If an account’s positions are public, it seems natural to ask what would happen if you held the same positions at your own size. A paper copy answers that question in simulation: it mirrors a tracked account’s positions, scaled down, using imaginary capital. It is a useful research tool if you know exactly what it models and what it leaves out, because the gap between a paper copy and a real follower’s account can be large.
This part of the series explains how SuperQuant’s paper copy sizes positions, which costs it charges, which real-world effects it does not model, and how to use the result as evidence about a trader rather than as a promise about your own returns. The scorecards guide covers whether the trader is any good; this guide covers how much of their result could reach a follower.
How the scale works
Copying a trader dollar for dollar is impossible for almost everyone, so a copy scales by account size. SuperQuant’s guide says each paper copy is sized to the trader’s account at its first check. In practice, the simulation divides your paper capital by the tracked account’s value at that check and uses the ratio, called the scale, for every later trade. In a fictional example with $10,000 of paper capital and a tracked account worth $2,000,000, the scale is 10,000 ÷ 2,000,000 = 0.005. When the tracked account adds 40 ETH, the copy adds 0.2 ETH.
Because the scale is set once, later deposits and withdrawals in the tracked account change your exposure. If the trader deposits enough to double their account and doubles every position, the copy doubles too, even though your paper capital did not change. Check the copy’s exposure relative to its own equity from time to time, not only its profit.
What the simulation charges
SuperQuant’s guide publishes the fill model. Copied accounts are checked about every two minutes, and every fill uses the mid price at that check plus a 0.045% taker fee and 0.05% slippage. The mid price is halfway between the best bid and the best ask. Slippage of 0.05%, or 5 basis points, means a simulated buy fills 0.05% above the mid and a simulated sell 0.05% below it. Trades under $10, or under 0.2% of the copy’s capital, are skipped, although a full close always goes through. No real orders are placed.
The 0.045% matches the base-tier perpetual taker rate on Hyperliquid’s fee page, which lists 0.015% for makers at the same tier. Hyperliquid says fees are based on your rolling 14 day volume and are assessed at the end of each day in UTC, so a real account’s rate depends on its own tier and order types. The guide to crypto trading fees shows how maker and taker rates combine with spread and slippage.
What the simulation does not model
- Detection delay. A copied trade is priced at the check that detects the change, not at the trader’s own fill. With checks about two minutes apart, the move in between can be a large part of the opportunity in a fast market.
- Queue position and partial fills. A real order competes with other orders, and a large one can walk the book, filling at progressively worse prices, or fill only partly. QuantConnect’s fill-model documentation makes a related point: in live trading, orders can partially fill, while its pre-built backtest fill models assume orders completely fill.
- Funding. Funding is the periodic payment between longs and shorts on a perpetual contract. The guide’s fill rules list a taker fee and slippage, not funding, so funding paid or received while holding a position is outside the simulation.
- Margin and liquidation of your copy. The published rules cover sizing and fills, not margin. The copy does not model your own margin mode, maintenance margin or a liquidation of your account.
- Skipped small changes. Because trades under $10 or 0.2% of capital are skipped, gradual scaling in or out by the trader can leave the copy slightly different from a true proportional mirror.
A worked comparison
The table uses fictional prices to show how a paper copy and the tracked trader can differ on the same idea, per unit, before funding.
| Step | Tracked trader | Paper copy |
|---|---|---|
| Entry price | 100.00 (own fill) | 100.40 mid at detection × 1.0005 = 100.45 |
| Exit price | 105.00 (own fill) | 104.60 mid at detection × 0.9995 = 104.55 |
| Gross move per unit | +5.00 | +4.10 |
| Taker fees at 0.045% each side | −0.09 | −0.09 |
| Net per unit | +4.91 | +4.01 |
Rounded to cents. The trader’s line assumes both of their fills paid the same taker rate; their actual fees depend on their own tier and order types.
The copy keeps about 82 percent of the trader’s net result in this example. Of the 0.90 gap, 0.80 comes from detection delay: the mid moved from the trader’s 100.00 fill to 100.40 before the entry was detected, and from 105.00 to 104.60 before the exit was detected. The other 0.10 comes from the modeled slippage, and fees are the same on both lines. A real follower acting on alerts would also face a detection delay, plus their own reaction time and the queue and liquidity effects the paper copy ignores. The guide to evaluating copy trading explains why a leader’s displayed record cannot show what a follower’s own fills would be.
Using a paper copy as research
SuperQuant’s guide starts a paper copy with /whales copy #12 10k, where #12 is a trader’s rank and 10k the paper capital, and checks on it with /whales copies. Without an amount, /whales copy uses $10,000 of paper capital. The guide lists 3 paper copies for Holder accounts and 10 for Pro. /whales copies shows each copy’s equity, return, number of fills and fees paid.
A paper copy is most useful when you ask it specific questions.
- How much of the trader’s result survives the delay? Compare the copy’s return with the trader’s own account change over the same period.
- Could you hold through it? Note the copy’s equity at regular intervals, then look at its largest drawdown (the biggest fall from a previous peak) and the longest stretch without a new high. A record that is good on paper can be hard to follow in practice.
- Is it better than a simple benchmark? Compare the copy with holding the same coins, or with not trading at all, over the same weeks.
- Is the sample long enough? A few weeks of results can be dominated by one or two large moves.
Boundaries
A paper copy is simulated only. It does not place orders, and its result is not evidence that a live follower would earn the same. Real copy-trading services are a regulatory subject in their own right. The European Securities and Markets Authority (ESMA) published a supervisory briefing in 2023 that includes guidance on the qualification of copy trading services as an investment service. It also covers information requirements, including costs and charges, and the qualifications of traders whose trades are being copied. The broader differences between simulated and live trading are covered in paper versus live crypto trading, and the effect of thin books on fills in the guide to spread, liquidity and slippage. This is educational research context, not investment advice.
The last part of this series, a whale-alert research routine, turns everything so far into a repeatable process for reviewing alerts.