Project Enduring Pulse — restored long-horizon archive, website release 3, 14 September 2026.
ARCHIVAL SOURCE — READ THESE CORRECTIONS FIRST
The baseline CSVs rank PULSE first in 114 of 120 scenario/value comparisons and FOURTH in the other six. Statements saying second in source prose are incorrect. All six losses occur in scenario 19 (Weak institutions). The Scenario Atlas and Top-2 counts already reflect this vulnerability.
The restored v2 campaign is reproducible with the provided runner, not the original scripts’ defaults. Data are aggregate outputs, not individual trajectory logs. The older non-v2 baseline is retained for provenance; a sample rerun showed small discrepancies and that earlier campaign is not certified here.
2,100,000 reduced-form samples are randomized parameter draws with analytically accumulated hazards, not simulated Markov paths. 138,840,000 dynamic periods is a nominal loop count, not independent civilizations.
All parameters, scores and thresholds are model assumptions. Reproduction is internal computational verification, not empirical forecasting, external validation or proof of constitutional effectiveness. The v1.2 integration variable bundles effects and does not test the exact v1.2.1 non-coercion boundary. PULSE-Lite is a post-hoc hypothesis, not an adopted Charter.
External references in this historical source are not all verified; use the website claim register and citation audit for bounded current claims. In particular, its vague Paluck citation and numerical constitutional-lifespan claim are not endorsed by the website. Original source wording follows unchanged.

--- ORIGINAL SOURCE TEXT; TABLE CELLS IN READING ORDER ---
PROJECT ENDURING PULSE
Long-Horizon Civilization Simulation, Scenario Forecasting & Robust Decision Analysis
Project Enduring Pulse | Long-Horizon Scenario Simulation | September 2026
Methodological status: scenario simulation and robust-decision analysis. It is empirically anchored where human evidence exists, but it is not an empirical forecast of a future organic-synthetic civilization. Absolute probabilities are assumption-dependent. Comparative behavior, failure modes, tradeoffs, and breakpoints are the intended evidence.
1. Research question
The prior PULSE simulation program mostly asked whether individual constitutional mechanisms behaved as intended. This pass asks a broader question: across radically different plausible futures, does the full PULSE architecture tend to preserve a free, integrated, individually sovereign, recoverable civilization more robustly than credible alternative governance systems?
A second objective is adversarial: identify worlds in which PULSE loses, determine whether those losses arise from a specific doctrine or from implementation complexity, and estimate how strongly results depend on favorable assumptions about PULSE-specific mechanisms.
2. Why this is not an empirical forecast
No society of recognized conscious synthetic persons exists, so there is no empirical distribution for synthetic political preferences, reproduction, cognition, durability, merge behavior, or long-horizon economic choices.
Human evidence can constrain signs and plausible ranges—for example, contact generally reduces prejudice, perceived political voice strongly tracks institutional trust, wealth returns differ persistently, infrastructure dependencies cascade, and weak institutions amplify fragility—but it cannot supply future synthetic parameters.
The model therefore uses wide uncertainty envelopes and multiple archetypes rather than one “best estimate.” The robust claim is comparative: which architectures remain acceptable across many assumptions?
3. Empirical anchors
Domain
Human evidence
Use in model
Intergroup contact
713 samples / 515 studies; contact typically reduces prejudice
Sign and broad strength of Functional Integration feedback
Political voice / trust
OECD reports a 47-point trust gap by perceived political voice
Legitimacy sensitivity to voice/fairness
Constitutional endurance
Historical constitutions are often short-lived; design and flexibility matter
Non-zero institutional mortality and adaptability value
Power sharing
Literature distinguishes inclusive, dispersive, and constraining forms
Different effects for parity, federalism, checks, and rights
Wealth returns
Persistent return heterogeneity; returns correlate with wealth
Long-lived-capital concentration pressure
Automation
Productivity and displacement effects coexist
Synthetic productivity can raise output while creating distributional strain
Compute/energy
IEA projects rapid near-term data-centre electricity growth under multiple sensitivity cases
Resource/energy bottleneck scenarios
Infrastructure
CISA/NIST document physical, cyber, geographic and logical dependencies
Common-mode failure and resilience parameters
Fragility/conflict
Weak institutions and overlapping shocks produce persistent fragility and large economic losses
Compound-shock and capacity scenarios
Human demography
UN projects slowing growth/aging with large regional divergence
Organic demographic uncertainty envelope
4. Experimental design
20 baseline archetypes × 7 architectures × 8,000 dynamic worlds = 1.12 million 300-year architecture trajectories in the core comparison.
