Copy trading is a mechanism, not a strategy. You link your account to another trader’s. When they open a position, software opens a proportional position in your account, sized to your balance rather than theirs. When they close, you close. Some platforms let you set a maximum loss on the relationship, and all of them let you stop at any time. That is the entire product.
What it does not do is transfer skill. It transfers risk — at your position size, against your balance, on your timeline. A strategy that a $500,000 account absorbs as a routine 8% drawdown can close out a $500 account entirely, because the same percentage move meets a different margin buffer. The person you copy does not know your balance, does not know what else you are holding, does not know when you need the money back, and owes you no duty of care. They are running their book. You are along for it.
This shortlist covers nine brokers whose copy or social trading offering appears in our own published testing. Every figure below comes from a review we wrote after funding a live account, not from a broker’s marketing page. Where our testing did not cover something — and on copy trading specifically, there is a lot it did not cover — this article says so rather than filling the gap. We hold no affiliate relationships with any broker named here, and our commercial arrangements are set out at how we make money.
Most retail traders lose money on leveraged products. Copying someone else does not change that, and in several documented ways it can make it worse.
What you are buying, and what you are not
Copy trading is the most heavily promoted product in retail trading, and the promotion works because the pitch is genuinely appealing: skip the learning curve, attach yourself to someone who already knows what they are doing. Five things get left out of that pitch, and all five matter more than the choice of platform.
Risk transfers; competence does not
The copied trader’s edge, if they have one, exists inside their own risk framework — their capital, their diversification, their tolerance for a bad quarter, their ability to stop. None of that copies across. What copies across is the position: the entry, the size as a proportion of equity, the holding period, and the drawdown. If their approach involves surviving a 40% peak-to-trough decline on the way to a good year, you inherit the 40% decline. Whether you can sit through it is a question about you, not about them.
Leaderboards are survivors, not populations
Every copy platform ranks its providers by past performance. Every ranking is therefore a list of the accounts that did not blow up. Traders who lost their capital stop appearing, because there is no account left to rank. The result is a display that systematically overstates how copy trading tends to go, without anyone having to publish a false number.
We have a concrete measurement of this. When we reviewed FXTM Invest for our FXTM assessment, we looked at the top 75 listed strategy managers. Every single one showed a positive return. A ranking in which one hundred percent of the visible population is profitable is not a description of trading; it is a description of a filter. Nothing about that is unique to FXTM — it is what every leaderboard does — but it is rare to have the count written down.
The costs stack
A copied trade carries the broker’s normal spread and commission, exactly as a manual trade does. On top of that, copy programmes commonly add a provider’s share of profits, and in some designs a markup applied to the spread on copied volume specifically. You pay the base cost of trading plus the cost of not making the decision yourself. The all-in figure is higher than trading the same positions manually — that is arithmetic, not opinion.
You are the client of record
The account is yours. The margin is yours. The losses are yours. If a copied position runs against you and triggers a stop-out, the broker closes your positions and it is your balance that carries the result. The provider loses their own money in their own account and, at worst, some reputation. There is no arrangement in retail copy trading under which the person you copied makes you whole.
Regulatory treatment varies by regime
Copy trading sits on a boundary. Where a client makes each decision, it is execution-only. Where the platform or provider is effectively making decisions on the client’s behalf, some regimes treat it as discretionary portfolio management, which requires permissions beyond a standard broking licence and brings additional conduct obligations with it. The same product can therefore be structured, disclosed and protected quite differently at two brokers with equally valid licences. Our regulation section covers what individual authorisations do and do not require.
The shortlist at a glance
Scores are our overall broker ratings out of 10, derived from the six weighted categories described in how we rate brokers. They are ratings of the broker as a whole, not of its copy product. A high score means the firm handled our money and our withdrawals well; it does not mean its leaderboard is a good place to pick a trader.
Two columns need a caveat before you read them. Our reviews are built around funding a live account and measuring spreads, fills, withdrawals and support. None of that testing captured a documented performance fee or a documented spread markup on copied volume at any of these nine brokers. Rather than reproduce figures we did not verify, those cells say so.
| Broker | Our score | Copy system | Minimum to copy | Performance fee | Spread markup on copied trades | Regulation tier |
|---|---|---|---|---|---|---|
| eToro | 7.8 | CopyTrader, proprietary | $200 per trader copied | Not documented in our testing | Not documented; base spreads were the widest we measured on every instrument | Tier one — FCA, CySEC, ASIC, plus ADGM and MAS |
| HFM | 7.9 | HFcopy | $25 to follow a strategy | Not documented in our testing | Not documented; HFcopy accounts are commission-free with cost in the spread | Tier one — FCA, CySEC, DFSA; no ASIC licence |
| AvaTrade | 7.6 | AvaSocial, separate app | Not documented; we allocated $200 | Not documented in our testing | Not documented; fixed spreads from 0.9 pips on EUR/USD | Tier one — Central Bank of Ireland, ASIC, Japan FSA; closed to new UK retail |
| FxPro | 7.6 | cTrader copy, built into the platform | Not documented in our testing | Not documented in our testing | Not documented; Standard from 1.2 pips, Raw+ from 0.0 plus $3.50 per lot per side | Tier one — FCA, CySEC; no ASIC licence |
| FXTM | 7.1 | FXTM Invest | Not documented in our testing | Not documented in our testing | Not documented; Advantage from 0.0 pips plus volume-tiered commission | Mixed — FCA licence being surrendered; Mauritius FSC is client-facing |
| RoboForex | 6.5 | CopyFX | $100 | Not documented in our testing | Not documented; ECN and Prime from 0.0 pips | Offshore only — Belize FSC; no UK, EU or Australian access |
| FBS | 6.5 | FBS CopyTrade, separate app | Not documented in our testing | Not documented in our testing | Not documented; 0.5 pips advertised on Standard, 1.3 pips measured | Mixed — CySEC and ASIC entities; most clients onboarded in Belize |
| Deriv | 6.4 | Deriv cTrader copy | Not documented in our testing | Not documented in our testing | Not documented; 0.6 pips measured on EUR/USD | Mixed — MFSA for EEA; most clients elsewhere routed to an unlicensed SVG entity |
| LiteFinance | 5.1 | Built into the LiteFinance app | Not documented in our testing | Not documented in our testing | Not documented; 1.8 pips advertised on Classic, 2.0 pips measured | Weak — CySEC for EEA, Mauritius FSC, unlicensed St Vincent entity; no FCA, no ASIC |
1. eToro — the category leader, and the most expensive way to trade, at 7.8/10
eToro is the broker most people meet first, and usually they meet it through CopyTrader rather than through a spread table. Our eToro review found the copy product to be the strongest implementation in the field, and the trading costs underneath it to be the weakest of any broker we measured.
