🔬 applied physics

From transient shocks to unexpected outcomes: disruptive drivers in scenario pathways

This paper extends the Cross-Impact Balance (CIB) method to better analyze scenario pathways by introducing four distinct run types that track disequilibrium and structural uncertainty, thereby enabling more robust stress-testing and the exploration of rare or surprising outcomes in socio-technical transitions like decarbonization.

Original authors: Andrew G. Ross

Published 2026-04-24

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Andrew G. Ross

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to plan a long road trip across a continent. You have a map (your model) and a destination (a clean energy future).

Most traditional planning tools give you one main route or perhaps a wide band showing the "best case" and "worst case" scenarios. It's like looking at a weather forecast that says, "It will be between 60 and 80 degrees." That's helpful, but it doesn't tell you why the temperature might spike, or if a sudden storm could knock your car off the road, or if your map itself might be wrong.

This paper, written by Andrew G. Ross, argues that for big, complex challenges like the energy transition, we need a better way to look at the future. Instead of just one line on a map, we need to simulate four different types of road trips to see how our plan holds up.

Here is the simple breakdown of the paper's ideas, using the analogy of a giant, interactive board game where the pieces influence each other.

The Game Board: Cross-Impact Balance (CIB)

First, the paper uses a method called Cross-Impact Balance (CIB). Think of this as a giant board game with 15 different pieces (like "Solar Power," "Government Policy," "Public Opinion," and "Oil Prices").

In this game, every piece has a rulebook on how it affects the others.

If "Public Opinion" goes up, "Government Policy" might go up.
If "Oil Prices" crash, "Investment in Renewables" might drop.

The goal is to see how the game plays out over time (from 2025 to 2050). Usually, players just run the game once to see the "most likely" outcome. Ross says that's not enough. We need to stress-test the game in four specific ways.

The Four Ways to Play (The Four Scenarios)

Ross proposes running the game four different times, each time changing the rules slightly to see what happens.

1. The "Sudden Pothole" Test (Disequilibrium)

The Analogy: Imagine you are driving smoothly, and suddenly, a massive pothole hits your car in the year 2030. You swerve, the car shakes, and you might even spin out for a moment before you get back on track.
The Science: This tests transient shocks. It asks: What happens if one specific year gets hit by a huge, one-time disaster (like a pandemic or a war)?
The Insight: It measures how "wobbly" the system gets. Does the car recover quickly, or does it crash? This helps us see how resilient our plan is to sudden, short-term shocks.

2. The "Parallel Universes" Test (Extremes/Regimes)

The Analogy: Imagine you are playing the board game, but you decide to play three different versions of the game at the same time:

Universe A (The Dream): Everyone loves green energy, and politicians are super supportive.
Universe B (The Nightmare): Everyone hates change, and oil companies block everything.
Universe C (The Middle): Things are just okay.
The Science: This tests different storylines or "regimes." Instead of just one set of rules, we change the entire "physics" of the game to match a specific story.
The Insight: This helps us see which parts of our plan are robust. If our plan works in the "Nightmare" universe, it's a strong plan. If it only works in the "Dream" universe, it's fragile.

3. The "Foggy Map" Test (Unexpected Outcomes)

The Analogy: Imagine you are playing the game, but you realize your rulebook might be wrong. Maybe you don't actually know how "Public Opinion" affects "Policy." So, in this version of the game, you randomly rewrite the rulebook every single turn. Sometimes the rules are logical; sometimes they are weird and surprising.
The Science: This tests structural uncertainty. It acknowledges that we don't fully understand how the world works. By sampling thousands of different "rulebooks," we can find rare, surprising outcomes (like a "Black Swan" event) that a standard map would never show.
The Insight: This reveals the "unknown unknowns." It shows us the weird, unlikely futures that could happen if our understanding of the world is slightly off.

4. The "Static Noise" Test (Exogenous Shocks)

The Analogy: This is the control group. Imagine driving on a normal road, but there is just a little bit of static noise in the radio and a few small bumps. Nothing huge changes the rules; nothing changes the map.
The Science: This is the baseline. It runs the game with just random, small disturbances to see what happens when there are no big shocks, no alternate universes, and no rule changes.
The Insight: It gives us a "normal" baseline to compare the other three tests against. It helps us separate "real surprises" from just "random noise."

Why Does This Matter?

The author argues that if we only look at the "average" road trip, we are blind to the dangers.

The "Band" Problem: Traditional models show a range (e.g., "We will reduce emissions by 40% to 60%"). But they don't tell you why it might be 40% or 60%. Is it because of a war? Because people hate solar panels? Because our math was wrong?
The Solution: By running these four types of simulations, policymakers can say:
- "Our plan is robust because it works even in the 'Nightmare' universe."
- "We need a backup plan because if a 'Sudden Pothole' hits in 2030, we will crash."
- "We are overconfident because when we mess with the rulebook (Foggy Map), we find a 5% chance of total failure."

The Bottom Line

This paper is a toolkit for stress-testing the future. It moves us from asking "What is the most likely future?" to asking:

How do we handle a sudden shock?
Does our plan work in a bad world?
What weird surprises are hiding in our blind spots?
How much of this is just random noise?

By using these four lenses, we can build energy plans that aren't just "likely" to work, but are ready for anything.

1. Problem Statement

Current scenario and pathway analysis (e.g., for energy transitions) often relies on a single central trajectory or a narrow band of outcomes (min-max ranges). This approach suffers from three critical limitations:

Lack of Disequilibrium Quantification: It fails to measure short-lived departures from a stable, self-consistent state along a pathway, obscuring the system's resilience or vulnerability to transient shocks.
Conflation of Uncertainty Sources: It does not distinguish between variation caused by a one-off shock, a different underlying narrative (storyline/regime), or structural uncertainty about how factors interact.
Blindness to Rare Outcomes: It often overlooks "grey swan" or "black swan" events that are plausible under structural uncertainty but unlikely under a single central matrix.

