only once in a great while. Because it hasn't flooded in a long time, more and more people build houses there. When the flood finally comes, it ruins way more houses than last time, even though the flood is the same size.

Piling Up In Harm's Way

Some dangers, like a big flood or earthquake, only happen rarely. In between, the place stays calm, so people think it must be safe and they keep moving valuable things there: homes, farms, factories. Each new thing makes it cheaper and easier for the next one to move in too. So when the rare disaster finally hits, it destroys a giant pile of stuff, not because the disaster got stronger, but because so much more was sitting in its path.

Stake Creep Between Shocks

Exposure Creep is when valuable things slowly pile up inside the danger zone of a rare-but-severe hazard during the long calm between hits. The quiet stretch gets misread as proof the place is safe, and each thing added there lowers the cost of adding the next, so the stock grows steadily. By the time the hazard strikes, the total at risk is many times bigger than last time, so the loss is far larger. The key point is that the hazard itself never intensified; only the amount of stuff exposed did. This is different from a place simply becoming more dangerous over time.

Exposure Creep is the pattern where, between rare-but-severe shocks, valuable assets (people, capital, infrastructure, organisms, data) gradually accumulate inside a known hazard's impact zone, so the eventual loss is far larger even though the hazard process is unchanged. It runs on five roles: a hazard with a long mean inter-arrival time and severe impact; an impact zone where damage concentrates; a protective measure that flattened the short-horizon damage history; a value gradient pulling assets inward (good farmland on floodplains, cheap land near faults, yield in opaque markets); and a recency-weighted decision rule that over-weights the recent flat record against the long-run hazard distribution. Because each added unit locally lowers the marginal cost of the next, accumulation is monotonic between shocks. When the shock arrives, the apparent jump in losses is real but mechanically driven by the stock placed in harm's way, not by any change in the hazard. It is therefore distinct from a behavioral response to a safety measure and from the slow decay of stock already present. It is specifically the gradual placement of additional stock into a known impact zone.

Structural Signature¶

the rare-but-severe hazard with long inter-arrival time — the impact zone where its damage concentrates — the protective measure that flattens short-horizon frequency — the value gradient drawing assets into the zone — the recency-weighted placement rule — the monotonic accumulation of stake between shocks

A system exhibits this pattern when each of the following holds:

A rare, severe hazard. A hazard with a long mean inter-arrival time and severe impact when it arrives.
An impact zone. A region in which the hazard's damage concentrates and into which assets can be placed.
A protective measure. A physical, regulatory, financial, or cognitive defence that lowers the short-horizon frequency or apparent severity, leaving short-horizon observers with a flat damage history.
A value gradient. An incentive — good land, cheap access, high yield — that draws valuable assets into the impact zone.
A recency-weighted decision rule. Asset-placers weight the recent flat history more heavily than the long-period hazard distribution, so the quiet period reads as evidence of safety.
Monotonic stake accumulation. Between shocks the stake exposed grows monotonically, so when the shock arrives the loss is driven by the grown stake, not by any change in the hazard process.

These compose so that expected loss decomposes into hazard-frequency × severity × stake-exposed, with stake the non-stationary term; the protective measure is an upstream cause of the larger eventual loss, and the structural lever is to break the recency-weighting that feeds the accumulation.

What It Is Not¶

Not risk. risk is the general probability-weighted prospect of loss; exposure creep is the specific non-stationary growth of the stake term between rare shocks while frequency and severity look stationary. Risk names the whole prospect; exposure creep names which term is silently moving.
Not risk-return tradeoff. risk_return_tradeoff is a deliberate acceptance of more risk for more reward; exposure creep is an inadvertent accumulation driven by recency-weighting reading quiet as safety — no one chose the larger stake.
Not risk pooling. risk_pooling spreads exposure across many to reduce variance; exposure creep concentrates stake in one impact zone. Nearly opposite operations on exposure.
Not layered accumulation. layered_accumulation is value-neutral stratification over time; exposure creep is specifically accumulation inside a hazard's impact zone whose loss the next shock realizes — the hazard and the recency-weighted placement rule are essential, not incidental.
Not arbitrage. arbitrage_finance exploits a price discrepancy for low-risk gain; exposure creep is the build-up of stake at risk under a misleadingly quiet history — no riskless spread, just a growing exposed stock.
Common misclassification. Reading a larger loss as evidence the hazard intensified ("the world got more dangerous"). Catch it by decomposing loss into frequency × severity × stake and asking which term moved — on the way out of a quiet period it is almost always stake.