Each trajectory advances in 5-year steps for 60 periods. Common random seeds make competing architectures face comparable exogenous shocks in the core experiment.
Additional experiments include a 19-level PULSE-benefit skeptic sweep; five doctrine ablations plus full PULSE; institutional-capacity, homophily, and private-concentration sweeps; six compound extremes; PULSE-Lite exploratory tests; and 2.1 million trajectories in a separate reduced-form hazard model.
New-campaign total: 2.314 million dynamic architecture trajectories, 138.84 million dynamic architecture-period updates, plus 2.1 million reduced-form trajectories.
5. World archetypes
World
Initial synth share
Institutional capacity
Dominant stress profile
01 Gradual emergence
0.080
0.820
interplanetary load 0.05, illicit replication 0.10, disaster 0.12, external threat 0.12
02 Rapid synthetic growth
0.100
0.800
interplanetary load 0.08, disaster 0.12, external threat 0.14, illicit replication 0.22
03 Synthetic demographic majority
0.550
0.820
interplanetary load 0.10, disaster 0.12, illicit replication 0.14, external threat 0.15
04 Stable near parity
0.480
0.860
illicit replication 0.08, interplanetary load 0.08, disaster 0.10, external threat 0.12
05 Organic demographic contraction
0.280
0.820
interplanetary load 0.08, illicit replication 0.10, disaster 0.12, external threat 0.12
06 Cheap compute energy boom
0.200
0.860
scarcity 0.08, technology 0.90, disaster 0.10, external threat 0.10
07 Compute and energy scarcity
0.220
0.720
interplanetary load 0.05, illicit replication 0.12, disaster 0.18, external threat 0.20
08 Synthetic productivity leap
0.180
0.820
productivity asymmetry 0.90, disaster 0.10, interplanetary load 0.10, external threat 0.14
09 Durable-capital concentration
0.250
0.800
interplanetary load 0.08, wealth-return concentration 0.90, illicit replication 0.10, disaster 0.10
10 High functional integration
0.300
0.880
homophily 0.08, illicit replication 0.08, interplanetary load 0.08, disaster 0.10
11 Low integration and homophily
0.300
0.800
interplanetary load 0.08, homophily 0.90, illicit replication 0.10, disaster 0.10
12 Coordinated illicit replication
0.240
0.780
illicit replication 0.95, interplanetary load 0.08, disaster 0.12, external threat 0.18
13 Frequent cognitive attacks
0.320
0.780
cognitive attack 0.95, interplanetary load 0.08, disaster 0.14, illicit replication 0.18
14 Private infrastructure concentration
0.260
0.820
private concentration 0.95, interplanetary load 0.08, illicit replication 0.12, disaster 0.14
15 Repeated external conflict
0.300
0.760
external threat 0.95, interplanetary load 0.10, illicit replication 0.15, disaster 0.22
16 Species-targeted catastrophe
0.340
0.780
disaster 0.95, interplanetary load 0.08, illicit replication 0.12, wealth-return concentration 0.28
17 Common-mode catastrophe
0.300
0.740
interplanetary load 0.08, illicit replication 0.12, disaster 0.88, wealth-return concentration 0.28
18 Interplanetary expansion
0.340
0.820
interplanetary load 0.95, illicit replication 0.10, disaster 0.12, technology 0.86
19 Weak institutions
0.250
0.320
interplanetary load 0.08, disaster 0.18, illicit replication 0.20, external threat 0.28
20 Post-scarcity abundance
0.380
0.900
scarcity 0.02, technology 0.98, abundance 0.98, disaster 0.08
The complete normalized parameter matrix is supplied in scenario_archetype_parameters.csv. All archetype parameters are modeling inputs, not measured future quantities, and each is further randomized inside a broad distribution.
6. Governance architectures
Architecture
Modeled concept
PULSE
Five-doctrine architecture; protected constituencies; Functional Integration; cognitive sovereignty; resilience and mutual preservation; layered federalism.
Liberal constitutional democracy
Species-blind rights, checks, elections and adaptable institutions without strong permanent species-power sharing or mandatory functional integration.
