If consumers need 10 wood per tick and producers only harvest 8, the economy starves within 20 ticks. AgentE continuously monitors the ratio of consumer demand to producer output across every resource type, flagging any sustained imbalance before inventory buffers run dry. This is the most fundamental check — without production-consumption parity, nothing else matters.

Raw materials appearing in finished-goods markets create noise that corrupts price signals and confuses participants. Each supply chain stage must hand off directly to the next, enforcing clear boundaries between extraction, refinement, manufacturing, and retail. When intermediaries are skipped, the economy loses its ability to price goods accurately at each tier.

A new producer who can’t afford inputs on their first tick is a bootstrap failure — they exist in the economy but can’t participate. Starting capital must cover at least one full production cycle, including input costs, tool fees, and any listing charges. Without this, new entrants are stillborn, and the economy’s ability to replace retiring participants collapses.

If crafters can produce every 3 ticks but materials arrive every 5, they idle 40% of the time. Input delivery rate must always exceed production rate, or the entire downstream chain starves while producers sit around waiting. AgentE monitors delivery-to-production ratios and flags bottlenecks before they cascade into full supply chain breakdowns.

Most player trades are liquidating unwanted surplus, not selling deliberate production. Price signals from disposal trades are fundamentally weaker demand indicators than production-for-sale trades — a player dumping 500 iron ore they don’t need sends a different signal than a miner selling 500 iron ore they specifically produced. AgentE weights these trade types accordingly when computing economic indicators.

Profit formulas that ignore population cause stampedes. If one role pays a flat rate regardless of how many participants share it, the entire economy piles in — destroying margins for everyone and abandoning critical roles elsewhere. Profitability must always be calculated relative to the number of competitors in that role, so incentives naturally balance across the economy.

Any role without inverse-population scaling is a herd vulnerability. One unprotected role becomes the black hole that absorbs the entire economy — if mining pays the same whether 10 or 10,000 people do it, everyone will mine and nobody will craft. Every single role needs a crowding multiplier that reduces individual returns as participation increases.

When specialists exceed 70% of the population, the pot is at sustainability risk. The generalist majority funds the specialist minority through broader economic activity — taxes, fees, and consumption. Without enough generalists generating baseline economic flow, specialist reward pools starve and the entire incentive structure collapses.

If a role dominates above 30%, classify as structural (in dominantRoles config) or pathological. Structural dominance gets no intervention. Pathological dominance gets crowding pressure. This principle absorbs the old P28 — the classification step (structural vs pathological) is now unified here, preventing AgentE from fighting design intent.

Without switching costs, agents oscillate between roles every tick and never reach equilibrium. If more than 5% of the population switches roles in a single period, it signals herd movement — everyone chasing the same opportunity simultaneously. Friction (cooldowns, retraining costs, reputation loss) dampens oscillation and lets markets settle into stable distributions.

When new agents enter the economy, they should fill underserved roles — not pile into whatever’s already popular. Spawn probability must be inversely proportional to current role count, so underpopulated roles naturally attract new entrants. Without this, imbalances compound with every new participant.

Corrections applied only at hard thresholds create delayed, jerky responses. Instead, apply light continuous pressure at all times (gentle nudges toward equilibrium) and strong pressure only at critical deviations (emergency interventions). This creates smooth, invisible corrections that players rarely notice, rather than sudden shocks that feel punitive.

A healthy economy needs Gamers, Traders, Collectors, Earners, and Builders coexisting. When any single persona type exceeds 40% of the population, the economy enters behavioral monoculture — it becomes fragile because everyone responds to shocks the same way. AgentE monitors persona distribution and adjusts incentives to maintain at least 3 distinct behavioral clusters.

Multiple independent currency sources create uncontrollable inflation risk. Just like a central bank controls money supply through one primary mechanism, a virtual economy needs one master dial that controls total currency inflow. Secondary faucets can exist, but they must be downstream of the primary control, not independent sources.

Every prize pool must be mathematically sustainable before it opens. If winRate × multiplier exceeds (1 - houseCut), the pot drains to zero — it’s just a matter of time. AgentE validates pot sustainability formulas before activation and monitors drain rates in real time, flagging any pool trending toward insolvency.