Broad Use¶

The pattern recurs wherever a protective measure produces a quiet period that is then used as an accumulation window. In floodplain and coastal development — the origin — levees and seawalls reduce flood frequency, protected land is developed, and total assets-at-risk grow even as the return period appears favourable: the "levee effect" or "safe development paradox." In the wildland–urban interface, decades of fire suppression reduce visible fire frequency while subdivisions push deeper into fuel-loaded forest, so one ignition exposes thousands of homes. In earthquake-prone urban cores, long inter-event times let high-density development concentrate in liquefaction zones, so the next major quake exposes a stake an order of magnitude larger. In financial systems, long quiet volatility regimes shift capital into yield-chasing, correlated, or opaque assets, so the downturn stake is many times the last cycle's. In public health, decades without a pandemic concentrate pharmaceutical and PPE manufacturing in one or two clusters, so a single regional disruption exposes the world's supply. In infrastructure, undersea cables, GPS, and single-provider cloud regions accumulate dependent services over quiet years until an outage exposes a stake unrelated to the original service's stated reliability. In ecology, a refugium habitat accumulates species and biomass during quiet periods until a single bad year wipes out a far larger stake than it once held.

Clarity¶

Naming exposure creep makes visible the distinction between hazard frequency — a property of the environment — and stake exposed — a property of how assets are allocated. It reframes rising disaster losses from "the world is getting more dangerous," which is sometimes true and sometimes not, to "we have placed more value where the hazard hits," which is almost always true on the way out of a quiet period. By separating the hazard process from the stake process, the frame prevents the analyst from misreading a loss driven by accumulation as evidence of an intensifying hazard.

The frame also makes visible an uncomfortable structural fact: the protective measure that reduced short-horizon frequency is part of the cause of the larger eventual loss, not external to it. The levee did not fail; the levee bought time, and the time was used to accumulate stake. Recognising the protection as an upstream driver of the stake, rather than as an unalloyed good, is the clarifying move — it converts "the defence let us down" into "the defence performed exactly as designed and, by doing so, enabled the accumulation that the next shock converted into loss."

Manages Complexity¶

The pattern absorbs a wide catalogue of seemingly unrelated "the surprise was bigger this time" stories — major floods, financial crises, tsunami inundations, supply shortages, wildfire town losses, large-scale software outages — and reads them through a single mechanism: a known hazard's quiet period was used as the accumulation window for a stake whose size at the next event is the dominant driver of loss. The analyst no longer memorises each event separately; the mechanism predicts the same pattern in domains where the next event has not yet occurred.

The compression also supports quantitative reasoning, because expected loss in any period decomposes into hazard frequency times hazard severity times stake exposed, and exposure creep is precisely the term that grows monotonically between shocks while the first two appear stationary or favourable. Modelling the stake-placement process as a function of recency-weighted loss history is then a standard step — in catastrophe modelling, financial stress testing, and ecological vulnerability assessment — and the mechanism predicts that naive extrapolation of a historical-loss series will systematically underestimate future loss, because the stake is non-stationary while the historical series anchors expectations to past stakes. The frame thus tells the analyst which term in the loss decomposition is moving and why the obvious extrapolation misleads.

Abstract Reasoning¶

The mechanism licenses inference about when exposure creep can be stopped: it stops when the asset-placement decision is keyed to the hazard's long-period distribution — via mandatory disclosure, insurance pricing that reflects the distribution, zoning that does not respond to short-horizon quiet periods, or capital-adequacy rules that hold against the long-period distribution — rather than to recency-weighted loss history. The structural lever is always the same: break the recency-weighting that feeds the accumulation.