Majoritarian democracy
Efficient one-person-one-vote majoritarianism with weaker minority/group vetoes and thinner anti-capture architecture.
Strict consociational parity
Strong parity/group veto architecture across ordinary government; high minority protection but high group salience and gridlock.
Species-separated federalism
High autonomy and separate institutions for species communities with a thinner shared layer.
Technocratic guardian governance
High expert/technical capacity, resilience and efficiency with weak electoral voice and concentrated guardian authority.
Minimal-rights market order
Strong negative rights, private ordering and innovation; weak public anti-domination/integration and life-essential dependency controls.
7. State variables and feedbacks
Population: organic/synthetic shares and illicit-origin synthetic share.
Social: Functional Integration, trust, species cleavage, homophily, political voice.
Rights: liberty, individuality/cognitive autonomy, anti-domination.
Political: legitimacy, constitutional health, gridlock/governance efficiency, authoritarian pressure.
Economic: prosperity, concentration/inequality, private dependency, productivity asymmetry.
Resilience: infrastructure resilience, restoration capacity, catastrophe recovery, external-defense pressure.
Exogenous shocks: economic disruption, cognitive attack, illicit replication, disaster, external conflict, species-targeted catastrophe, and common-mode failure.
The source code and full parameterization are included in the package. No hidden fitted coefficients were used.
8. Value-system robustness
value_system
liberty
individuality
integration
anti_domination
legitimacy
resilience
prosperity
governance_eff
trust
inequality
PULSE-aligned
0.170
0.170
0.120
0.150
0.080
0.130
0.070
0.050
0.030
0.030
Classical-liberal
0.280
0.180
0.030
0.140
0.080
0.080
0.100
0.050
0.020
0.040
Procedural-democratic
0.160
0.100
0.070
0.120
0.230
0.070
0.060
0.100
0.050
0.040
Capability-utilitarian
0.100
0.080
0.050
0.080
0.050
0.180
0.280
0.100
0.030
0.050
Stability-first
0.080
0.070
0.080
0.120
0.120
0.270
0.080
0.100
0.040
0.040
Individuality-first
0.180
0.300
0.100
0.140
0.050
0.080
0.040
0.030
0.040
0.040
Every value system independently penalizes catastrophic constitutional failure. The weighting exercise is intended to detect whether PULSE wins only when judged by PULSE’s own ideals.
9. Core results

Figure 1. Mean robust-decision value score across 20 archetypes and six value systems.
Architecture
Mean value
Worst combo
10th pct.
Max regret
Wins
Top-2
No catastrophic failure
Free & integrated
PULSE
0.579
0.159
0.495
0.080
114
114
0.627
0.568
Liberal constitutional democracy
0.525
0.222
0.444
0.102
6
120
0.631
0.285
Majoritarian democracy
0.440
0.202
0.375
0.201
0
0
0.535
0.075
Minimal-rights market order
0.426
0.205
0.356
0.215
0
6
0.497
0.078
Species-separated federalism
0.425
0.043
0.330
0.206
0
0
0.508
0.049
Technocratic guardian governance
0.400
0.130
0.338
0.232
0
0
0.341
0.058
Strict consociational parity
0.377
-0.002
0.283
0.255
0
0
0.450
0.139
PULSE ranked first in 114 of 120 scenario-value combinations and second in the remaining six. Those six are not six different failures: all are the weak-institutions archetype viewed through six different moral weightings. Liberal constitutional democracy won that world.

Figure 2. Raw modeled survival and free-integration are different objectives. Liberal democracy slightly exceeded PULSE on raw catastrophic-failure avoidance, while PULSE substantially exceeded it on free integration.