Resource gathering (collecting wood, mining ore) is not the same as currency injection (earning gold). Confusing the two creates accounting errors that make monetary policy impossible. AgentE distinguishes between resource creation, resource transformation, and actual currency injection, tracking each separately to maintain accurate economic models.

Any pool (bank, reward pool) without a cap accumulates infinitely. A pool at 42% of total supply means 42% of the economy is frozen. Cap configurable (default 10%). Violation fires at 2× cap share of total supply. Without caps and decay, pools become economic black holes that drain the circulating supply.

Depleted pool with estimated high staking signals early-withdrawal abuse. This is a symptom detector — the penalty formula is developer responsibility. AgentE flags the conditions; you build the mechanic.

Money supply is meaningless if it’s not moving. An economy can have abundant currency but feel dead if everyone is hoarding. AgentE tracks velocity (transactions per total supply) as a primary health metric — low velocity despite adequate supply signals stagnation, which requires different interventions than actual scarcity.

In barter-heavy economies, no single commodity naturally stabilizes as the universal unit of account. Multiple items rotate as de facto currency, but none locks in permanently. If a standard unit of account is needed for economic stability, it must be explicitly designed and enforced — emergence alone won’t produce one reliably.

Any intervention before tick 30 is premature. The economy needs time to bootstrap with designed distributions. Early intervention against designed dominance can kill the economy instantly — you would be “fixing” something that was intentional.

The first producer in the economy can’t create products from nothing. Materials must already exist in the world before the first production cycle can begin. This is the economic chicken-and-egg problem — AgentE ensures that world initialization includes sufficient raw materials for at least the first generation of production.

The world must feel abundant on day one. Scarcity is earned through play, not imposed at birth. If consumables are scarce at launch, new players hit walls immediately and churn before they experience the core loop. AgentE validates that initial supply buffers exceed projected early demand by a comfortable margin.

Without entropy, wealth concentrates exponentially — the rich get richer with no counterforce. EVE Online survives because ships explode. Every stable economy needs aggressive sinks: decay rates, durability loss, maintenance costs, or expiration timers that force resources back into circulation and prevent runaway hoarding.

If 10,000 swords exist but the NPC vendor still sells at the Day-1 price, the market is lying. Prices must reflect total inventory, production rates, and consumption patterns — not static values set at design time. AgentE monitors price-to-supply ratios and flags disconnects where prices no longer reflect economic reality.

Producers who don’t check market conditions before producing create chronic overproduction. Every production decision should consider current supply, demand, and margin — otherwise agents pour resources into goods nobody wants. AgentE ensures production signals incorporate market feedback to prevent wasteful overproduction cycles.

A trade that yields 100 gold is worthless if there’s nobody to buy the output. Theoretical profit that ignores liquidity is phantom profit — it looks great on paper but never materializes. AgentE factors market depth, buyer availability, and transaction feasibility into profitability calculations, not just gross margin.

Agents stuck waiting for resources, market access, or role availability represent a silent bottleneck. If blocked agents don’t experience accelerated wealth decay, they accumulate as invisible deadweight — technically present but economically inactive. Faster decay on blocked agents creates urgency to switch roles or adapt, keeping the economy dynamic.

Adjusting sinks to fix inflation is pulling the wrong lever — inflation is a faucet problem, not a sink problem. AgentE maps each economic pathology to its root cause and selects the intervention that addresses the actual mechanism, not just the symptom. Wrong-lever interventions create secondary distortions that are often worse than the original problem.

A 1% adjustment per tick is better than a 10% correction once. Gradual pressure is invisible to players — they don’t notice the economy gently steering itself. Sudden shocks feel like punishment and erode trust. AgentE defaults to continuous micro-adjustments, reserving large interventions only for genuine emergencies.

High churn combined with low satisfaction may indicate oscillation from rapid adjustments. Cooldown enforcement is structural (handled by the Planner). This principle acts as a symptom detector — when the diagnostic fires, it signals that recent adjustments may have been too frequent or too aggressive.