The frame also sharpens its own boundaries by contrast. Exposure creep is not risk compensation: there the agent changes behaviour in response to a safety measure, whereas here additional stock arrives and need not "behave" at all. It is not gradual deterioration, which damages the same stock over time, whereas exposure creep adds stock that is fine until the hazard event. It is not systemic risk, which concerns local failures propagating system-wide, whereas exposure creep is the upstream mechanism that raises the stake on which any systemic event operates. And it can include but does not require a moral-hazard component, since even fully internalised, recency-weighted decision rules produce exposure creep on their own. The reasoning carries a mild normative tinge from its risk-management lineage — the implicit "exposure creep should be prevented" — which places the prime toward the structural end of a mixed-structural classification: its roles travel cleanly, but its home contexts are human asset-placement settings, and some natural analogues are weaker.

Knowledge Transfer¶

The interventions transfer because the roles map cleanly across substrates: the hazard maps to a flood, fire, earthquake, market crash, pandemic, or outage; the impact zone maps to a floodplain, a fuel-loaded interface, a liquefaction basin, a correlated-asset cluster, a manufacturing concentration, or a single-vendor dependency; the protective measure maps to a levee, fire suppression, an insurance regime, or a reliability guarantee; and the recency-weighted decision rule recurs identically wherever a flat recent history governs placement. Because the roles correspond, the intervention family — decouple asset-placement from recency-weighted loss history and key it to the long-period hazard distribution — is the same move in every domain.

The documented transfers run in many directions. The hazards-geography "levee effect" and "safe development paradox" generalise out of flood-control vocabulary to non-physical hazards without losing their shape. Catastrophe modelling's exposure-data construction is the step that operationalises the prime, and the prime is the conceptual scaffold for why that step is load-bearing. Behavioural finance supplies the recency-weighted decision rule itself — "this time is different" is its verbal form — linking the geography cases to the financial-stability ones. Infrastructure-dependency analysis ports the move from "did this service have an outage?" to "what stake depends on this service?", which is this prime under another name. And conservation biology's refugium-concentration risk is the same mechanism with biological assets. Across these, the canonical intervention recurs unchanged: zone or price or capitalise against the unprotected, long-period hazard, so that placement decisions cannot be captured by the quiet period. The transfer is largely structural — the loss decomposition and the recency-weighting are the same in every substrate — but it is mixed rather than purely structural because the asset-placement decision is a human-practice category and the prime's risk-management origin gives it a faint normative charge that travels alongside the bare mechanism.

Examples¶

Formal/abstract¶

The "levee effect" in floodplain development is the origin instance and makes the loss decomposition explicit. The hazard is a major river flood with, say, a 100-year mean inter-arrival time and severe impact in the floodplain — the impact zone. A levee is built (the protective measure); it does not change the flood's physics but lowers the short-horizon frequency of inundation, so for decades the observed damage history in the protected zone is flat — zero floods. The value gradient is strong: protected floodplain land is flat, fertile, well-watered, and now apparently safe, so it is highly desirable for farms, then suburbs, then commercial development. The recency-weighted decision rule does the damage: each new developer and zoning board weights the recent flat history ("no flood in living memory behind this levee") far more heavily than the long-period flood distribution, so stake accumulates monotonically between shocks. Write expected loss as hazard-frequency × severity × stake-exposed. The levee nudged frequency down and left severity unchanged, but it drove the third term — stake — up by an order of magnitude, and stake is the non-stationary term. When a flood finally exceeds the levee's design (or the levee is overtopped), the loss is enormous, not because the hazard intensified but because the protected interval was used as an accumulation window. The prime's uncomfortable structural claim is exact: the levee is an upstream cause of the larger eventual loss — it performed as designed, and by doing so enabled the accumulation the next shock converted into catastrophe. The lever is to break the recency-weighting: price flood insurance against the long-period distribution and zone against the unprotected hazard, so placement decisions cannot be captured by the quiet period.

