We tried to build a passive arb-yield launcher for small token projects. Here's what the simulator says.

The premise

Most token launches force a project's founder into a uncomfortable spot: they hold a lot of the token they just created, and the only way to make money from it is to sell some of their bag. That's extractive. It makes the founder a counterparty to their own community.

We started thinking about a different design: what if the project owner deploys liquidity infrastructure that generates passive arb income, without ever needing to sell the underlying token? You set up curves between $TOKEN and a few volatile pair assets (WETH, cbBTC, a memecoin). External arbitrageurs notice price spreads whenever the pair assets move, do their thing, and pay fees on the way through. Those fees are the founder's income. The token itself accrues value from supply burns + scarcity. Founder doesn't have to touch their bag.

For this to work, the design needs to be viable at small scale — projects launching with $0–$50 of expected daily trading volume, growing slowly. If you need $10k/day in volume to break even, you've reinvented Uniswap and you're not building anything new. The interesting question is whether the burns + cross-pool arb dynamic generates meaningful income at the very bottom of the project lifecycle.

What we tried first: discrete bands

The first design was a Uniswap V4 hook that managed liquidity in discrete bands — pre-planned tick ranges with different L in each. The hook materialized one band on-chain at a time, and transitioned to the next band when price crossed the boundary. The theoretical upside: you can shape the price-vs-supply curve any way you want by tuning the L per band.

We deployed this for $TRINI and $EPIC. Both tokens have 18 bands per pool and several pools per token. And both ran into the same problem.

The leak

Bands look elegant on paper but they have a property that hurts real users: at any moment, only the EPIC committed to the active band is "in the pool." The other 95%+ of the supply sits in the hook contract as reserves, waiting. From the perspective of a trader, the pool's effective depth is just the active band's depth — not the total committed liquidity.

We measured this with a $300 simulated buy on the $EPIC/USDC pool: the trade walked through 5 bands, the price pumped 16×, and after arbitrageurs equilibrated the side pools the buyer had effectively lost 64% of their $300 to arber profit. Not fees, not slippage — just pure value extraction enabled by artificial shallowness.

That's not a small project's friend. If your supporters lose 64% every time they buy any meaningful amount, you don't have a project, you have a leak.

The fix: continuous positions

The fix turns out to be straightforward: don't use bands. Use a single big concentrated liquidity position per pool. All the EPIC committed contributes to L at every price within the range. The curve is the standard x×y=k Clanker shape. A $300 buy now creates a ~2× pump (instead of 16×), which means much smaller spread for arbs to extract. The leak drops from 64% to single digits.

You give up the ability to "shape" the curve (it's just a fixed x×y=k inside the range) but you keep multi-pool dynamics, you keep the cross-pool burn engine, and you stop extracting value from your own users.

With this redesign in hand, we wanted to know: does it actually generate enough fee income to make sense as a launcher product for small projects? Time to simulate.

The simulation

We built a year-long simulator that models four scenarios under identical crab-market conditions (zero net drift on all pair assets):

External pair asset prices follow geometric Brownian motion with annualized vols of: WETH 60%, cbBTC 55%, CLANKER 150% (memecoin volatility). Daily timestep. Zero drift means the year ends with prices statistically unchanged on average — a true crab market. Each scenario uses a deterministic RNG seed for reproducibility.

The pair-asset random walks

Just so you can see the underlying volatility: here's what the simulated WETH, cbBTC, and CLANKER prices looked like over the year. Note CLANKER is plotted with prices ×100 to share an axis with WETH (which starts at $2,200) — the actual CLANKER prices in the sim hover around $25.

The system has to capture value from these movements via arb — and ONLY via arb. No directional bet on the pair assets is built in.

The results

Here's the year-end summary:

What we learned

1. Multi-pool architecture is real money — when there's some activity

At $50/day organic volume, the multi-pool system generates 2.2× more treasury income than the single-pool design ($216 vs $98), AND it burns nearly 4× more EPIC, AND it accumulates over $1,000 of multi-asset treasury reserves in the side pools. The cross-pool arb engine genuinely amplifies the returns from the same organic activity.

2. The bootstrap problem is the killer

With zero organic volume, the multi-pool system generates exactly $0. Not "small numbers" — literally zero. This is because the side pools start single-sided in EPIC (they hold only the token, no quote asset), so when an arb opportunity opens from external pair asset moves, the arber has nothing to BUY from the side pool against. The chicken-and-egg loop never closes.

The fix is deployment-time pre-funding: the project owner puts $20 of WETH, $20 of cbBTC, and $20 of CLANKER into the side pools at deploy. ~$60 of capital and the bootstrap goes away. Type B arbs fire from day one.

3. Bands cost more than they gave us

The discrete-band hook design we built first looks elegant on paper but extracts ~50–65% of every meaningful trade as arber profit. That's a tax we never wanted to charge our supporters. Continuous positions remove the tax entirely. There's no good reason to use bands in this design — the curve-shaping benefit is too small to pay for the leak.

4. The design needs volatile pair assets

The Type B arb income comes from external pair asset moves. Stable-stable pools (EPIC/USDC paired with EPIC/DAI, say) would generate near-zero income because there's no spread to extract. Pairing with WETH (60% annual vol), cbBTC (55%), and CLANKER (150%) gives the system three independent volatility sources, which compounds nicely. The more volatile the pair, the more cycles per unit time, the more income.

5. It's not a get-rich scheme

$216/year on $50/day volume is ~12% annualized on the daily volume flow. That's good for a passive system but it's not life-changing. For a small site coin or community project, it's a real income stream that scales linearly with growth. For someone trying to make a full-time business, it's not enough.

What we're doing about it

The current $TRINI and $EPIC deployments use the band-based hook, which we now know leaks too much value to be a good deployment of the architecture. The tokens themselves stay live, but we're going to drain the existing pools (the band-based ones) and stop running them.

The longer-term plan: rewrite the hook contract to use continuous concentrated positions, add the deploy-time pre-funding capability, and turn the whole thing into a templated launcher that other projects can use. The simulator gives us confidence that this is worth building. The data above is the proof-of-concept.

If you're a small project that wants to run a curve like this and would like to be an early test deployment, get in touch via the link in the nav.