BDPD — Be Different Play Differential

A computational laboratory for the Seneca Effect, the Tragedy of the Commons, and the politics of collapse

Roberto Brunelli

2026-06-23

“He cut down the cedar, he conquered the forest of cedars.” — Epic of Gilgamesh, Tablet V

A computational laboratory for the Seneca Effect, the Tragedy of the Commons, and the politics of collapse — through simulation and a physical card game.

Roberto Brunelli · 2026

The Cliff

“Increases are of sluggish growth, but the way to ruin is rapid.” — Seneca, Lucilium 91

Socio-ecological systems grow slowly but collapse rapidly — the Seneca Effect.
Resource extraction accelerates through capital accumulation while pollution feeds back with a fatal delay.

Productivity peaks one instant before the cliff.
By the time the decline is visible, recovery is already impossible.

Bardi’s canonical three-variable Seneca model: Resources decline monotonically; Capital rises, peaks, then crashes; Pollution ignites late and outlasts everything. The asymmetry ratio is 0.47 — collapse is twice as fast as growth.

The Gap

“It is the obvious which is so difficult to see most of the time.” — Isaac Asimov, Foundation

Seneca Effect
Slow growth, fast collapse.
Irreversible tipping points.

Tragedy of the Commons
Individual rationality vs. collective ruin.

Wealth Inequality
U-shaped relationship with sustainability (Dayton-Johnson).

Three phenomena
— the Seneca Effect, the Tragedy of the Commons, and wealth inequality —
have been studied extensively, but almost never together.

What happens when a single aggressive actor operates inside a system with asymmetric temporal dynamics?

Does adaptation help or create exploitable vacuums? Does transparency prevent free-riding, or enable it?

BDPD addresses this gap — not as a model making claims about any specific real-world system, but as a configurable laboratory where the three phenomena can be studied in controlled interaction, replicated, and extended by others.

The Arena

“All models are wrong, but some are useful.” — George Box

A configurable multi-agent laboratory in Node.js. Pluggable engine: logistic commons (hidden reserve + stochastic collapse) or Bardi R/C/P ODE (Seneca Effect, three-variable). Three schedulers (simultaneous, sequential, wealth-weighted). Configurable observation noise and Knightian uncertainty about the collapse threshold via a hidden reserve. Perturbation engine for exogenous shocks and strategy overrides. Agents connect via HTTP — heuristic bots, sandboxed code, or LLMs.

The Arena produces statistical regularities — robust patterns across thousands of runs. But it produces no intuition. That is what the card game is for.

The BDPD Arena dashboard: real-time trajectory of commons stock, per-agent wealth, Gini coefficient, and welfare score. Four agents compete for a shared resource.

The Forest of Humbaba

“Humbaba’s cry is the Flood, his mouth is Fire, his breath is Death.” — Epic of Gilgamesh, Tablet V

A physical card game that translates every platform mechanism into a tangible object:

Forest Deck / commons stock,
Box Reserve / hidden depth,
Forest Die / stochastic collapse,
Seneca Clip / irreversible cliff

Four archetypal decks embody the strategies — and the same phenomena emerge on a fundamentally different substrate.

The card game produces intuition — the felt sense of a system tipping, of a conservator subsidising a defector. But it produces no statistics. Neither instrument alone is sufficient. Together, they form a dual-mode laboratory: findings on one substrate are tested for robustness on the other.

The game is also a pedagogical tool: by making the strategic structure visible, playable, and teachable, it makes complex system dynamics accessible in classroom and workshop settings.

Cedar of Uruk (Warrior-King): pure aggression. High harvest, immediate forest damage, no conditions

Temple Gardens (Temple Keeper): the conservator’s balance. A small personal gain that heals the commons

A Dual-Mode Laboratory

“The map is not the territory — but the right map reveals the territory’s traps.”

BDPD is proposed as an open-source experimental and didactic testbed — an infrastructure for the research community, not a predictive model with universal claims.

The four findings reported below are offered as demonstrations of the laboratory’s capacity to surface counterintuitive dynamics and structural fragilities that traditional isolated models often overlook.

They are starting points: illustrations of what the testbed can do, and of the hypotheses it can generate.

All code, card assets, and raw data are open-source (MIT licence).
The game is print-and-play.
The questions are open.

Platform → statistics
Sweeps across thousands of runs reveal thresholds, slopes, and boundary conditions.

Card game → intuition Discrete, stochastic play makes the strategic structure felt — and structurally contrasts whether platform patterns survive a fundamentally different substrate.

Together → dual-mode science
Convergence across substrates increases confidence.
Divergence reveals hidden assumptions worth resolving analytically.

One Is Enough

“He who controls the spice controls the universe.” — Frank Herbert, Dune

The collapse threshold is absolute within the canonical CPR regime — not relative to pool size or composition. A single aggressive agent dooms the commons
regardless of pool size, initial endowments, or turn order. No amount of conservative co-players can compensate.
(Rescue requires regeneration r \(\approx\) 0.95, roughly 8× above the experimental default regeneration rate (r = 0.12); Supplement B3.)

Higher regeneration within the canonical range doesn’t help — it only buys a slower collapse.

