Most disappointment with automated trading comes from a mismatch between what a system promises and what it structurally is. This note removes the ambiguity by listing what HyperMirror deliberately does not do.
In short
HyperMirror is not a vault and never pools your capital: funds stay in your own Hyperliquid account under a trade-only agent approval that cannot withdraw or transfer. It does not promise alpha, publish forward-looking return figures, or guarantee that any leader keeps performing. It does not chase the leaderboard — the Elite basket is sticky, so a hotter wallet appearing elsewhere does not evict a current leader. It does not net your leaders against each other, because each one runs in its own isolated sub-account. And it does not remove market risk, leverage risk or liquidation risk; it bounds where those risks land.
It is not a vault, and it does not pool capital
A vault takes deposits. Your USDC leaves your control, joins other depositors' capital, and a manager trades the pool. You receive a share of the result, and you carry the operator's solvency and honesty as a risk on top of the market risk.
HyperMirror does none of that. Your capital stays in your own Hyperliquid account for the entire lifetime of the subscription. What you grant is an agent approval — a scoped execution permission that can place, modify and cancel orders on your account and nothing else. It cannot withdraw. It cannot transfer. It cannot move funds to another address, including ours. You revoke it on-chain whenever you want, and delegation ends at that moment.
The practical consequence is that there is no redemption queue, no lock-up, no gate, and no scenario where an operator failure strands your balance. The failure modes that remain are trading failure modes.
No deposits into a pooled account, ever.
Trade-only agent approval; withdrawals require your own signature.
Revocation is on-chain and immediate.
It does not promise alpha
Every number published on this site is a measurement of something that already happened, drawn from public Hyperliquid fill data. None of it is a forecast, and none of it is a commitment. Past performance is not indicative of future results — not as a legal formula, but as a description of how perpetual futures markets actually behave.
What the system claims is narrower and testable: it selects leaders on a documented composite of activity, realised performance, risk and survivability; it weights capital in proportion to that score; and it removes leaders according to a fixed, published rule set rather than a judgement call. Those are process claims. A process can be sound and still lose money in a given month.
If you are looking for a system that tells you what it will return, this is not it, and you should be suspicious of anything that does.
It does not chase the leaderboard
This is the point most often misread. The Elite Top 10 is a sticky basket. It is not reshuffled on every rank snapshot, and a wallet that suddenly posts a stronger score does not displace a current leader by being hotter.
The reason is anti-overrotation. Rank-chasing systematically buys the peak of a regime and sells the trough of a drawdown: you rotate into an account at the top of its cycle, and out of one closest to mean-reverting back. Doing that repeatedly also multiplies turnover, so you pay spread, fees and funding for the privilege of being late twice.
So score does two jobs, and only two: it ranks candidates for entry, and it weights how much capital each leader receives. Membership itself is decided by the replacement policy — emergency conditions that remove a leader immediately, and soft issues that must repeat across three separate UTC strike-days before a replacement happens. A clean evaluation day resets the strike counter to zero.
A higher score elsewhere alone never forces a replacement.
Emergency removal is immediate: account value below roughly $1,000, or no fill for 96h+ with zero trades in 7 days.
Soft issues accrue at most one strike per wallet per UTC day; three strike-days trigger replacement.
It does not net your leaders against each other
A single perpetual futures account holds one net position per market. Mirror two leaders with opposing views in one account and you hold neither view — you hold the difference, having paid fees on the full turnover of both.
HyperMirror avoids this structurally rather than cleverly: one isolated Hyperliquid sub-account per mirrored leader. A long in one sleeve and a short in another coexist. Margin is scoped per sleeve, per-leader PnL stays measurable, and replacing one leader is a local operation that leaves the other sleeves untouched.
What it does not do is make those positions cheaper. Isolation preserves exposure; it does not remove the cost of holding it.
It does not remove market risk
Perpetual futures are leveraged instruments. A leader can be liquidated. A market can gap through your stop-equivalent. Funding can turn against a carry position and grind a sleeve down without a single bad entry. Correlation can rise so that several leaders lose together in the same session, which is precisely when diversification helps least.
Isolation bounds where a liquidation lands — inside one sleeve, not across the whole book. Score-weighting bounds how much any one decayed edge costs. Strike-based replacement bounds how long a deteriorating leader stays in the basket. None of these bound the market itself.
There is also an honest cost to the deliberate pace of the soft policy: while a leader accumulates strikes, you continue to hold their sleeve. That is a chosen trade-off — cutting on the first weak day converts noise into realised loss — but it is a real cost, not a hidden benefit.
It does not do a handful of smaller things worth naming
For completeness, so nothing here is a surprise later.
It does not trade any venue other than Hyperliquid.
It does not give you manual control of individual mirrored orders — you control the delegation, not each fill.
It does not offer advice; nothing here is a recommendation about your capital.
It does not charge a management or performance fee. The only revenue is a 0.1% builder fee on mirrored notional volume, settled natively by Hyperliquid.
It does not unlock 10 leaders immediately: Starter mode mirrors one leader until $100,000 of mirrored volume, after which Full mode unlocks up to 10.
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.
Keep reading
How the diversified approach is implemented
If the structural argument above holds, the interesting question is the implementation: how leaders are scored, how weights are set and how replacement is triggered.