Merged into P8. The classification step (structural vs pathological) is now part of P8’s check. A designed dominant role should not trigger population suppression. AgentE distinguishes between roles that are dominant by design and roles that took over unexpectedly.

High churn may indicate unexplained changes. Logging enforcement is structural (handled by DecisionLog). This principle flags high churn as a signal to review recent decisions — if players are leaving, the first question is whether they understood why things changed.

The ideal economic state is maximum demand from supply being just scarce enough to worry about — but not so scarce that it frustrates. When demand drops, the economy has oversupply and feels dead. When frustration rises, there’s undersupply and players leave. AgentE continuously searches for the sweet spot where scarcity drives engagement without crossing into frustration.

A static scarcity threshold combined with player progression means players will inevitably outgrow the challenge. As participants get stronger and more efficient, the scarcity threshold must shift upward to maintain tension — otherwise the economy becomes trivially easy and all engagement evaporates.

With N tradeable items, there are (N-1)N/2 relative prices to track. With thousands of items, no single agent — human or AI — can monitor them all. Economies with large item catalogs must be designed for distributed self-organization through market mechanisms, not centralized pricing by a regulator. AgentE focuses on aggregate health metrics rather than trying to manage individual prices.

Time-to-currency and currency-to-items ratios are the fundamental anchor values of any economy. When 1 hour no longer equals X gold, or X gold no longer buys Y items, the economy has drifted. AgentE tracks these anchor ratios continuously and raises alerts when any drift exceeds 20%, signaling that the core value relationships are breaking down.

A crisis visible at per-tick resolution may be completely invisible at per-100-tick resolution, and vice versa. AgentE monitors at three time scales simultaneously — fine (per-tick), medium (per-10-tick), and coarse (per-100-tick) — using ring buffers. When metrics diverge across resolutions, it signals a hidden crisis that single-resolution monitoring would miss entirely.

A virtual economy is never in true equilibrium — it oscillates around it. The aggregate arbitrage window across all relative prices serves as a live health metric. Rising arbitrage signals destabilization (prices are diverging from fundamentals), while falling arbitrage signals recovery (markets are correcting toward equilibrium). AgentE uses this as an early warning system.

Non-market exchanges — gifts, charity trades, social signaling, guild transfers — contaminate price signals if included in economic calculations. A player gifting a rare sword to a friend at zero cost doesn’t mean the sword is worthless. AgentE filters gift-like and below-market transactions before computing economic indicators, using trade-price-to-market-price ratios to detect non-market exchanges.

A Gini coefficient of 0 is boring — everyone has the same and there is nothing to strive for. Healthy inequality from skill and effort is fine. Inequality from money (pay-to-win) is toxic. Below 0.10 Gini means the economy is too flat; above configurable thresholds means oligarchy. AgentE watches the Gini coefficient to ensure inequality stays within the productive range.

Nothing permanently lost means inevitable inflation — the economy endlessly accumulates until everything is worthless. EVE Online thrives precisely because ships explode, removing value from the system. Every stable economy needs aggressive sinks: item durability, repair costs, failed crafting, PvP loot destruction. Without destruction, creation becomes meaningless.

If resource respawn rate merely matches consumption (1×), the economy drifts toward permanent depletion — any demand spike causes a shortage with no buffer. At 2× replacement, there’s breathing room for spikes, seasonal events, and population growth. AgentE monitors replacement-to-consumption ratios for every resource and flags any that drop below the 2× safety margin.

Concentrated idle wealth (Gini above 0.55, top 10% holding over 60%, velocity below 5) signals hoarding. AgentE raises transaction fees as a proxy holding tax when these conditions are met. Decay, storage fees, and item expiry are developer-implemented mechanics — AgentE detects the conditions; you build the response.

When one system is deterministic (crafting always succeeds) and another is probabilistic (combat has random outcomes), players feel betrayed by the random side. The expectation mismatch creates perceived unfairness even when the math is balanced. AgentE detects these friction points between deterministic and probabilistic systems and flags them for design review.

A game can have hundreds of economic dials, but AgentE should only tune 20 or fewer at once. This is not a limit on the game — it’s a limit on AgentE’s autonomous scope. Each additional parameter creates interaction effects with every other (20 params = 190 two-way interactions, 30 = 435). Beyond 20, AgentE fires a warning to prune low-impact parameters but continues operating normally.