Mapped back: the 100-year flood is the rare severe hazard, the floodplain is the impact zone, the levee is the protective measure that flattens short-horizon frequency, fertile flat land is the value gradient, the "no flood in living memory" rule is the recency-weighted placement rule, and the order-of-magnitude growth in floodplain development is the monotonic stake accumulation.

Applied/industry¶

A long low-volatility regime in financial markets reproduces the identical structure in a capital-allocation substrate. The hazard is a severe market crash or liquidity freeze with a long, irregular inter-arrival time and severe impact when it hits. The impact zone is the set of correlated, illiquid, yield-chasing, or opaquely-leveraged positions that get destroyed in a crash. The protective measure is partly real and partly perceived — central-bank backstops, low realized volatility, risk models calibrated on the recent calm — all of which flatten the short-horizon loss history, so for years the observed drawdowns are small. The value gradient is the relentless search for yield: when safe assets pay little, capital flows toward the higher-returning positions inside the impact zone. The recency-weighted decision rule is the verbal form "this time is different" and the literal form of value-at-risk models that weight recent calm and systematically under-price tail risk, so leveraged exposure accumulates monotonically through the quiet years. When the crash arrives, the loss is many times the previous cycle's — driven by the grown stake, not by a novel hazard. The prime's diagnosis prevents the standard misreading: the bigger loss is not evidence that markets "got more dangerous," but that more value was placed where the hazard hits. The intervention is the same structural lever as the levee case: capital-adequacy and stress-testing rules keyed to the long-period distribution rather than recency-weighted volatility, so allocation cannot be captured by the calm. The same mechanism governs wildland–urban-interface fire losses (suppression flattens visible fire frequency while subdivisions push into fuel-loaded forest) and pandemic-supply concentration (quiet decades concentrate PPE manufacturing in one cluster).

Mapped back: the market crash is the rare severe hazard, correlated illiquid positions are the impact zone, central-bank backstops and calm-calibrated risk models are the protective measure, the yield search is the value gradient, "this time is different" and recency-weighted VaR are the recency-weighted placement rule, and the buildup of leveraged exposure is the monotonic stake accumulation — the same mechanism across flood geography, financial stability, and pandemic logistics.

Structural Tensions¶

T1 — Hazard Frequency versus Stake Exposed (scopal). Expected loss decomposes into hazard-frequency × severity × stake, and the prime insists stake is the non-stationary term. The failure mode is misreading a loss driven by accumulation as evidence of an intensifying hazard — "the world is getting more dangerous" when the truth is "we placed more value where it hits." Diagnostic: separate the hazard process from the stake process and ask which term moved; on the way out of a quiet period it is almost always stake, not frequency.

T2 — Protection as Good versus Protection as Upstream Cause (sign). The uncomfortable structural fact is that the protective measure which reduced short-horizon frequency is part of the cause of the larger eventual loss — the levee bought time that was used to accumulate stake. The failure mode is treating the defence as an unalloyed good and "the defence let us down," when it performed exactly as designed and enabled the accumulation. Diagnostic: ask whether the protection flattened the damage history and thereby opened an accumulation window — if so, it is an upstream driver of the next loss, not external to it.

T3 — Quiet Period as Safety versus Accumulation Window (temporal). The recency-weighted rule reads the flat history as evidence of safety, but the quiet period is precisely the window in which stake grows monotonically. The failure mode is naive extrapolation of a historical-loss series, which underestimates future loss because the stake is non-stationary while the series anchors to past stakes. Diagnostic: the longer the quiet, the larger the eventual loss — treat duration-since-last-shock as an accumulation clock, not a stability record.

T4 — Exposure Creep versus Adjacent Risk Patterns (scopal). The prime sharpens its boundary against three neighbors: it is not risk compensation (no behavior change — stock just arrives), not gradual deterioration (it adds healthy stock, not damages existing), and not systemic risk (it is the upstream mechanism that raises the stake any systemic event operates on). The failure mode is diagnosing one when another applies and prescribing the wrong fix. Diagnostic: ask whether new stock was placed (creep), behavior changed (compensation), existing stock degraded (deterioration), or failures propagated (systemic).

