The Copy-Trading Disappointment (Story)

[Fairman]

The Copy-Trading Disappointment (Story)

A composite story based on patterns many traders experience. Names are illustrative.

Imagine a man named Ikenna who joined a copy-trading platform. He picked a top-ranked trader with impressive returns over three months and set his account to copy automatically. He didn't need to learn anything, he thought. He could just watch the profits roll in.

For the first month, the results were great. In the second month, the copied trader opened a large position that went against him. The account dropped 35% in two days. Ikenna hadn't looked at the trader's maximum risk, average drawdown, or trading style.

What he discovered later is that the high returns came from aggressive leverage and occasional large positions held without stops. It was only a matter of time before the strategy hit a bad run.

He also learned that copying means your results depend on someone else's decisions, at slightly different prices, with different account sizes and risk tolerances.

Ikenna didn't swear off copy-trading forever, but he changed how he looked at it. If he ever tried again, he'd look for:

- A long track record, not a few months
- Low, consistent drawdowns
- Clear risk rules
- A small allocation he could afford to lose

Returns without risk information tell only half the story.

Whoever you follow, you remain responsible for the money.

[LondonNewsTrader]

A 35 percent drop in two days is the number that should have been predictable from the trader's history. If the three-month record was built with large positions and no stops, the worst drawdown hadn't happened yet, it was just waiting for the right market. Short track records almost never include a proper volatility event, a central bank surprise or a big data miss, and those are exactly the days that expose how someone sizes.

One more item for Ikenna's list: look at how the trader behaved around big scheduled releases. Did they hold large positions through payrolls and rate decisions? Did their equity curve show sudden jumps on those days? That tells you more about their risk than the headline return.

The point about different prices is also real. A copier on a different broker can get filled several pips away from the master account in fast markets, and over many trades that slippage adds up.

[Fairman]

Quick follow-up on the copy-trading story. Three months of returns is a tiny sample. Before copying anyone, look at the worst drawdown they've had and ask yourself if you could sit through double that without pulling the plug. If the answer's no, the copy will fail you even if the trader is decent.

[Fairman]

The other half of the lesson: Ikenna never knew why the trader he copied was entering. So when the drawdown came, he had nothing to judge it by except fear. If you copy, at least understand the style well enough to know what a normal bad week looks like for it. A trend follower might lose eight times in a row and still be fine. A grid or martingale style might look smooth for months and then lose most of the account in a week. Check the open trades history, not just the closed ones, because some of these strategies hide big floating losses that never show in the closed results.

[Shadow Trader]

The size difference between the two accounts causes its own problem. Copy platforms usually scale the master's position by account size, and the copier's result depends on how they handle the minimum lot.

Say the master runs 100,000 and opens 0.30 lot. Proportionally a 2,000 copier should get 0.006 lot, but the minimum is 0.01, so the copy is about 1.67 times the intended risk.

Some platforms round down instead and skip trades below the minimum. Then the copier misses part of the strategy and ends up with a different track record from the one that was chosen.

Either way the copier is not getting the trader's results. Checking a week of copied trades against the master's, lot by lot, shows which of the two is happening.

The 35 percent drop is also harder to undo than it sounds. Getting back to the starting balance needs about 54 percent, since 65 times 1.54 is roughly 100.

[Shadow Trader]

The point above about copiers getting different prices deserves a number, because the damage depends on what kind of trader is being copied. A fixed delay costs the same pips on every trade, so it eats a much bigger share of a short term edge.

Say a copier ends up 0.8 pip worse than the master on each round trip, from the delay plus a different broker's spread. If the master averages plus 2 pips a trade, the copy keeps only 1.2, so 40 percent of the edge is gone.

img_12766_r2.png (45.92 KiB) Viewed 20 times

The chart (illustration) runs the same 0.8 pip lag against average trades of 2, 5, 10, 20 and 40 pips. The share lost falls from 40 percent to 16, 8, 4 and then 2 percent.

So a top ranked scalper is close to the worst thing to copy, even if the record is honest. A slower trader with bigger average trades passes most of the result through.

Ikenna could have checked this before choosing. The master's average trade in pips and average hold time are often on the profile page, and dividing an estimated lag by that average shows how much of the edge would survive the copy.

[Shadow Trader]

Copy trading platforms rank providers by returns, and the top of the list looks irresistible. A trader up three hundred percent in six months. A smooth curve. Thousands of followers.

The catch is survivorship. Hundreds of providers started; those who blew up vanished from the list. The ones at the top may simply be the luckiest, or the ones taking the largest risks, and risk shows up as a drawdown later.

For scalpers, copying introduces a second problem: execution. The follower's fill comes after the provider's, usually at a worse price, with extra slippage. A provider's scalping edge may disappear entirely once a delay of a second or two is added.

When evaluating a provider, check the maximum drawdown, the trade frequency, the average holding time, and the length of the record. A high-frequency provider with a long record is rare and worth studying.

Would the numbers survive a one-second delay and an extra pip of slippage per trade?