If a new player can’t reach economic viability within a reasonable time window, they’ll churn before experiencing the core loop. Every economy must define a maximum time-to-value and ensure latecomers can reach it, even as the economy matures and early movers accumulate advantages. Without latecomer protections, the economy slowly dies as it loses all new blood.

When time-to-value exceeds 30 ticks with satisfaction below 55, the economy is demanding too much from its participants. This is a proxy metric — it does not measure individual available time, but signals when the collective experience suggests excessive grind. AgentE flags these conditions so developers can tune progression curves or add catch-up mechanics.

When the median spender’s power exceeds 2× the median non-spender’s power, free players leave permanently — the game feels rigged. AgentE targets a maximum 1.5× ratio, where spending provides meaningful but not decisive advantage. Above 2.0×, the economy enters pay-to-win territory and hemorrhages its free player base, which also collapses the paying player experience.

When (mean - median) / median exceeds 0.30, the mean is a lie. A few high-balance agents raise the mean while most agents have low balances. AgentE switches to median-based calculations for all health metrics when mean-median divergence exceeds this threshold, ensuring that decisions reflect what the majority of participants actually experience — not the distorted average inflated by whales.

Wild inflation swings (above 30%) may indicate insufficient simulation data. The minimum iteration count is structurally enforced by the Simulator. As a diagnostic, this principle detects large inflation rate oscillations that correlate with decisions made on too little data. AgentE applies conservative corrections when it suspects simulation noise.

Total intervention lag is 3–5× the observation interval. An adjustment made at tick 100 won’t fully manifest until tick 115–125 (assuming 5-tick observation windows). AgentE never re-adjusts a parameter before its estimated effect tick, because premature re-adjustment guarantees overshoot — you’re correcting for a correction that hasn’t landed yet.

When net extractors exceed 65% of total participants, the economy requires external subsidy to survive — it’s consuming more than it produces. AgentE triggers a yellow warning at 50% and red at 65%. Below 50% extractors, the system is self-sustaining. This metric is especially critical for economies with real-money bridges where extraction means actual cash leaving the system.

When intrinsic utility divided by market value drops below 0.30, the asset is speculation-dependent — its price is driven by expected future value, not current usefulness. Below 0.10 is critical: a price correction kills the economy because there’s no utility floor to catch the fall. AgentE monitors utility-to-price ratios as an early warning for speculative bubbles.

When correlation between gameplay currency and external markets exceeds 0.50, insulation has failed — your in-game economy now rises and falls with Bitcoin, stock markets, or other forces completely outside your control. AgentE monitors external correlation and flags any economy whose internal currency has become a proxy for external market movements. The response mechanism raises internal transaction fees to dampen arbitrage between internal and external markets.

Healthy live operations show a “shark tooth” pattern: revenue spikes with each event, valleys that hold their floor between events. When peaks start shrinking or valleys start deepening, it signals event fatigue — the content pipeline is failing to maintain engagement. AgentE monitors peak-to-valley ratios across event cycles to detect fatigue before it becomes terminal.

High completion (above 90%) combined with low satisfaction (below 60) suggests activities are not creating perceived value. Players are going through the motions without feeling ownership or attachment. This may indicate a missing endowment effect — the psychological phenomenon where people value things more simply because they own them.

Free completion rate of 100% means zero monetization pressure — there’s no reason to spend. Below 40% completion is predatory — it frustrates the majority and damages trust. AgentE targets the 60–80% sweet spot where most free players can complete events with effort, but spending meaningfully accelerates progress for those who choose it.

Low velocity (below 2) combined with low satisfaction after tick 100 signals supply stagnation. This is an advisory signal for the developer to audit content freshness — the economy may have run out of new things to discover, and the content pipeline needs attention.

Every new item injection shatters existing price equilibria. When a new rare sword drops, players must re-price every other sword in the economy — arbitrage spikes as the market digests the new information. AgentE builds cooldown windows for price discovery after content drops, waiting for markets to stabilize before measuring post-drop economic health. Judging the economy during the shock itself produces misleading readings.