Architecture
Every architectural choice has a bill. This one buys faithful exposure, measurable attribution and contained failure, and it pays for that with capital efficiency and an operational floor on account size. Both sides are worth stating.
In short
HyperMirror allocates one Hyperliquid sub-account per mirrored leader. Each sleeve holds its own collateral and its own positions, so leaders never net against each other, per-leader performance stays measurable, and a failing sleeve is bounded by what was allocated to it. The cost is that collateral is committed per sleeve rather than shared, which is why Full mode unlocks at $100,000 of mirrored volume.
| Dimension | Shared single account | One sub-account per leader |
|---|---|---|
| Capital efficiency | Higher — margin is shared | Lower — each sleeve carries its own buffer |
| Replication fidelity | Degraded by netting | Faithful per leader |
| Attribution | Blended | Per leader |
| Failure containment | Losses pressure every position | Bounded by the sleeve's collateral |
| Minimum practical size | Lower | Higher — hence Starter mode |
The three benefits are related but distinct, and each of them is something a single-account design cannot recover later.
Shared margin is efficient. Splitting collateral across sleeves means each sleeve carries its own buffer, so the same capital supports less gross exposure than it would in a single cross-margined account. That is a real trade-off and it is accepted deliberately: efficiency that comes from letting sleeves fund each other is exactly the coupling isolation exists to prevent.
The second cost is a practical floor. Ten sleeves out of a small balance produce positions too small to track their leaders faithfully, which is why Starter mode mirrors one leader well rather than ten badly.
Allocation is score-weighted: Capital weighted in proportion to each leader's composite score. Equal weighting treats a leader with three years of consistency the same as one that qualified last month, and score weighting exists to avoid that.
Weights change as scores change, and membership itself is sticky — a higher score elsewhere does not by itself displace a current leader. Selection and replacement rules are documented in full on How it works.
Isolation contains sleeve-level failure; it does not prevent simultaneous losses across sleeves, and it does not make leveraged perpetual positions safe. A basket of correlated leaders is still a correlated portfolio.
Past performance is not indicative of future results. Perpetual futures are leveraged instruments and carry a substantial risk of loss, including the loss of your entire position.
Methodology
Scoring and replacement are documented in full on How it works and in the Docs (Policy v3). In short: the Elite basket is sticky, emergencies remove a leader immediately, and soft issues accrue at most one strike per UTC day with three strike-days triggering replacement. Read how it works or the documentation for the full table.
Questions
Because the shared-margin efficiency comes precisely from sleeves being able to fund and offset each other, which is the coupling that makes a basket behave like one position.
No. The same capital is divided across sleeves by score. It does mean each sleeve is smaller, which is why very small balances mirror one leader instead of ten.
It is closed and the capital is reallocated to the replacement under the documented policy — no other sleeve is disturbed.
Score-weighted allocation across up to 10 elite Hyperliquid traders, each isolated in its own sub-account. Your funds never leave your account.
Non-custodial · Agent cannot withdraw · Cancel delegation anytime
