A launchpad token does not sit still. The reserve behind its price grows, the impact of a fixed trade size falls, the remaining curve capacity shrinks, and then a pool replaces the curve entirely. Curve Campaign asks the same planning questions at all four depths and writes down why the answers differ.
There is a layer between owning a tool and being able to say afterwards what a run did. This desk sits in that layer, and it treats the depth of the curve as the variable that changes every answer.
How large should a trade be? How often should one land? How many accounts should carry the flow? Each question is answered once per phase, because a reserve that has quadrupled turns the previous answer into a different instrument.
Open the phase sectionNo setting has a correct value on its own. Every parameter here is written with the failure it causes when the value is too low and the different failure it causes when it is too high, so the choice is a bet you can defend.
Open the parameter sheetsMost avoidable damage happens between phases: a size band carried across a doubling of the reserve, or a route still pointed at a curve after the pool has taken the liquidity. The handover gets its own written checklist.
See the handoverEach column names what the curve looks like at that depth, the question the campaign has to re-answer there, and the parameter that shifts most when the token arrives.
The phase plan explains the whole document. The other three are the decisions that most often get carried from one depth of curve to the next without anyone re-deriving them.
The whole campaign laid out against four depths of curve, with the parameter block re-derived at each one and a written transition between every pair.
Phase planAt the bottom of a curve the reserve is small, impact is high and fees dominate small trades. What that combination does to swap size, interval and the readable record.
Phase notePacing turns a budget into a shape, and the middle of a curve is where the shape stops being obvious. Four pacing patterns, the interval derivation, and when to slow down.
ParameterThe last stretch before a curve completes is the part most often run on yesterday settings. What changes on approach, what the handover needs, and where the run should stop.
TransitionPhases covers where the token is and what that state does to a campaign. Parameters covers the individual fields that have to be re-derived every time the state changes.
A curve token is never in one state for long. These notes take the campaign decisions phase by phase, because the same swap size that is sensible at the bottom of a curve is a different instrument two thirds of the way up it.
Open the sectionPacing, swap size, depth and the venue split are the fields that have to be re-answered whenever the curve moves. Each one is written here with the failure it causes at both ends of its range.
Open the sectionThis is the sheet the desk works from, and it gets filled in once per phase rather than once per campaign. Each line names what goes wrong at the bottom of its range and what goes wrong at the top, because the two failures are different and only one of them is obvious while the run is happening.
| Field | Unit | What it decides | Too low | Too high |
|---|---|---|---|---|
| Swap size | SOL per trade | The band each trade is drawn from, set against the reserve sitting in front of it rather than against the budget behind it. | Fees and priority fees take a larger share than the trade itself, so the run buys cost instead of turnover. | The quote moves on every trade and the campaign pays its own slippage on both sides of a round trip. |
| Interval | seconds | The gap between trades, and whether that gap is fixed, drawn from a range, or triggered by an event. | A burst that empties the allocation in minutes and leaves the rest of the window silent. | Activity so sparse that no chart interval and no observer window ever contains enough of it. |
| Wallets | count | How many funded signing accounts carry the flow, and how many trades each one is allowed to contribute. | A short list of accounts recycling, which reads as one participant and concentrates every failure. | Account creation, rent floors and sweep transactions on wallets that never carry readable flow. |
| Buy and sell mix | ratio | How the flow is split between the two directions, which on a curve is also a decision about where the quote ends. | A one-directional run that pushes the quote up and leaves the campaign holding the position it created. | A perfectly balanced round trip that pays two sets of costs to arrive back where it started. |
| Phase allocation | share of budget | What proportion of the total is committed to each depth of curve before the run begins. | A phase funded too thinly to produce anything readable, which wastes the phase rather than saving the budget. | The budget exhausted at one depth, leaving the transition that mattered most with nothing behind it. |
| Priority fee | micro-lamports | What each transaction offers for inclusion, which is a congestion decision rather than a campaign decision. | Transactions that expire unlanded, so the pacing curve on paper stops matching the record on chain. | A fee line that grows faster than the flow it protects and quietly rewrites the cost per swap. |
| Venue | curve or pool | Which mechanism the flow is routed through, which changes at migration whether or not the plan noticed. | A route pinned to a venue that no longer holds the liquidity, so trades fail or fill somewhere unintended. | A split so wide that no single venue shows a change worth reading in any window. |
| Stop rule | condition | The written condition that ends a phase early, agreed before anyone has an opinion about how it is going. | No rule, so the phase ends when its allocation does and nobody ever takes a decision. | A trigger so sensitive that ordinary variance halts the campaign in its first quarter hour. |
A phase is ready when the depth is described in one sentence, the size band has been derived against the reserve currently sitting in front of it, the pacing pattern is drawn rather than implied, the allocation for the phase is ring-fenced from the rest of the budget, and the condition that ends the phase early has been written down. At that point the work moves from a sheet to a console, and the console only has to execute what the sheet already decided.
The remaining notes, in the order a curve token usually needs them.
Size is chosen against the reserve in front of it, not against the budget behind it. Why the same trade behaves differently at three depths, and how the band is derived.
Read the noteMigration replaces one price mechanism with another. What the campaign has to re-point, which parameters survive the move, and what the first post-pool hours are actually for.
Read the noteThree commitments that decide what can be published here and what cannot.