vellus runs on a single premise: the more evenly it is held, the more it pays, and the relationship is not a slogan but a function. a fee pool releases to holders at a rate equal to the square of the ledger's evenness, recomputed every breath from the actual shape of every balance on chain. no person sets the rate. no vote adjusts it. the distribution is the dial, and only the distribution can turn it.

in 1912 an italian statistician named corrado gini published a single number that could describe how unequally anything is shared across a population. economists have used it for a century to measure nations. it has never once been used as a price signal, because no market has ever been able to compute it live, trustlessly, at the moment of every trade. a blockchain can. vellus takes gini's coefficient off the shelf, wires it into the transfer path itself, and makes it the throttle on everyone's cash flow.

here is the machinery. vellus is a token 2022 mint whose transfer hook maintains a sixty four bucket logarithmic histogram of all holder balances. when a transfer executes, the same instruction that moves the funds rebuckets the sender and the receiver, so the histogram is never stale by even one transaction. from those buckets the program derives G, the concentration index, in fixed point arithmetic, exact to the bucket, with nothing sampled and nothing estimated and no oracle consulted anywhere in the pipeline.

the fee side is deliberately boring. thirty basis points on every transfer, collected by the transfer fee extension, swept into a vault called the follicle at each breath boundary. one inflow, one vault, no exceptions. the interesting part is the outflow: at every breath, the follicle releases R = P · (1 − G)² pro rata across all holders, where P is everything it currently contains. at G of 0.2 the vault is a pipe and 64% of it flows out per breath. at G of 0.85 the vault is a tomb releasing barely two percent. the difference between those two markets is not policy. it is arithmetic.

the exponent matters. a linear penalty would let concentrated markets keep half the flow at G of 0.5, which is roughly where most tokens live and die. squaring the term makes concentration progressively cruel: each additional point of G costs more yield than the last one did. the regime where whales operate is precisely the regime where the curve is steepest, and that is not an accident of the math, it is the reason this exponent was chosen over every other.

releases do not fire themselves. the breath closes and then waits for a call to groom(), which any address on the planet may send. the groomer's transaction snapshots the index, computes the release, distributes it, and keeps one percent as payment for being first. this converts system maintenance into an open race: when the pending release is fat, bots will fight over the boundary in milliseconds, and when concentration has strangled the release to dust, the follicle compounds unattended until dispersion makes grooming worth racing for again. the system does not need anyone in particular. it needs anyone at all, and it pays whoever shows up.

consider the whale's position, because the design is really a letter addressed to them. accumulating vellus is legal, easy, and self-defeating. every million tokens gathered into one account pushes G upward, which cuts the release rate, which reduces the income on the very stack being built. the whale is not fined, not blacklisted, not confiscated. their yield simply thins with every buy, on a public curve, while the locked pool they are creating waits for whoever eventually spreads the supply back out. a corner in vellus is a donation with extra steps.

the obvious counterattack fails on arithmetic. splitting one enormous balance across ten thousand fresh wallets feels like dispersion but is not, because a gini coefficient measures the lorenz curve of the distribution, not the count of addresses holding it. dust wallets add population at the bottom while the mass of the supply stays exactly where it was, which moves the curve almost nowhere, and in bucketed form moves it even less. lowering G requires transferring real value to hands you do not control. there is no cheaper trick, and a century of statistics literature is the audit.

so what is a holder actually long on. not emissions, there are none. not scarcity theater, the supply of 33,554,432 is fixed and nothing ever burns. a vellus position is a bet on breadth: that this asset will be held by more hands tomorrow than today, because every new small holder mathematically raises the income of every existing one. it is the only position in this market where you rationally want strangers to have some, which inverts the usual psychology of holding so completely that it takes most people a minute to believe it.

two structural facts deserve plain statement. first, custodial platforms are one holder as far as the histogram is concerned, so float parked on an exchange raises G and suppresses the release for everyone, including that exchange's own customers. vellus therefore pays a standing, measurable bounty for self custody, which no other asset has ever managed to price. second, the system has edges it does not pretend to smooth: a market of five wallets will read concentrated because it is concentrated, and no formula owes anyone a good number.

the program was sealed at deployment. upgrade authority gone, mint authority disabled, and the follicle carries no withdraw path in its account structure, so the questions people usually ask about rugs do not have a subject here. our own wallets obey the same histogram, pay the same thirty basis points, and wait on the same curve as everyone else's. the machine reads the ledger, the ledger reflects the holders, and the holders decide with every transfer what the machine pays. it needed builders once. it does not anymore.