10. Failure-mode comparison
architecture
no_catastrophic_failure
free_integrated_rate
domination_rate
authoritarian_rate
bifurcation_rate
individuality_failure_rate
species_loss_unrecovered_rate
avg_liberty
avg_individuality
avg_integration
avg_anti_domination
avg_legitimacy
avg_resilience
avg_governance_eff
avg_prosperity
Liberal constitutional democracy
0.631
0.285
0.001
0.196
0.030
0.000
0.019
0.701
0.680
0.428
0.804
0.660
0.414
0.877
0.306
Majoritarian democracy
0.535
0.075
0.025
0.257
0.036
0.001
0.020
0.578
0.599
0.377
0.696
0.626
0.354
0.887
0.304
Minimal-rights market order
0.497
0.078
0.015
0.300
0.036
0.001
0.020
0.588
0.615
0.356
0.697
0.595
0.334
0.885
0.298
PULSE
0.627
0.568
0.000
0.185
0.028
0.000
0.009
0.725
0.760
0.594
0.855
0.646
0.610
0.804
0.299
Species-separated federalism
0.508
0.049
0.000
0.212
0.116
0.001
0.015
0.608
0.628
0.293
0.797
0.535
0.458
0.663
0.260
Strict consociational parity
0.450
0.139
0.000
0.196
0.081
0.001
0.014
0.586
0.601
0.339
0.764
0.476
0.400
0.514
0.232
Technocratic guardian governance
0.341
0.058
0.002
0.425
0.032
0.000
0.011
0.498
0.669
0.424
0.727
0.456
0.536
0.832
0.292
Domination is deliberately modeled as a persistent multi-period condition rather than a single election loss. The absolute rates are highly specification-sensitive; the relative pattern—higher risk under majoritarian/minimal-rights systems and lower risk under strong group/check architectures—is the useful inference.
11. Institutional capacity is PULSE’s clearest vulnerability

Figure 3. Full PULSE’s complexity becomes beneficial only after moderate institutional capacity is available.
The capacity sweep produced a crossover between approximately 0.50 and 0.55 on the model’s normalized state-capacity index. Below that range, simpler liberal constitutionalism performed better. This finding recurred in compound worlds combining weak institutions with scarcity/homophily and with interplanetary-frontier conditions.
This is consistent with the broader human evidence that weak institutions amplify fragility and that constitutional endurance depends partly on workable, adaptable institutions. It suggests a potential implementation principle: complexity should be staged rather than imposed faster than institutions can competently administer it.
12. Exploratory PULSE-Lite hypothesis

Figure 4. A deliberately simpler PULSE-Lite parameterization was tested only as a hypothesis for low-capacity contexts.
The PULSE-Lite experiment preserved strong rights, cognitive protections, integration, resilience, and federal adaptability while reducing constitutional complexity and ordinary veto costs. In the stylized model it beat both full PULSE and liberal democracy across the low-capacity sweep and in both low-capacity compound extremes. This result should not be read as validation of a specific PULSE-Lite constitution: its parameters were hand-designed after observing the capacity failure, so it is a hypothesis generator, not a clean out-of-sample test.
13. Functional Integration and social bifurcation

Figure 5. As voluntary homophily rises, Functional Integration materially extends the range over which a shared society remains plausible.
At moderate homophily, PULSE retained much higher free-integration rates than liberal, strict-parity, or species-separated models. At extreme homophily, every architecture failed. This is important: constitutional design can change incentives and create repeated common work; it cannot guarantee affection or erase powerful social preferences.
14. Skeptic sweep: how much self-favoring assumption can PULSE tolerate?

Figure 6. Multiplier applied to PULSE-specific benefits while competitor assumptions were unchanged.
On an average of all six value systems, PULSE crossed the liberal-democracy benchmark between roughly 0.80 and 0.85 of the baseline assumed institutional effect. Put differently, if PULSE-specific mechanisms deliver less than about four-fifths of their modeled benefit, the simpler liberal comparator catches or exceeds it. This is a useful anti-self-confirmation result: PULSE is not so advantaged in the model that its mechanisms can fail badly and still “win.”
15. Doctrine ablation

Figure 7. Each foundational doctrine was weakened in turn.
ablation
avgscore
no_catastrophic_failure
free_integrated_rate
bifurcation_rate
species_loss_unrecovered_rate
avg_liberty
avg_individuality
avg_integration
avg_resilience
Civilizational Resilience & Mutual Preservation
0.544
0.622
0.562
0.029
0.019
0.727
0.762
0.596
0.349
Equal Personhood & Sovereignty
0.493
0.599
0.000
0.032
0.010
0.470
0.650
0.581
0.604
Full PULSE
0.581
0.630
0.570
0.029
0.009
0.727
0.762
0.596
0.612
Functional Integration
0.543
0.618
0.167
0.064
0.010
0.723
0.735
0.373
0.573
Least Reversible Harm
0.566
0.620
0.562
0.028
0.010
0.723
0.757
0.592
0.552
Shared Continuity
0.551
0.597
0.520
0.031
0.009
0.675
0.752
0.560
0.602
The ablation is most useful directionally. Equal Personhood & Personal Sovereignty was the largest overall loss. Functional Integration was uniquely important to the shared-society outcome. Civilizational Resilience & Mutual Preservation drove catastrophe/recovery performance. Least Reversible Harm has the smallest numerical effect partly because the integrated annual model cannot fully represent one-off tragic-choice reasoning; this should not be interpreted as evidence that the doctrine is morally unimportant.