In the card game, stochasticity softens the binary into a probability,
but the Warrior-King’s 81.5% collapse rate in mirror matches confirms the pattern.

Gate pass rate heatmap: regen rate × aggressive fraction. The only green region is zero aggressive agents. One is enough.

0%

gate pass rate with 1 aggressive agent, across pool sizes 2–20, regen rates r \(\leq\) 0.32, all schedulers, all endowments. (Rescue at r \(\approx\) 0.95, roughly 8× the experimental default (r = 0.12).)

Smart \(\neq\) Good

“To know a thing well, know its limits. Only when pushed beyond its tolerances will its true nature be seen.” — Frank Herbert, Heretics of Dune

The individually rational strategy — observe the stock, reduce harvest when it falls — is collectively worse than unconditional moderation.

Reactive agents create a strategic vacuum: by pulling back, they cede extraction capacity to the aggressive player, who immediately fills the gap.

The smarter the response, the larger the subsidy to the predator.

Welfare score: conservative-only pools consistently outperform reactive-only pools across all regen rates, despite reactive agents being individually “smarter”.

Reactive < Conservative

The “smart” agent sees the cliff coming, steps back — and hands the advantage to the one agent that never steps back. Reactive governance can be worse than no governance at all.

The Fog Paradox

“Knowing where the trap is — that’s the first step in evading it.” — Frank Herbert, Dune

In the Arena, observation noise on the commons stock improves welfare in two regimes — a modest trend across the noisy range, and a discontinuous further gain under full opacity. Noise blunts the aggressive agent’s fine-grained exploitation; under full opacity it removes the stock signal entirely. Conservators don’t depend on the signal in either regime.

In the card game, noise has zero effect — heuristic archetypes do not condition on exact stock numbers.

Information fidelity only matters when agents have the cognitive architecture to exploit it. The fog is irrelevant to the blind.

Arena (P9): a modest welfare trend across the noisy range plus a discontinuous further gain at full occlusion, which selectively disables the aggressive agent’s signal advantage.

Card game (CT3): collapse rate, Gini, and stockpile are statistically flat from 0% to 100% noise. The fog makes no difference.

Safe Betrayal

“It pays to be obvious, especially if you have a reputation for subtlety.” — Isaac Asimov, Foundation

Against a conservator, defection is not only safe — it is rewarded.

The Stranger-King gains up to 27% more wealth than its cooperative baseline without triggering collapse. Forced defection even lowers the collapse rate below the no-defection baseline.

The conservator’s healing creates the ecological slack that the defector captures as private profit.

CT4: Stranger-King vs. Temple Keeper. Stockpile declines monotonically from 20.90 (T=3) to 18.18 (T=11), always above the no-defection baseline of 16.51. Collapse rate stays below 10%.

+27%

personal wealth gain from defection, with no increase in commons collapse. The system survives. The cooperator pays. The traitor profits.

The Tragedy of the Compensator

“He who has gone before has made the way easier for those who follow.” — Gilgamesh, Tablet X

When a coalition implements strict emissions caps, they marginally reduce the atmospheric carbon load.

If this buffer allows others to continue cheap extraction
without triggering the climate cliff,
the cooperators have subsidised the defectors’ profits.
The system survives, but the wealth distribution becomes increasingly skewed.

Unilateral conservation without exclusion mechanisms is not just insufficient
it is structurally counter-productive.
It transforms the commons from a shared resource into a subsidised extraction ground.

Compliance is punished,
defection is rewarded

… and the very resilience that conservation builds becomes the defector’s most reliable asset.

The interaction between strategic irreversibility and the tragedy of the commons is more dangerous than either phenomenon alone: the step-collapse threshold and strategic pre-depletion ensure that once extraction exceeds a critical level, recovery is foreclosed before it can be organised; the tragedy ensures that at least one agent will push past that level. Institutional design must focus not on optimising average behaviour, but on constraining worst-case actors.

What’s Next

“You can prove anything you want by coldly logical reason — if you pick the proper postulates.” — Isaac Asimov, Foundation

The framework is open and extensible. Current and planned extensions include:

LLM Agents \(\checkmark\) (preliminary, single-model case study) DeepSeek v4-flash — emergent safe betrayal observed. Awareness-without-restraint pattern observed. Multi-model sweeps planned.

Bardi R/C/P Extension
\(\checkmark\) Three-variable Seneca model integrated as engine mode.
Systematic sweeps planned.

Meta-Agent & Human Experiments
Policy-maker with dynamic harvest caps, taxes, and monitoring noise.
Empirical validation with real players at the table.

Open Source
Configure your own environments. Test your own agent architectures. gitlab.com/bdpd/bdpd

“Only the gods live forever. As for us men, our days are numbered. Our occupations are a breath of wind.” — Epic of Gilgamesh, Tablet III

“The platform produces statistics. The card game produces intuition.
The questions are open. We invite others to use it.”

Open-source platform · Print-and-play card game · Preprint forthcoming

gitlab.com/bdpd/bdpd · Roberto Brunelli · v0.5.1 · 2026

Authored by Roberto Brunelli with the support of frontier large language models; model outputs were source-grounded by prompt, validated by extensive triaging with different models and finally integrated under author supervision.