Timed vs permanent locks

XPower Banq supports two flavours of position lock. They look similar from the outside but behave differently in important ways.

Timed locks

A timed lock pins your principal for a specified duration, then expires automatically. The protocol stores up to 16 active timed locks per user, each in a quarterly slot. Each slot represents one calendar quarter (~91 days).

• Minimum term: any positive seconds offset; in practice locks land in the next slot at minimum (~3 months remaining of the current quarter).
• Maximum term: 15 slot-offsets from now, ≈ 45 months. The ring buffer's total horizon spans 16 slots ≈ 48 months, but the slot already holding the next-expiring lock can't be re-targeted — only the next 15 slots are addressable from now.
• Granularity: quarterly. A lock created at any time during a quarter expires at the end of that quarter (after the chosen number of slots).
• Expiry: automatic. No action required from you. Once a slot's epoch has passed, the locked tokens become liquid again.

You can have multiple timed locks at once — different amounts with different expiry dates. The 16-slot ring buffer means you can structure a "ladder" of locks expiring at staggered intervals.

Slot semantics

The ring buffer reuses slots after ~4 years. If you create a new lock that maps to a slot containing an already-expired lock, the protocol cleans up the old one automatically (a "self-healing" operation). You don't need to manually clear expired locks, but doing so via the Position's free(user) operation reclaims gas in subsequent operations.

Permanent locks

A permanent lock has no expiry. The principal is locked forever. You created it; you can never redeem it.

• Term: infinite.
• Storage: the protocol uses a single packed field for permanent locks rather than ring slots, so adding to a permanent lock is cheaper than adding to a timed lock.
• Cascade attenuation: the (1 − ϕ) attenuation in the cascade theorem applies only to the permanent-locked fraction. Timed locks contribute attenuation only until their expiry.

When to use which

Situation	Recommendation
You have a known holding period (e.g., locked employer tokens, a tax planning horizon)	Timed lock matched to your horizon
You want maximum bonus and don't care about flexibility	Permanent lock
You're a long-term APOW holder using XPower Banq as yield infrastructure	Probably a permanent lock or a multi-year timed lock
You're using XPower Banq tactically (leverage, arb)	Don't lock — keep your liquidity
You want some lock for the bonus, some for flexibility	Lock part of your position; leave the rest liquid

Bonus differences

The interest bonus/malus scales with your lock ratio — the fraction of your balance that's locked, weighted by remaining time. Permanent locks count at their full value forever; timed locks contribute proportionally to their remaining duration.

This means a permanent lock of 1 token earns more bonus than a 1-quarter timed lock of 1 token, integrated over time. See Bonus and malus for the formula.

Cascade-protection asymmetry

The Cascade Attenuation Theorem (in the protocol whitepaper) bounds cascade severity by (1 − ϕ), where ϕ is the permanent-locked fraction.

Timed locks attenuate cascades only while active. A timed lock that expires the day before a crash provides zero protection. Permanent locks always count.

If your goal is cascade protection — e.g., you're a major supplier who wants the pool to survive — a permanent lock is the only way to commit credibly.

Practical advice

• Most users supplying for yield should pick a timed lock with a multi-quarter horizon. The bonus is meaningful and the commitment is bounded.
• Borrowers should be conservative with locks. Locking your debt prevents early repayment, which can hurt if your circumstances change.
• Permanent locks are an ideological choice as much as a financial one. Don't make it lightly.

Where to go next

• Bonus and malus — the rate adjustments
• Transfers and exits — how to get out of a lock
• Locking positions — the practical UI/call walkthrough