16. Compound extreme worlds
Compound world
Cross-value winner
PULSE score
Liberal score
PULSE no-failure
Liberal no-failure
X1 Scarcity + homophily + weak institutions
Liberal constitutional democracy
0.030
0.086
0.133
0.169
X2 Rapid synth boom + private concentration
PULSE
0.565
0.497
0.624
0.609
X3 War + cognitive attack + common-mode failure
PULSE
0.444
0.417
0.480
0.501
X4 Interplanetary frontier + weak institutions
Liberal constitutional democracy
0.192
0.222
0.213
0.245
X5 Species catastrophe during external war
PULSE
0.516
0.477
0.551
0.556
X6 Abundant high-tech hive temptation
PULSE
0.687
0.597
0.710
0.684
PULSE led four of six compound extreme worlds. Liberal constitutional democracy led the two worlds dominated by very weak institutional capacity. In the scarcity + homophily + weak-institutions perfect storm, every system performed poorly; the correct inference is not that liberalism solved the world, but that elaborate constitutional architecture cannot compensate for simultaneously weak state capacity, severe scarcity, and deep social separation. Full architecture-by-architecture compound results are in compound_extreme_scenarios.csv and the Scenario Atlas.
17. Reduced-form cross-check
architecture
no_catastrophic_failure
bifurcation_rate
reduced_form_quality
PULSE
0.817
0.195
0.715
Strict consociational parity
0.651
0.954
0.346
Species-separated federalism
0.608
0.984
0.309
Liberal constitutional democracy
0.483
0.654
0.282
Technocratic guardian governance
0.388
0.532
0.240
Minimal-rights market order
0.193
0.810
0.085
Majoritarian democracy
0.150
0.775
0.068
The reduced-form model changes the mathematics substantially: instead of evolving continuous social/economic state variables, it accumulates architecture-conditioned hazards. PULSE again ranked first, but the ordering and absolute event rates of other architectures changed. That divergence is evidence that absolute probabilities are model-dependent; it is also modest evidence that PULSE’s top ranking is not purely an artifact of one feedback equation.
18. What the model says about each alternative
Liberal constitutional democracy is PULSE’s strongest comparator. It is simpler, efficient, and robust under weak capacity, but it is less effective in the model at maintaining cross-species integration and protected mutual sovereignty over long horizons.
Majoritarian democracy remains efficient but incurs substantially more domination/bifurcation pressure when demographics or power become asymmetric.
Strict consociational parity protects groups but accumulates gridlock and group salience. It performs especially poorly on efficiency and long-run integration.
Species-separated federalism preserves autonomy and some resilience but tends toward exactly the parallel-civilization outcome PULSE is designed to avoid.
Technocratic guardian governance is strong on technical resilience/efficiency but consistently suffers legitimacy and authoritarian-capture penalties.
Minimal-rights market governance remains comparatively efficient and innovative, but weak public anti-domination/integration mechanisms expose it to inequality, exclusion, and legitimacy costs.
19. Results that should change how PULSE is reviewed
Institutional capacity should be treated as a precondition for the full complexity of the mature architecture. This does not imply changing the five doctrines; it suggests scrutinizing transitional and frontier implementation.
Functional Integration is not decorative. In the model it is one of the principal mechanisms preventing a formally unified state from becoming two civilizations.
PULSE’s resilience advantage matters most when infrastructure and restoration mechanisms actually have independent failure domains. The prior micro-model conclusion remains important.
No architecture succeeds under near-total social homophily plus weak institutions plus severe scarcity. Cultural cooperation and competent administration are irreducible inputs, not outputs that constitutional text can guarantee.
PULSE’s benefit is not “more survival at any cost.” The strongest modeled advantage is preserving survival together with individuality, liberty, anti-domination, and integration. Its costs are complexity and some governance/economic efficiency.
20. Major limitations and bias risks
No empirical synthetic-citizen dataset exists. Synthetic behavioral parameters are scenario assumptions.
