WARLANDS

Most token-incentivised games die because they mint rewards faster than they burn them; the token inflates to zero. WARLANDS inverts the model. Every unit of $HEXAR paid out as a seasonal reward is first collected from a real sink — staking fees, market fees, upkeep, construction, training — in that same season. The protocol can never distribute more than it took in:

Players earn by being economically and militarily productive relative to other players, never by extracting freshly printed supply. The result is a closed, conserved economy: a fixed 1,000,000,000 $HEXAR supply that only ever moves between players, locked stake, allegiance treasuries, and the burn furnace — an invariant we enforce in code and verify with over a million randomized simulation steps (§10).

One logical world per season: a large axial hex grid with a deliberate center→edge risk gradient. The outer ring is the newbie belt (gentle starter land); the mid-ring is the heartland(the full economic mix); the core is the Crucible — high-value, contested, perpetually at war. Land is finite (≈250,000 plots at 100k-CCU scale), which makes scarcity the core economic primitive: not everyone gets a Warzone plot.

You secure a plot by staking $HEXAR against it (locked, never spent). Higher-tier terrain demands more stake and yields more — and is harder to defend. Unstaking returns your principal minus a small early-exit fee (a sink); losing a plot to conquest is principal-safe — you lose built assets and stored goods, not your locked stake.

Twenty resources across three tiers form a genuine supply chain. Extractors gather eight raw resources from terrain that supports them; factories refine and manufacture the rest. Production follows Base × Terrain × Level × Workforce × (1 − congestion) × DiminishingReturns(plot) — super-linear upkeep and diminishing returns are deliberate anti-whale pressure.

Food and Water are universal upkeep — armies and workforces consume them every tick, creating permanent demand for farmers. Storage is capped (warehouses raise the ceiling), and starvation softly degrades a plot’s defenses, so a war machine without a food economy rots from within.

Six unit classes locked in a counter triangle. There is no single dominant army — every composition has a hard counter, so scouting and combined-arms beat brute force.

Battles are resolved by a seeded, deterministic engine: the same inputs always produce the same auditable round-by-round log. RNG is bounded (a controlled damage roll), and decisive territory battles run more rounds so variance averages out and skill dominates at the top. Raids loot stored goods (a vault protects a share); sieges can break defenses. Conquest is principal-safe — the loser keeps their locked stake.

A single shared, player-driven order book. Sellers list goods (escrowed into the order for a flat listing fee); buyers sweep the cheapest asks, and $HEXAR flows directly from buyer to seller. A transaction fee skims every trade into the sink router — split between the burn furnace, the season reward pool, and allegiance/region tax. The market is the economy’s primary sink engine and its price-discovery layer: scarcity and war shocks ripple straight into prices.

Players band into allegiances with a shared treasury. Members contribute $HEXAR (raising their contribution score) to fund allegiance buildings that buff every member’s plots — a Research Center grants +12% production, a Fortress +15% defense, a Trade Hub −25% market fees.

Construction runs through on-chain-style governance: any member proposes a building, members vote for or against, and a proposal whose window closes with a majority is auto-resolved and built from the treasury — settled deterministically inside the world tick. Founders retain a direct-build fast path. Disband returns the remaining treasury to the last member out; nothing is destroyed silently.

The world runs in 30-day seasons. Throughout a season, sinks fill a reward pool. At season’s end, a four-factor score — economy (value produced & traded), military(raids/sieges won), territory (control, weighted by plot tier), and allegiance(contribution) — ranks every commander, and the sink-funded pool is distributed accordingly. Then the map resets and the meta shifts, while account-permanent layers (commanders, reputation) carry forward.

Because the pool is funded only by collected sinks, the season cannot pay out more than the economy produced. Rewards are a redistribution of productivity, never an emission.

Every fee, upkeep charge, construction, training order and early-unstake penalty routes through the Sink Router, which splits collected $HEXAR three ways:

This is the entire sustainability thesis: the protocol literally cannot distribute more than it collected, and a slice of everything is burned forever — so net supply trends flat-to-deflationary with activity.

The simulation is server-authoritative. A pure, deterministic core (src/sim) owns the rules — world generation, a 1 Hz tick, and validated commands — with zero IO, no randomness outside seeded paths, and no framework entanglement. The same module runs on the server and ships to the client.

• Authoritative server — Node + ws: holds the world in memory, ticks it, validates every command, and broadcasts state to all clients.
• Persistence — JSONB world snapshots to Postgres every few ticks; the world reloads on boot, surviving restarts.
• Client — Next.js 16 / React 19. The polished single-player prototype lives at /play; the live shared world at /world.
• Determinism — battles and ticks are reproducible; snapshots round-trip losslessly, which is what makes authoritative persistence safe.

Value events settle on an EVM L2 (deployed to Base Sepolia); high-frequency gameplay stays off-chain and authoritative. Four self-contained Foundry contracts, no external libraries:

Staking is principal-safe: the StakingManager escrows stake and only ever returns it to the owner (minus sinks) or transfers a plot via conquest — it has no power to seize principal. The treasury settles claims only from a sink-funded pool — paying out solely pool-earned $HEXAR, capped per commander, never minting. Genesis supply is fixed at the $HEXAR mint and burnable through the sink router.

The economy makes a hard promise — a fixed, conserved supply — so we hold the code to it. The simulation core is built test-first, and a property-based fuzz harness throws random valid and invalid command streams at it, asserting after every step that:

• $HEXAR is conserved: Σ balances + Σ locked stake + Σ treasuries + burned = constant.
• resources never go negative, no value is NaN, structural integrity holds, and no input ever throws.
• a JSON snapshot round-trips to an identical subsequent tick (persistence safety).

A 1.2-million-step run across many seeds and players surfaced five places where $HEXAR was silently destroyed without being recorded as burned — building, upgrading, training, a market rounding residue, and treasury loss on disband. All five are fixed (the spent $HEXAR is now correctly booked as a sink or refunded), and the fuzzer is a permanent regression guard. The supply is now provably conserved under a million-plus randomized operations.

Shipped — the full single-player feature set, now server-authoritative multiplayer, deployed and verified live in six slices:

On the roadmap to a full launch:

• Merkle reward pipeline (season scoring → on-chain claims) atop the deployed contracts.
• Sector sharding & autoscaling for 100k-CCU scale.
• Wallet-based identity + anti-cheat signatures (anonymous sessions today).
• Contract audit hardening, then mainnet launch on an EVM L2.

Stack: Next.js 16 · React 19 · TypeScript · Zustand · wagmi/viem · Node + ws · Drizzle ORM + Neon Postgres · Foundry / Solidity. Open /world in two tabs to watch the authoritative world sync in real time.