The seam
It empties, and the grid does not move.
Difficulty only rises, so eventually the best machine available cannot cut a block at all. Where a block belongs is fixed by the grid it was seeded in.
Nothing is deployed. No address exists on any chain.
There is no factory address, no vault address and no mine address, on mainnet or on any testnet. No transaction has been broadcast. What exists is source, a test suite, the measurements on these pages — and a deploy script that has not been run.
A harder seam pays more per block, not less.
This is the part that runs against the intuition the word “mining” invites. The pot is the tax a period’s trading produced. Its size depends on how much the token traded — not on how hard the seam is. The pot does not shrink as the seam hardens.
Difficulty does not decide who wins a race. It decides which machines can still finish inside a 200-block window at all, so slower rigs fall out first. Fewer rigs finish, fewer blocks are cut — and the settlement divides the same pot among those fewer blocks.
The limit is not a small wage. It is no wage at all: a period nobody cut has no wage to pay, and its bucket goes to the project instead. That is a branch in the contract anyone can read, not a rule bolted on.
Source: MECHANISM.md — “the pot is the same and _settle divides it with wageRate = wagePool / cut, among fewer blocks”.
200 blocks, and periods that are whole multiples of them.
A puzzle window is 200 blocks wide, and each miner may cut one block per window. A payout period is a separate, longer cadence set at launch: between 200 and 1,000,000 blocks, and required to be a whole multiple of 200.
The multiple-of-200 rule is enforced by a second require, which is why the reachable set is 200, 400, … 1,000,000 and the stated lower bound of 20 is unreachable. Without the alignment, a cadence like 250 would give some periods one window and others two, and any cadence below 200 would leave most periods with no window at all — structurally unpayable, sweeping to the project while the miners’ 60% could never be reached.
Everything here is counted in blocks, never in seconds. No part of the contract assumes a seconds-per-block figure. Elapsed wall time moves with BSC consensus, and the contract does not care.
Source: MECHANISM.md, FACTORY.md, README.md
Block numbers follow the same worked example as the console: a launch block, 200-block windows on a fixed grid from it, and a 1,200-block payout period, which is six whole windows. A cadence that is not a whole multiple of 200 would leave some periods holding no window at all — structurally unpayable, which is why the factory requires the multiple.
Runtime size against the 24,576-byte EIP-170 limit.
Puzzle parameters shown
Not connected to any chain. Nothing is deployed; these are the bounds and constants the code enforces.