The same research process that helped design PULSE built the model. Six alternative value functions, deliberately strong comparators, a skeptic sweep, compound extremes, and a different mathematical model reduce but do not eliminate design bias.
The model does not resolve consciousness science, detailed criminal justice, exact labor markets, family formation, culture, religion, or future scarcity technology.
The private-sovereignty outcome is less well quantified than other outcomes; the current model is better at showing concentration pressure and legitimacy effects than at forecasting a crisp threshold where private dependence becomes sovereignty.
Constitutional rules are represented as effectiveness parameters. Real institutions can be captured, ignored, reinterpreted, or administered incompetently in ways no compact parameter can fully capture.
Long-horizon compounding makes small coefficient choices matter. Breakpoints and comparative patterns are more credible than exact 300-year probabilities.
21. Recommended next use of these results
Do not treat the simulation as a mandate to rewrite v1.2. Use it to focus external review. The most valuable questions for a constitutional scholar or systems engineer are now concrete: how should a doctrine-rich architecture stage itself under low administrative capacity; what minimum institutional competence is required before full parity/complexity; how does Functional Integration avoid coercive overreach while remaining effective; and how can resilience obligations be audited without freezing technology?
References and empirical anchors
1. Pettigrew & Tropp (2006), intergroup contact meta-analysis. https://pubmed.ncbi.nlm.nih.gov/16737372/
2. OECD Trust Survey 2026, political voice and institutional trust. https://www.oecd.org/en/publications/oecd-survey-on-drivers-of-trust-in-public-institutions-2026-results_9eb63fec-en.html
3. Elkins, Ginsburg & Melton, The Endurance of National Constitutions. https://www.cambridge.org/core/books/the-endurance-of-national-constitutions/CD671879ADBFF7420C25C3707E6135F0
4. Andeweg (2000), Consociational Democracy. https://www.annualreviews.org/content/journals/10.1146/annurev.polisci.3.1.509
5. Safeguarding Democracy: Powersharing and Democratic Survival. https://www.cambridge.org/core/journals/american-political-science-review/article/abs/safeguarding-democracy-powersharing-and-democratic-survival/D31FFD5C1EE8DF14FA2C9BB8E73536EB
6. World Bank, fragile/conflict-affected situations. https://www.worldbank.org/en/research/publication/fragile-and-conflict-affected-situations-vulnerabilities
7. Fagereng et al., Heterogeneity and Persistence in Returns to Wealth. https://www.nber.org/papers/w22822
8. Daminato & Pistaferri, Returns Heterogeneity and Consumption Inequality. https://www.nber.org/papers/w32490
9. Acemoglu & Restrepo, Robots and Jobs. https://www.nber.org/papers/w23285
10. ILO, Generative AI and Jobs: 2025 Update. https://www.ilo.org/publications/generative-ai-and-jobs-2025-update
11. IEA, Energy and AI. https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai
12. UN DESA, World Population Prospects 2024. https://desapublications.un.org/publications/world-population-prospects-2024-summary-results
13. CISA Infrastructure Dependency Primer. https://www.cisa.gov/topics/critical-infrastructure-security-and-resilience/resilience-services/infrastructure-dependency-primer/learn
14. NIST Community Resilience Planning Guide. https://www.nist.gov/community-resilience/planning-guide
Appendix A — Experiment counts
Dynamic experiment
Architecture trajectories
5-year state transitions
Core 20-world × 7-architecture dynamic comparison
1120000
67200000
PULSE skeptic sweep
304000
18240000
Five-doctrine ablation + full PULSE
144000
8640000
Institutional-capacity sweep
128000
7680000
Homophily/integration sweep
120000
7200000
Private-concentration sweep
120000
7200000
Six compound extreme worlds
210000
12600000
PULSE-Lite capacity exploration
132000
7920000
PULSE-Lite compound extremes
36000
2160000
Dynamic total: 2,314,000 architecture trajectories and 138,840,000 five-year state transitions. Reduced-form Markov cross-check: 2,100,000 additional architecture trajectories. These counts are separate from the earlier PULSE constitutional/micro-model suite.
Appendix B — Files and reproducibility
The accompanying archive contains the complete Python source, architecture/scenario/value parameter CSVs, raw aggregate outputs for every experiment, empirical-source notes, model codebook, figures, and the unmodified Project Enduring Pulse v1.2 reference documents. The integrated model is deterministic given its seed and parameters.