dust
2% FLAP1,000 BPS THE VAULT9,000 BPS MEASURED ON A MAINNET FORK
The balanceWhat leaves, and what stays

06  /  Economics

The parameters we would fix at launch.

No token exists and nothing has traded. These are the values we would write into the launch payload, and the arithmetic that follows from them.

Status

Nothing is deployed. No factory, vault or ledger address exists on any chain, no transaction has been broadcast, and the repository carries no git remote. ./submission/check asserts all three.

01  /  The parameters we would fix at launch

A 2% tax, and no second leg to pay.

These are the values the launch transaction would carry. None of them has been set on chain, because no launch transaction has been sent.

Buy tax, to be set2%  ·  buyTaxRate 200
Sell tax, to be set2%  ·  sellTaxRate 200
Share of the tax Flap does not take that reaches the vaultmktBps 10000
Dividend legdividendBps 0
Commission receiveraddress(0)

With no dividend leg, the vault’s whole balance would be the payroll. There is nowhere else in the contract for a period’s tax to go.

02  /  The first cut

Flap’s fee is 1000 bps, measured on chain.

Read from the launch this suite actually executes on a mainnet fork, rather than quoted from documentation. So of every 100 units of trading tax, 90 units reach the vault — 9000 bps.

The number is no longer merely printed to a log. DustVault.mainnet.t.sol asserts feeRate == 1000 with equality, and that assertion was itself broken on purpose and watched to fail, so a change in Flap’s fee turns the suite red with a message telling you to re-derive every revenue figure on this page.

Commission receiver left at address(0)9000 bps to the vault
Any address named as commission receiver8700 bps to the vault

That 300 bps would be carved out of the vault’s own share — a cut of the wage bill, not new money and not out of anybody else’s slice. Which is why the receiver stays at zero. The difference is measured by a mainnet-fork test rather than assumed — one of the three that need an archive endpoint to run at all, which the audit page sets out.

03  /  What the pot buys

A period can only pay for as many addresses as its own tax stream supports.

Sixty per cent of a period is the wage bill, split by headcount and by nothing else, so the wage per address falls as the room fills. The entry fee does not fall with it. The two meet at one headcount, and that headcount is the cap — arithmetic rather than an allowlist, which is why the vault carries no Sybil defence at all.

a sweeper spends fee and receives about 0.6 × pot / N
so pressing stops paying once N > 0.6 × pot / value(fee)

The equilibrium  /  what a pot can pay for

Three periods, three pot sizes, one entry fee. The wage is 0.6 × pot / N; the fee is flat at every N, because the ledger carries no function that could move it. Schematic: the curves are that expression, the pots and the fee value are chosen to read against each other.

Wage per address, 0.6 times the pot divided by headcount, drawn for three pot sizes. Each curve crosses the flat entry-fee line at a different headcount: six addresses for a pot of half a BNB, twelve for one BNB, twenty-four for two. Past its own crossing, a press costs more than it returns. pot 2.0000 pot 1.0000 pot 0.5000 wage = fee at N = 6, 12, 24
one crossing per pot — twice
the tax in a period supports
twice the room, and no more
value(fee)  0.0500
  • pot 0.5000 BNB — wage meets the fee at N = 6
  • pot 1.0000 BNB — wage meets the fee at N = 12
  • pot 2.0000 BNB — wage meets the fee at N = 24
  • value(fee) — 0.0500 BNB, the same number at every N
This is the whole Sybil answer, drawn. The wage falls as the room fills, the fee does not move, and where they meet is the headcount that period can support — N > 0.6 × pot / value(fee). Nothing decides that number: a bigger pot pushes the crossing right, a dearer fee pulls it left, and a period whose crossing sits below one address is a period nobody presses in, which is the 98% tier further down this page.

Two consequences the rest of this page keeps running into. Quiet stretches produce periods nobody presses in, because their crossing sits below a single address — and those periods run at the 98% tier. And the 60/30 split is not neutral to any of this: the wage bill is what the crossing is computed from, so the share the project takes sets how many participants the same tax revenue can carry. That is why the open item at the foot of this page is an economic question and not a courtesy.

04  /  The project’s share is a range

30%, 38%, or 98%.

The vault’s 1.8% of traded volume would be divided by four constants that have no setter: wage 6000, project 3000, referral 800, bounty 200. The project reaches that money by three different paths, so the honest way to read the line is a range.

CaseReached byOf traded volume
30% tiersomebody pressed, and everybody had a referrer0.540%
38% tiersomebody pressed, nobody was referred — the 8% and the remainder fall through0.684%
98% tiernobody pressed — the period’s bucket goes to the project, less the bounty1.764%
The wage billthe 60% that is split equally by headcount1.080%

05  /  An extrapolation, not a measurement

The same arithmetic, at volumes nobody has traded.

This table is the one projection on the site. The percentages above it are fixed by constants in the source; the dollars below are those percentages multiplied by a daily volume we picked. No token has launched, no volume has been observed, and none of these rows is a forecast of what any of them would be.

Daily volume30% tier38% tier98% tier
$10,000$54$68$176
$50,000$270$342$882
$200,000$1,080$1,368$3,528
$1,000,000$5,400$6,840$17,640
$5,000,000$27,000$34,200$88,200

Two things to carry away from that table.

Empty periods run at the 98% tier. Headcount caps itself — pressing stops paying once N > 0.6 × pot / value(fee) — so quiet stretches produce periods nobody presses in by design, and those periods go almost entirely to the project.

The 38% tier is commoner than intuition suggests. One participant without a referrer sends their 8%, and the rounding remainder, to the project. At a cold start almost nobody has been referred, so a launch begins at the 38% tier rather than arriving there.