T5 — Recency-Weighting versus Distribution-Keyed Placement (sign/intervention). Creep stops only when asset-placement is keyed to the hazard's long-period distribution — disclosure, distribution-reflecting insurance pricing, zoning that ignores short-horizon quiet, capital rules against the long tail — rather than to recency-weighted history. The failure mode is intervening on the hazard or the protection while leaving the recency-weighted placement rule untouched, so accumulation resumes. Diagnostic: the structural lever is always to break the recency-weighting that feeds the accumulation — check whether placement decisions can still be captured by the quiet period.

T6 — Stake Measurement versus Loss-History Anchoring (measurement). Reasoning quantitatively requires modeling the stake-placement process and measuring current stake-at-risk, but the available data is the historical-loss series, which anchors expectations to past stakes. The failure mode is monitoring loss history and concluding stability while the stake silently grows beyond anything the series captured. Diagnostic: instrument the stake directly (assets in the impact zone now), not the loss record — the obvious extrapolation misleads precisely because it measures the wrong, stationary-looking term.

Structural–Framed Character¶

Exposure Creep sits just on the structural side of the middle of the structural–framed spectrum — mixed-structural, aggregate 0.3 — a genuine relational mechanism carrying a faint normative charge from its risk-management lineage and a binding to human asset-placement contexts. The mechanism itself is bare: expected loss decomposes as frequency × severity × stake, with stake the silently non-stationary term, driven up by a recency-weighted placement rule that reads a protected quiet period as safety.

Two diagnostics read zero and three carry half-points, distributed differently from the typical framed-prime pattern. vocab_travels is 0 because the working description — a rare hazard, an impact zone, a protective measure flattening short-horizon frequency, a value gradient, monotonic stake accumulation — needs no home lexicon to travel; the levee-effect language strips cleanly to a loss decomposition that finance, ecology, and infrastructure each state in their own terms. import_vs_recognize is also 0 because invoking the prime RECOGNIZES a stake-accumulation dynamic already present in the system — the assets really did pile into the impact zone whether or not anyone decomposed the loss — rather than IMPORTING an interpretive frame.

The three half-points capture the genuine frame. evaluative_weight (0.5) reflects the mild but real normative tinge: the risk-management lineage carries an implicit "exposure creep should be prevented," so the construct is not quite value-neutral the way "feedback" is. institutional_origin (0.5) reflects that its home and operating context are risk-management institutions — insurance pricing, zoning, capital-adequacy rules, catastrophe modeling — even though the underlying decomposition is not itself institutional. human_practice_bound (0.5) reflects that the core cases are human asset-placement decisions — developers behind levees, capital chasing yield, manufacturers concentrating supply — and while there is a biological analogue (a refugium accumulating biomass until a bad year), some natural analogues are weaker, so the pattern leans toward human and organizational substrates without being confined to them. Because the bare loss-decomposition mechanism is genuinely portable and value-neutral on two diagnostics, while the risk-management origin, normative charge, and asset-placement home supply the other three half-points, the aggregate lands at 0.3 — structural-side, mixed.

Substrate Independence¶

Exposure Creep is strongly substrate-independent — composite 4 / 5 on the substrate-independence scale. Its domain breadth is broad (4): the pattern in which a protective measure produces a quiet period that is then used as an accumulation window — so total exposure climbs even as visible frequency falls — recurs in floodplain and coastal development (the "levee effect" or safe-development paradox), the wildland–urban interface (fire suppression masking growing fuel loads while subdivisions push into the forest), finance (risk controls inviting larger positions), insurance, public health, and infrastructure. Its structural abstraction is high (4): the bare skeleton — a suppressor lowers an event rate, the lowered rate is read as safety, and assets-at-risk grow until one event clears a far larger accumulated stock — is medium-neutral, with the structural roles (protection, quiet window, accumulating exposure, eventual large release) traveling across settings. What holds the composite below the ceiling is that the strongest cases are framed by human asset-placement decisions: someone must choose to develop the protected zone, and the purest physical analogues are weaker than the socio-technical ones. Transfer evidence is the strongest component (5): the levee effect, the fire-suppression paradox, and risk-control-induced position growth are all concretely documented with the same accumulation-behind-a-suppressor mechanism. Within its range the prime is recognized rather than translated wherever protection silently licenses accumulation.

