1. Choosing leaders on recent ROI
Recent return is the most available number and the least informative one. Over a short window it is dominated by leverage and by whether the trader's style happened to suit the regime. A leader who ran 20x into a trending week will top any seven-day ranking, and the same behaviour is what produces the eventual liquidation.
The correction is to select on properties that persist: realized PnL consistency across periods, profit factor rather than win rate alone, position discipline, and account survivability through drawdowns. Because Hyperliquid is on-chain, all of these are computable from public history rather than self-reported.
The practical test: would this leader still look good if you shifted the window back one month? If the answer depends on the window, you are looking at a regime, not an edge.
2. Trusting the raw leaderboard
Exchange leaderboards rank on a single metric over a fixed window, usually PnL or ROI. That construction rewards large accounts taking large risk, hides accounts that recovered from near-liquidation, and says nothing about how the returns were produced.
It also has a survivorship problem. Accounts that blew up leave the board, so the visible distribution is systematically better than the real one. Ranking within that filtered set feels like evidence and is not.
Use the leaderboard as a candidate generator, never as a decision. Every candidate needs a second pass against stated criteria before capital follows it. The research note on why raw leaderboards mislead works through the specific distortions.
3. Ignoring position netting
This is the most expensive silent error in multi-leader copying. An exchange account holds one net position per market. If leader A goes long ETH and leader B goes short ETH inside the same account, the exchange nets them. Your exposure moves toward zero while you have paid taker fees and crossed the spread on both legs.
The visible symptom is a portfolio that seems inert: leaders are trading, fees are accumulating, and equity barely moves in either direction. Attribution also becomes impossible, because there is no per-leader position left to attribute.
The fix is structural: one sub-account per mirrored leader, so both strategies survive intact and margin problems stay contained to the sleeve that caused them.
4. Undersizing capital for the number of leaders
Proportional mirroring only works when the resulting position clears the market's minimum order size. Spread a small account across many leaders and each sleeve becomes too small to size positions correctly: some orders round up, taking more risk than the model intends, and others are skipped, so your basket quietly diverges from the strategy you thought you were running.
Margin fragmentation compounds it. Each sub-account needs its own margin buffer, so capital sitting as collateral in one sleeve cannot support another sleeve's position. Below a certain account size, adding leaders reduces effective diversification rather than increasing it.
The correction is to match leader count to capital rather than maximising it. Start with fewer sleeves at a workable size and add breadth as the account grows. The capital requirements page covers the size bands in detail.
5. Over-leveraging the mirror
Mirroring a leader at higher leverage than they run is not a way to earn more from the same edge. It changes the position's liquidation distance, so a drawdown the leader survives comfortably can liquidate your copy of it. At that point you stop tracking the strategy entirely — you have realised the loss and you are not there for the recovery.
The related error is treating the leader's leverage as safe because it has not failed yet. A trader running high leverage successfully for months is not demonstrating that the leverage is safe; they are demonstrating that the failure has not happened during the observed window.
Apply your own leverage cap independently of the leader's, and size so that the worst historical drawdown of that strategy would not liquidate you.
6. Underestimating slippage, spread and funding
The headline fee is rarely the largest cost. Crossing the spread on entry and exit, slippage when the book is thin or the move is fast, and funding paid while holding a crowded perp position all subtract continuously and none of them appear on a fee schedule.
Funding in particular is easy to miss because it accrues quietly. A strategy holding the popular side of a heavily-skewed perp can pay meaningful funding over days, which turns a modestly profitable directional call into a flat or negative outcome.
Account for the full cost stack when you evaluate whether a strategy is worth mirroring: exchange fees, the platform fee, spread and slippage, funding, and the opportunity cost of margin sitting idle. The fees page breaks each component down.
7. Having no exit or revocation plan
People test the entry flow carefully and never test the exit. Then something goes wrong and they are learning the revocation process under stress, or discovering that exit requires the operator's cooperation.
In a non-custodial agent model, exit is a signature from your own wallet that revokes the approval; the funds do not move because they were never anywhere else. Test that path deliberately while nothing is wrong, so you know exactly what it takes.
Decide in advance what would make you leave — a drawdown threshold, a change in how the system behaves, a leader replacement policy you disagree with — and write it down before capital is at stake.
8. Treating automation as an excuse not to look
Automation removes the need to place orders. It does not remove the need to know what is happening in your account. Leaders decay, strategies drift away from what earned them their score, and market regimes change faster than any selection process can fully anticipate.
A weekly check is usually enough: which leaders are in the basket, how the weights have moved, whether realised results are diverging from the model more than usual, and whether anything about your own margin position has changed.
The opposite error is over-supervision — pausing after every losing day and resuming after every good one. That converts a systematic strategy into discretionary timing with worse information than the leaders have.
9. Sizing up after a good stretch
The most common capital-allocation error in the category: fund small, observe a strong month, then multiply the account. This reliably concentrates maximum exposure at the point where the recent evidence is most flattering and the regime is most likely to be near its favourable extreme.
The discipline is to decide your target size in advance, based on what you can afford to lose, and step toward it on a schedule rather than in response to returns. If you would not have added capital after a losing month, you should be sceptical of the impulse to add after a winning one.
Nothing here is financial advice. Perpetual futures are leveraged instruments: a position can be liquidated in full, and past performance of any trader is not indicative of future results. Copy trading does not remove that risk — it changes who makes the decision, not what the market can do to it.
