Staking With Notice, Not Locks

"Show me the incentive and I will show you the outcome." - Charlie Munger

Munger meant it as a warning: design a system around what you hope people will do, and the system will fail exactly where their real incentives point elsewhere. Most staking designs fail this test at the moment it matters most.

---

What a Lock Buys, and When It Stops Buying It

The standard model is the hard lock. Commit your tokens for a year, receive a boosted reward, wait. It looks like commitment. Walk through the incentives and it comes apart in three places.

A lock pays for a decision made in the past. The holder chose to commit once, on day one. On day 300 they collect the same full multiplier while providing 65 days of remaining protection. The reward stays constant while the thing it pays for melts away.

The unlock date is a coordination point. Every holder in the same lock cohort becomes liquid at the same moment, and every one of them knows it. The rational move at the cliff is to exit before the others do. The protocol spends a year paying for loyalty and receives a stampede on a known date.

Commitment must be re-sold forever. When the lock expires, the default state is liquid. The system has to convince every holder to sign up again, every cycle, usually by raising the reward. Retention becomes a recurring cost.

Banking solved this decades ago. A certificate of deposit is a hard lock: fixed term, penalty to break. A notice account is the alternative: your money stays put by default, and withdrawing requires you to file notice and wait out an agreed window. The bank pays for the warning, because warning is what it actually needs. BASIS staking is a notice account for tokens.

How the Notice Ladder Works

At TGE, 48% of platform fees route to the BASIS staking contract and are paid out to stakers in USDB. Your portion of each distribution is your tokens times your weight, divided by the sum of everyone's tokens times weights. Weight comes from the notice period you choose:

Five mechanics complete the picture:

1. Filing notice starts a linear release. Tokens unlock at 1/W per day across the window. No cliff, no single unlock block.
2. Cancelling notice restores full weight instantly. The notice clock resets. Changing your mind costs nothing beyond the reduced earnings while the notice ran.
3. Weight decays during notice. From the moment you file, your weight walks linearly from your rung down toward 1.0x over the window.
4. Released tokens keep earning at 1.0x until you claim them.
5. Each position runs its own clock. Stake in batches, exit in batches.

The ladder doubles as the vesting schedule. Each presale round is assigned a rung as its minimum: public at Monthly, private at Quarterly, seed at Half-Year, angel at Annual. The founders' allocation is burned into Permanent. Any holder may elect any rung, Permanent included. Airdrop tokens land at Open, liquid from day one.

The reward pays for participation and for keeping tokens out of circulation over the notice period, not for ownership itself.

Four Properties That Fall Out

The reward pays for warning, and the warning never expires. An Annual staker provides 365 days of notice on day one and 365 days of notice on day one thousand. Protection stays constant instead of decaying toward a cliff, so the payment for it can stay constant too. Commitment is the default state. There is nothing to renew and no re-lock to sell.

There is no date to trade against. Exits release 1/W per day, so even a large Annual position reaches the market over a year, not in a block. There is no moment when a cohort becomes liquid together and no calendar date for the market to front-run.

Leaving pays the people who stay. Payout shares sum to 100%. When a holder files notice, their weight starts decaying, and the difference flows to everyone still committed. The premium for commitment is funded by the impatient, not printed as emissions. Fixed pot, shifting shares: the system rebalances instead of inflating.

The pot is external. Distributions are USDB from platform fees, not newly minted tokens. Reward and dilution are not the same lever, which removes the reflexive loop where paying stakers debases the thing they staked.

The Exit Queue That Comes Back

In a lock system, a departing holder is gone. In the notice ladder, the exit queue is a recovery pool, because everyone in it is still holding, still earning, and still receiving USDB payouts every distribution.

Those payouts do quiet work. A holder halfway through Annual notice earns at roughly 3x instead of 5x, and their own shrinking payout tells them precisely what cancelling would recover. No dashboard needed: the payout stub is the advertisement.

The pull grows as the exit proceeds. Early in the window, cancelling recovers little, since weight has barely decayed. Near the end, the holder earns close to 1.0x, and one click restores the full 5x. The system argues hardest for return exactly when the holder is closest to leaving.

And the door swings one way faster than the other. Entry is instant at full weight. Exit is a priced, published wait. The pool moves fast in the direction that helps it and slow in the direction that hurts it.

