Copy Trading Explained: When It Works and When It Doesn't

Copy trading is easy to describe and hard to do well. Someone else's account places trades; software replicates those trades into yours in proportion to the size you have allocated. You never read a chart, never set a stop, and never close a position by hand. The signal provider does all of that, and you pay a performance fee on the profits.

That pitch is accurate as far as it goes. What it leaves out is that your results depend less on the provider's win rate than on three things that never appear in a marketing screenshot: how correlated the drawdowns are, how much leverage the strategy is secretly running, and how much of the track record was selected after the fact. This article is about those three things, and about the narrower set of cases where copy trading genuinely earns its place.

Nothing here is advice to copy anyone in particular. It is a framework for deciding whether to copy at all, and a list of the failure modes you should be able to name before you allocate capital.

What copy trading actually is

Copy trading is a relationship between two accounts, usually at the same broker. One account, the master, or signal provider, trades normally. The other account, yours, the follower, receives a mirror of every position: same instrument, same direction, same entry and exit time, sized to your equity or balance rather than the master's.

Two mechanisms show up in practice, and the distinction matters:

• Copy trading links your account directly to a master on the same platform. Execution is automatic and immediate. You cannot see the master's reasoning, only the resulting orders.
• Signal following broadcasts trade alerts that each follower's account acts on. This introduces a delay and is sensitive to how fast your account receives and executes the alert.

Direct copy trading is the tighter of the two because there is no broadcast hop between the master's fill and yours. It is also the more opaque: you get the trades, not the thesis, and you are exposed to whatever the master does next.

The mechanics you actually control

Once you have chosen a provider, four decisions remain yours. They matter more than the choice of provider, because they determine how a losing streak translates into a loss on your account.

1. Allocation method. Fixed size copies every trade at a constant lot size regardless of your balance. Proportional-to-equity scales the copy so your risk stays a fixed fraction of your account value. Proportional-to-balance does the same against balance, which drifts less when you have open trades. The default you want is anything that keeps the copy proportional, fixed size quietly turns a drawdown into a much larger fraction of your account than it was on the master's.
2. Maximum allocation per provider. If you copy three providers at 100% each, you are not diversified. You have three copies of the same market exposure and triple the position size you thought you had.
3. Drawdown stops. Most platforms offer a soft stop (pause new trades at a set drawdown) and a hard stop (close everything at a set drawdown). Set them before you copy, not during a losing week. A live decision made at −15% is not a decision, it is a hope.
4. Exit behaviour. When you stop copying, you choose between closing the master's open positions in your account or letting them run unmanaged. Letting them run is how a copy account acquires positions nobody is watching.

When copy trading makes sense

There are real cases where copying beats trading your own account:

• You have no time or no system. If the alternative is discretionary trading you cannot supervise, a mechanical copy of a tested strategy is the lesser risk.
• You want exposure to a market or method you cannot build yourself. Futures, certain index strategies, or a market-hours routine that does not match your timezone.
• You are testing an allocation framework. Running a small copy alongside your own strategy is a legitimate way to learn how correlation behaves in your portfolio.
• The provider is an algorithm you can inspect. If the strategy is published, runs on a defined rule set, and its live track record matches its backtest, you are copying a process rather than a personality.

The four failure modes nobody advertises

These are the reasons copy accounts underperform the providers they follow. None of them is rare.

1. Drawdown correlation

Copying three providers feels like diversification. It usually is not. If all three trade EURUSD longs during US session momentum, their drawdowns arrive on the same day, from the same cause. Your combined account takes a hit roughly the size of the sum, not an average. Diversification in copy trading comes from uncorrelated exposure, not from a longer list of names. Before adding a provider, ask what it holds during a dollar rally and during a risk-off day, and compare that to the providers you already follow.

2. Hidden leverage

A provider can post a smooth equity curve and be carrying far more risk than the curve suggests. A strategy holding losing positions open across weeks, adding to a loser, or running many correlated pairs at high per-trade size can look stable right up to the point where the curve breaks. Two numbers expose most of it: the maximum lot size relative to the stated account, and the ratio of maximum drawdown to average monthly return. A return-to-drawdown ratio below about 1:3 means the strategy is surviving on timing, not edge.

3. Track-record selection bias

The marketplace you browse is a survivor list. Providers who blew up have been removed. Providers who are currently in a good quarter are promoted. The history you see has two biases baked in: the length of the record is chosen by whoever survived, and the ranking is sorted by recent performance, which is close to noise over short windows.

The practical defence is to ignore the default ranking and look at the second year of a record, not the last quarter. A strategy with 30 months live and a flat-ish curve is more informative than one with 4 months and a 40% gain, the second is usually variance dressed as skill.

4. Fee drag and copy slippage

You pay the provider's performance fee on gross profits, on top of your broker's spread and commission. A provider charging 25% on a strategy that returns 30% gross returns roughly 22.5% net before your own trading costs, and less after them. On thin-margin strategies, the fee is not a haircut, it is the difference between positive and negative expectation.

Copy execution is also not free of timing. Your fill happens after the master's, so fast strategies, scalping, news entries, lose a small amount on every trade. This cost is invisible on the master's statement because the master did not pay it. Slow strategies are far more robust to being copied.

How copy trading interacts with algorithmic strategies

Most providers worth following now run Expert Advisors rather than manual discretion, and that changes the calculus in two directions.

On the positive side, an algorithmic provider has a documented rule set. It does not revenge trade, it does not widen a stop because it "feels" like a reversal, and it will keep trading through a drawdown because the code says to. If you were going to copy anyone, a system that executes consistently is the safer counterparty to your capital.

On the negative side, algorithms fail differently from humans. A model trained on one market regime degrades silently when the regime changes. It does not sit out the chop the way a discretionary trader would, it keeps firing signals that no longer have edge, and the equity curve decays before anyone notices. When you copy an EA, you are exposed to its model risk as well as its market risk, and you will not see the warning until the drawdown is already real.

That is not an argument against algorithmic providers. It is an argument for watching correlation between the strategy's live results and its backtest: when live results diverge from the backtest distribution, the model has stopped describing the market.

Judging a track record honestly

When you open a provider's page, work through this list before you look at the return figure:

• Record length. Under 12 months is a sample, not a history. Under 6 months is noise.
• Maximum drawdown. Compare it to average monthly return. If the worst month is several times the average gain, the strategy's risk is not being managed.
• Equity curve shape. A straight line up is a warning, not a feature. Real strategies have losing stretches; a curve with none is either very short or very levered.
• Trade count and frequency. A meaningful sample needs hundreds of trades. A provider with 40 trades and a good curve has not proven anything.
• Largest single position. Look for the biggest lot size and the pairs held simultaneously. This is the fastest way to spot hidden leverage.
• Fee structure. Performance fee, and whether the provider's own capital is on the line. Skin in the game is the only reason to trust that a provider will stop trading when the regime turns.

One more rule that saves more accounts than any metric: start with a small allocation and let the provider go through a losing month before you add anything. A provider's behaviour during a drawdown, whether it cuts size, keeps trading the same way, or doubles down, tells you more than a year of winning months.

What copy trading will not do

It will not make a losing strategy profitable, and it will not remove the risk of loss, it moves the decision of when to enter and exit from you to someone else. If a provider has no edge, copying it converts that absence of edge into real costs: spreads, commissions, performance fees, and copy slippage. The only thing copy trading reliably removes is the labour of trading. Whether that labour was worth removing is a question about your own habits, not about the provider's returns.