Because tokens are permanently unspendable once they're at 0x…dEaD. balanceOf(0xdEaD) is the canonical burn counter — every block explorer, every analytics dashboard, every chart treats it as supply destruction. We chose this path over ERC20._burn so accounting stays public, indexable, and provable on Basescan in one query — no internal state assumptions required.

The hook only has afterSwap permission — it can't observe price before swaps and can't reorder. The internal BURN→ETH swap when the unspecified side lands in BURN happens atomically in the same transaction; slippage is bounded by the pool's own liquidity, and the worst case is a smaller pot contribution, never a drain. The pot itself is held in BurnGame, not in the hook.

Math. Each burn costs 500,000 BURN. At current market price that's ≈ 0.0124 ETH — call it $43. Holding the lead for a full 10 minutes against a single competitor requires winning a war of attrition where the floor on cost is the BURN token's spot price, which strengthens every round as supply at 0xdEaD grows. Whales can burn — but they pay the same price as everyone else, and they don't get the timer cheaper.

Calibrated so a single burn is a non-trivial commitment (≈ $43 at launch) but not so steep that the average wallet is priced out. At 0.05%, the supply can absorb 2,000 burns before the dead-wallet share exceeds 100% — leaving room for many full rounds before the game becomes uneconomical and the protocol naturally retires.

It seeds the next round's pot immediately. Round 7 opened with 0.86 ETH already inside from round 6's rollover; the rest accumulated from hook fees and the burn-side conversions. There is no treasury, no team allocation, no admin withdraw — the 20% is hard-coded into BurnGame.settleRound() as a forwarding op.

It's a coordination game with on-chain payouts. Whether that's gambling in your jurisdiction is a question you should answer yourself before you click anything. Nothing here is investment advice. Tokens that go to 0x…dEaD do not come back.