Someone always sends me a wallet address around this time of quarter. A perp trader on Hyperliquid or dYdX who is up some absurd number, and the message is always the same. Can we just mirror this guy. And yeah, technically you can. You can watch their positions on-chain in near real time, you can size your own order to match the direction, and you can route it non-custodially so your keys never leave your control. The wiring is the easy part. The reason most people who try this end up flat or down while the whale is up is that copy-trading a perp wallet is a completely different game from copy-trading a spot buy-and-hold, and almost nothing about it is obvious from the outside.
Let me walk through what actually happens when you try to shadow one of these accounts, because the gap between the idea and the outcome is where all the money leaks out.
How mirroring works, and where the first leak is
Mechanically it is straightforward. Perp positions on the major venues are visible. You can subscribe to a wallet's fills and open positions through the exchange's websocket feed or an indexer, and the moment they open a long on some pair you get an event. Your bot reads that event, decides how much to put on, and fires your own order. Non-custodially this means you are signing your own transactions or authorizing your own orders against your own margin, so you are not handing funds to some pooled vault that promises to mimic a leader. You mirror the signal, not the capital.
The first leak shows up immediately. You are always late. The whale's order fills, then the feed propagates, then your logic runs, then your order hits the book. Even if all of that takes a couple of seconds, on a fast pair that is a real slippage tax, and it is worse than that because your entry is often nudging the price in the same direction they just pushed it. So you enter after them, at a worse price, into a market they already moved. On a trade that plays out over days this barely matters. On a scalp it can eat the whole edge before you have done anything wrong.
The whale can leave before you know they moved
This is the part that quietly wrecks people. A big perp account does not just enter and hold. They flip, they scale out, they hedge, they unwind a position in pieces so it does not show as one dramatic exit. And they can do all of that faster than your mirroring loop can see and react. You are reading a delayed picture of their book. By the time your system registers that they have started closing, they may already be flat or reversed, and you are holding a position they no longer want.
A few things I watch for when trying to read an unwind before the feed spells it out:
- Position size ticking down in small increments rather than one clean close. That is often deliberate scaling out, and it means the conviction is fading even though the position still looks open.
- Realized PnL on the wallet jumping while notional exposure shrinks. They are booking gains, not adding.
- A fresh opposite-side position on a correlated pair. Sometimes the "exit" is actually a hedge, and if you only mirror the original leg you are now naked to a move they have already neutralized for themselves.
- Funding flips going against the position. If the whale opened when funding paid them to hold and it has since inverted, their reason to stay may be gone even if the position is still open on your screen.
The uncomfortable truth is that a lot of a good perp trader's edge is in the exit. Knowing when the trade is done is the skill. If you copy only the entry and treat the position as something to hold until they close, you are copying the easy half and skipping the half that actually made them money.
Funding, leverage, and why your liquidation math is not theirs
While you hold, funding accrues. On a perp, the funding rate transfers money between longs and shorts every funding interval, and if you are on the crowded side you are bleeding a little every few hours just for being in the position. A whale might be fine paying that because they expect a move big enough to swamp it, or because they are actually short-term and out before it adds up. If you enter late and hold longer, you pay more funding on a smaller expected move. That drag is invisible on a screenshot of someone's win and very visible in your own balance a day later.
Leverage is the other trap, and it is a nasty one. If you match the whale's direction but not their leverage, your liquidation price is nothing like theirs. Say they are running 3x on a position they can babysit all day and top up with margin the instant it moves against them. If you copy the direction at 10x because you want the same dollar exposure on a smaller account, a wick that they shrug off liquidates you. You got stopped out of a trade that was, for them, still perfectly alive. You did not copy their position. You copied the direction and invented a much more fragile version of it.
Sizing is where you claw some of this back. I default to sizing down hard relative to the whale, both in leverage and as a fraction of my own account. The point of mirroring is not to feel the same thing they feel on every candle. It is to take a fraction of a position you have reason to believe is informed, at risk you can actually survive, so that a delayed exit or a wick does not end your run. If the whale is 3x, you are not going 10x to compensate for a smaller balance. You go lower and you accept smaller absolute wins.
Copy the entry, manage your own exit
After enough of these I landed on a rule that has held up. Treat the whale's entry as a signal worth acting on, and treat the exit as entirely your problem. You are borrowing their read on direction and timing of the setup, which is the part you can actually observe cleanly. You are not borrowing their exit, because by the time you can see it, it has often already happened, and because your funding, leverage, and entry price make the right exit for you different from the right exit for them anyway.
In practice that means every mirrored position gets its own stop and its own take-profit the moment it opens, set against your entry and your liquidation math, not theirs. If they are still in and you have hit your target, you take it. If they unwind and you are still in, that is a signal to reassess, not an instruction you are forced to follow at a lag. The whale gives you a reason to be in the trade. Getting out stays your job.
At Blockcircle we score whale-wallet signals like these alongside insider filings and prediction-market moves precisely because a raw address is not a strategy. Knowing who to follow is maybe a third of it. The rest is the boring risk plumbing, sizing down, owning your exit, respecting funding, and never letting a mismatched leverage number turn someone else's good trade into your liquidation. If you take one thing from watching these wallets, let it be that the entry is the part they are showing you, and the exit is the part you have to build yourself.