Composite substrate independence — 4 / 5
Domain breadth — 4 / 5
Structural abstraction — 4 / 5
Transfer evidence — 5 / 5

Relationships to Other Primes¶

Parents (1) — more general patterns this builds on

Exposure Creep presupposes Risk

Exposure creep operates on a pre-existing hazard exposure: it is the non-stationary growth of the STAKE term in expected-loss = frequency x severity x stake. Presupposes risk; the 0.81 'risk' neighbor is the genus, not a duplicate.

Path to root: Exposure Creep → Risk → Uncertainty

Neighborhood in Abstraction Space¶

Exposure Creep sits in a sparse region of abstraction space (86^th percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely rather than landing on a neighbor.

Family — Uncertainty, Risk & Proxy Distortion (22 primes)

Nearest neighbors

Computed from structural-signature embeddings · 2026-06-14

Not to Be Confused With¶

The embedding-nearest neighbor is risk, and the relationship is genus-to-species: exposure creep is a particular risk dynamic, not risk in general. Risk is the broad probability-weighted prospect of loss — it encompasses the hazard's frequency, its severity, and the stake exposed, all at once. Exposure creep makes a sharper, decomposed claim: that of the three multiplicative terms in expected loss, the stake term is silently non-stationary while frequency and severity appear flat or even favorable, and that this growth is driven by a recency-weighted placement rule reading the quiet period as safety. The decisive value-add over the generic "risk" frame is exactly this decomposition: a risk analysis that treats the historical-loss series as the object will conclude stability, because the series anchors to past stakes, whereas exposure creep tells you to instrument the current stake directly. Collapsing exposure creep back into "risk" loses the one thing it contributes — the identification of which term is moving and why naive extrapolation of loss history systematically under-predicts.

A second genuine confusion is with risk_return_tradeoff. Both involve more exposure preceding a possible large loss, and both arise in financial settings. But the tradeoff is a deliberate, priced acceptance of more risk in exchange for more expected reward — the agent knowingly moves up the risk-return frontier. Exposure creep is inadvertent: no one chooses the larger stake as a considered bet against a correctly-priced hazard; instead the stake accumulates because the protective measure flattened the damage history and the recency-weighted decision rule mistook quiet for safety. The discriminating question is whether the larger exposure was consciously taken on with the hazard distribution in view (tradeoff) or drifted into because the quiet period captured the placement decision (exposure creep). The fix differs sharply: a tradeoff problem is solved by re-pricing reward against risk; an exposure-creep problem is solved by breaking the recency-weighting so placement keys to the long-period distribution rather than the recent calm.

A third confusion worth drawing is with risk_pooling, because both manipulate exposure but in opposite directions. Risk pooling spreads exposure across many independent participants to reduce aggregate variance — the law of large numbers working in the system's favor. Exposure creep concentrates stake into a single impact zone, so the next shock hits a correlated stock all at once — variance working against the system. The tell is whether exposure is being diversified across uncorrelated endpoints (pooling) or massed into one hazard's reach (creep). Mistaking one for the other is consequential: a system that believes it is pooling ("we have many assets") when it has actually crept (those many assets all sit in the same floodplain, the same correlated trade, the same manufacturing cluster) carries a tail risk that pooling intuition entirely conceals.

For a practitioner the cuts route to different instruments. If the concern is the whole prospect of loss, that is risk — model all three terms. If the larger exposure was a priced, chosen bet, that is the risk-return tradeoff — re-examine the pricing. If exposure is genuinely spread across uncorrelated endpoints, that is pooling — and the variance reduction is real. Exposure creep specifically names the inadvertent, recency-driven concentration of stake in a known hazard's zone — diagnosed by decomposing loss and measuring current stake, fixed by keying placement to the long-period distribution.

Solution Archetypes¶

No catalogued solution archetypes reference this prime yet.