How curves differ
Constant product, linear, stepped and quote-denominated curves produce different price paths from the same demand. What the curve parameters control and what they cannot control.
Strip the branding off a launchpad and four decisions are left. What function prices the primary phase. Who pays what, when, and to whom. What condition declares the primary phase finished. Where the accumulated liquidity is deposited and who can touch it afterwards.
These four notes take one decision each. They are written so that a venue nobody has heard of yet can be read in an afternoon: find its answer to each of the four questions and you know how it will behave, whatever its interface looks like.
The four decisions are not independent. A steeper curve interacts with a reserve-denominated graduation target; a large migration deduction interacts with the depth of the destination pool. These are the interactions worth holding in mind while reading the notes below.
A reserve-denominated target with a steep curve ends the primary phase while a large share of supply is still undistributed. The same target on a flatter curve distributes far more of the supply before it triggers. Same threshold, different token.
Curve / thresholdA percentage fee on every primary trade compounds against the number of trades, not against the amount raised. Venues with lively secondary trading during the curve phase extract materially more from the same eventual reserve.
Fee / churnWhatever is retained at migration is depth the destination pool never receives. A pool seeded from a reduced reserve is thinner from its first block, and thin pools move further on the same order.
Deduction / depthAggregators support programs, not tokens. A destination on a widely integrated program is routable within minutes; a destination on a newer program can be tradeable and unroutable at the same time.
Program / routingA continuing creator fee is a mechanism and a signal at once. The mechanism is small; the signal is not, and it is read by exactly the people deciding whether to hold.
Share / perceptionBurning, locking or retaining the position produced at migration are three different risk profiles. Only one of them can be verified from the chain without taking anyone at their word, which is the point.
Position / trustThe four decisions every launchpad has to make: the shape of the pricing function, who pays what to whom, the condition that ends the primary phase, and where the resulting liquidity is put.
Constant product, linear, stepped and quote-denominated curves produce different price paths from the same demand. What the curve parameters control and what they cannot control.
Trading fee, creation fee, migration fee, creator share and the fees you pay to the network rather than the venue. Six places money leaves, and who receives it in each design.
Every launchpad defines the end of its primary phase differently. Reserve targets, supply targets, time windows and manual triggers, and what each choice does to the last hour of a curve.
The destination is a product decision, not a technical inevitability. Native AMMs, third-party AMMs, concentrated designs and what the choice does to depth, routing and fee accrual.
Two venues that made identical mechanical choices will behave identically no matter how different their front pages look. Two venues with identical front pages and different curve constants will not. That asymmetry is why this section is the largest one in the ledger.
Venue profiles. What each launchpad is as a product, who is standing in front of it, what its interface makes easy, and which of its properties are actually different rather than differently named.
Open this sectionTurning the property matrix into a decision. Which differences matter for a given launch, which are noise, and what genuinely changes if a team moves to a different venue next time.
Open this section