The Raise You Get for Staying

Every notice filed raises everyone else's payout, automatically. Shares sum to 100%, so decaying weight on one side is growing yield on the other. Stayers are paid more during exactly the periods when others are leaving.

The sequencing matters. In a cliff system, the unlock and the sell pressure arrive in the same block: the story breaks and the supply hits together. Here the stayers' raise arrives during the notice window, before a single released token reaches the market, and the exit itself is smeared across the window at 1/W per day. Stayers are compensated in advance of the supply, in USDB, at a rate they can read off their own distributions.

The Thermostat

Put the two together and the pool behaves like a thermostat.

Suppose activity spikes: a partnership, a product release, any catalyst. Fees rise, so USDB distributions rise the same week. Holders sitting in the exit queue see the increase in their own decayed payouts and can read exactly what cancelling recovers. Cancels fire, each one restoring full weight instantly. Some holders buy more, and new positions enter at full weight immediately. The queue drains backwards into the pool, and the growth gets converted into committed weight while it is happening.

Run it the other way. Activity falls, holders file notice, and every filing raises the payout rate for whoever remains, while the actual token release trickles out at 1/W per day. Exits slow themselves down.

One honest limit: the thermostat regulates the pool's composition, not its inputs. If platform activity itself falls, the fee pot falls with it, and rising shares of a shrinking pot cushion the decline rather than cancel it. The ladder redistributes fees; it does not create them.

What a Lock Still Does Better

A hard lock gives the protocol horizon certainty: tokens locked to a date cannot move before that date, full stop. The notice ladder's worst case is every holder filing notice on the same day, after which the largest positions become fully liquid over 365 days, earning decayed weight the whole way. That is a slower, priced, pre-announced version of the same exit, not an impossibility of it.

The ladder also asks more of the reader. "Lock for a year, get 5x" fits in a sentence. Notice windows, linear release, weight decay, instant cancellation, and per-position clocks take a page to explain. This article is that page.

Why Not a Cliff

Notice could have been built differently. Two dials define the design space: when tokens release (all at once when the notice period ends, or gradually during it) and what weight does when notice is filed (stays full, drops to 1.0x immediately, or decays gradually). Six combinations. Each was tested against five questions: is filing costly enough that the warning stays credible, is there free optionality, does supply ever land on a date, does the incentive to cancel survive the whole window, and does the exit stay cheap enough that holders still commit in the first place.

The cliff with an immediate weight drop earns the detailed answer, because it wins one test outright: it keeps 100% of exiting tokens out of circulation for the full window, where gradual release averages 50%. It fails three others.

Stress synchronizes cliffs. In calm conditions filings stagger, so cliffs scatter across the calendar. A panic makes holders file in the same week, which schedules a coordinated supply release exactly one window later, visible on-chain. The market gets a front-runnable unlock calendar in the worst state of the world. That is the lock failure mode reappearing precisely when it matters most.

Sunk cost kills the comeback. Cancelling resets the notice clock. Under a cliff, a holder at day 300 of 365 has paid ten months of reduced earnings and received nothing; cancelling discards all of it, so late cancellation is irrational and the queue hardens into exits. Under gradual release, the holder keeps every token already released. Cancelling never wastes what was already paid for, and the door back stays real to the last day.

Markets break at spikes, not averages. Price impact scales with instantaneous flow. A cliff concentrates the deferred supply into single moments; gradual release produces a flow the market absorbs continuously. Gradual release gives up half the token-days of restraint and removes every spike. That is the trade, and the system takes it.

Gradual release with gradual weight decay is the only combination that passes all five tests. That is the live design, arrived at by pricing every alternative in the same space.

Final tier parameters are subject to change before TGE.

---

The Outcome Follows the Incentive

Walk the ladder as a holder and check each position. Committed and the platform is growing: stay, your rate is rising. Committed and others are leaving: stay, their decay is your raise. In the exit queue when good news lands: cancel, one click restores everything. Sitting on liquid tokens watching payouts climb: enter, at full weight, instantly. At every node, the selfish move and the pool's interest point the same direction.

That is the design test Munger's line implies. Not whether the system sounds committed, but whether each participant's next best move builds commitment. Show me the incentive: warning is paid for continuously, exits are priced and reversible, and staying pays more precisely when others leave. The outcome follows.