Decision-makers need to understand not just how much outcomes vary, but why they differ and under what specific stress or uncertainty conditions they arise. Existing frameworks lack a unified method to separate these influences.

2. Methodology

The paper extends the Cross-Impact Balance (CIB) method, a well-established technique for generating multi-period scenario pathways based on a cross-impact matrix (CIM) of descriptors. The author introduces a systematic framework implemented in the open-source package PyCIB, defining four distinct scenario run types to isolate different drivers of variation.

Core Extensions to CIB

The paper formalizes three new dimensions:

Disequilibrium Metrics:
- Consistency Margin ( $\mu$ ): Measures the gap between the chosen state and the best alternative for a descriptor. A negative margin indicates inconsistency.
- Hamming Distance: Calculates the distance between the current scenario and the nearest attractor (fixed point or cycle) or the consistent set.
- Cumulative Disequilibrium Burden ( $B$ ): Sums the negative consistency margins over time to quantify the total transient stress along a pathway.
- Partial Convergence: Instead of running succession to full equilibrium, the model caps iterations ( $K_{max}$ ) to simulate transient states where the system has not yet settled.
Regime-Aware Extremes:
- Pathways are run under distinct Regimes (named structural modes), each with its own CIM.
- This allows for the comparison of outcomes across different "worlds" (e.g., Baseline vs. Stress vs. Moderate) where the underlying influence structure is fixed for the duration of the run.
Structural Uncertainty Sampling:
- Instead of treating the CIM as a known constant, the influence table is sampled from an uncertain specification.
- Cell values are drawn from distributions (e.g., Normal) centered on expert point estimates, with variance scaling over time ( $\lambda(t)$ ) to widen uncertainty.
- This exposes "structural-uncertainty tails"—outcomes that are rare under a central matrix but plausible when the influence structure itself is uncertain.

The Four Scenario Types

The framework organizes these extensions into four distinct run types:

Disequilibrium Runs: A single nominal CIM with a one-off amplified structural shock in a specific period. Used to stress-test resilience and quantify transient disequilibrium.
Extremes Runs: Pathways run under fixed, distinct regimes (e.g., "Stress" vs. "Baseline" CIMs) combined with regime-specific shock scaling. Used to compare robustness across narratives.
Unexpected Outcomes Runs: Pathways where the CIM is re-sampled per period from an uncertain distribution, combined with dynamic and structural shocks. Used to identify rare, surprising endpoints (structural-uncertainty tails).
Exogenous Runs: A baseline using a single nominal CIM with only stochastic structural and dynamic shocks (no regime change or CIM sampling). Serves as a control to isolate pure noise effects.

3. Key Contributions

Formalization of Disequilibrium: Provides the first formal metrics (margin, distance, burden) to quantify how far a pathway is from a stable state at each step, moving beyond static consistency checks.
Attribution of Variation: Creates a framework that qualitatively separates the causes of outcome variation into: transient shocks, narrative/regime differences, and structural uncertainty.
Unified Framework: Integrates these three extensions with standard exogenous shocks into a single CIB pathway framework, allowing analysts to toggle specific mechanisms on or off.
Tail Risk Identification: Distinguishes between "realisation tails" (extreme shock sequences), "narrative tails" (adverse storylines), and "structural-uncertainty tails" (plausible but neglected futures due to unknown interactions).

4. Results (Empirical Illustration)

The approach was applied to a socio-technical decarbonisation pathway (15 descriptors, 3 states each) using data from Ross and Ross (2026). The simulation ran 500 pathways for each of the four types over a 25-year horizon (2025–2050).

Structural Sensitivity: Figure 1 (in the paper) visualizes where the influence structure varies most across the four types, highlighting specific descriptors (e.g., Grid Development, Investment Availability) that drive divergence.
Outcome Distributions:
- Disequilibrium: Showed a transient spike in response to an amplified 2030 event, with a mean of 43% of periods being inconsistent (transiently unstable).
- Extremes: The "Low," "Medium," and "High" regime ensembles showed varying degrees of spread. When lines diverged, outcomes were sensitive to the combined narrative and shock intensity; when close, outcomes were robust.
- Unexpected Outcomes: This workflow produced the widest distribution. 2.8% of pathways ended in "High emissions," a state classified as "unexpected" (rare in the ensemble but plausible under structural uncertainty). This highlights endpoints a central matrix would underweight.
- Exogenous: Served as a baseline, showing that without structural uncertainty or regime changes, the system remained largely anchored to the "Net-zero aligned" attractor.
Interpretation: The results demonstrate that while early dynamics are often aligned (due to shared initial conditions and baseline attractors), the scenario-specific mechanisms determine the extent of spread and tail outcomes.

5. Significance and Implications

Policy Transparency: The framework moves beyond a single "min-max" band, allowing policymakers to distinguish between risks that are robust across narratives and those that depend on specific structural assumptions.
Stress-Testing: Explicitly quantifies the "burden" of transient shocks, helping identify where buffering or contingency planning is most needed.
Epistemic Risk Management: By sampling uncertain influence structures, the method addresses "unknown unknowns" regarding how factors interact, reducing overconfidence in a single central view.
Methodological Rigor: The paper clarifies that dynamic CIB outputs should be read as the propagation of expert-structured judgments rather than validated state-space models, improving the transparency of what the model claims to represent.

In conclusion, this paper provides a robust, generalizable toolkit for exploring the full spectrum of future pathways, from transient shocks to structural surprises, enabling more resilient and informed decision-making in complex socio-technical transitions.