CopyTrader lets you filter traders by risk score, historical return, copier count and asset mix, and each profile exposes win rate, maximum drawdown and average trade duration. That last set matters more than the filters: drawdown and holding period tell you what copying will feel like, where return alone tells you only what already happened. Copying starts at $200 per trader, you can run up to 100 relationships at once, and you can set a stop-loss across an entire copy relationship rather than trade by trade. In our testing, copied positions appeared within seconds and adjustments propagated automatically.
The cost stack. Base spreads are the problem. We measured EUR/USD at 1.0 pip against 0.8 at the specialists, and on commodities the gap was wider still. Because a copied trade pays that spread on every entry and exit the provider makes, an active provider compounds eToro’s cost disadvantage against you far faster than it would if you were placing a handful of trades yourself. There is a $5 withdrawal fee on USD accounts, waived at Platinum tier and above and not charged on GBP, EUR, AUD or DKK accounts, and the inactivity fee was removed entirely in May 2026.
Who it suits. Someone who wants breadth across shares, forex, indices, commodities and crypto in one place, who values real share ownership alongside CFDs, and who is copying at low to moderate frequency where the spread disadvantage stays small in absolute terms.
The drawback. There is no MetaTrader and no Expert Advisor support, so if you ever want to move from copying to running your own systematic approach, you cannot do it here. Live chat is gated behind a $5,000 account balance, which means the traders most likely to need help — the small ones — are the ones who cannot reach a human quickly. eToro disclosed a 51% retail loss rate at our July 2026 check, the lowest on this list, though other eToro entities have published figures up to 61%.
2. HFM — the lowest entry price to copy anything, at 7.9/10
HFM carries the highest overall score of any broker on this list, and HFcopy has the lowest barrier to entry we found anywhere: $25 to follow a strategy, against $200 at eToro and $100 at RoboForex. Our HFM review rated the firm 7.9/10 on the strength of its regulation and its withdrawals, with trading costs as the category that held it back.
HFcopy sits alongside MT4 and MT5 as an account type in its own right, administered from the same mobile login, so you can hold a manual account and a copy account under one client profile. Becoming a strategy provider yourself requires $100 to $500 depending on the tier.
The cost stack. HFcopy accounts are commission-free with the cost carried in the spread, and Premium EUR/USD starts from 1.2 pips — above average for the sector and the single reason HFM’s trading-costs category scored where it did. A $25 entry point into a wide-spread account is a genuinely mixed proposition: the door is cheap, the room is not. There is also an escalating inactivity fee of $5 per month from six months, rising to $10 after a year and by a further $10 each additional year, which is exactly the kind of charge that erodes a small copy account you stopped watching.
Who it suits. Someone who wants to see how copying behaves before committing meaningful capital, and who would rather risk $25 finding that out than $200.
The drawback. HFM holds no ASIC licence and does not accept Australian residents at all. The Android app carried 3.68 out of 5 from roughly 8,700 reviews when we tested — below what the better broker apps manage, and the app is where HFcopy is actually administered. Clients routed to HF Markets (SV) Ltd should note that the St Vincent entity holds a company registration rather than a financial services licence, so no regulator stands behind that account.
3. AvaTrade — predictable costs, punishing dormancy, at 7.6/10
AvaSocial runs as a separate app from AvaTradeGO, letting you filter signal providers by performance, risk level and asset class and then allocate capital to them. Our AvaTrade review tested it directly: we allocated $200 to a trader showing a 62% win rate over six months and tracked it for a fortnight, and the allocation ended marginally ahead.
We report that result only to describe the mechanism. A fortnight is not a sample, a marginal gain is not evidence of an edge, and a 62% win rate over six months tells you nothing about the size of the losses in the other 38%. Read it as confirmation that the plumbing works, and nothing else.
The cost stack. AvaTrade’s fixed spreads are the distinctive feature — 0.9 pips on EUR/USD, held near their advertised level through news events when variable-spread rivals widened sharply. For copying, that has real value: you can calculate the cost of a copied trade before it happens, which you cannot do at a variable-spread broker when your provider trades the release. The minimum deposit is $100, with no deposit or withdrawal fees charged by the broker.
Who it suits. Someone who wants cost predictability over cost minimisation, and who is copying a provider that trades around economic releases where variable spreads would otherwise blow out the cost of every fill.
The drawback. The inactivity fee is the harshest on this list and it is aimed squarely at accounts like copy accounts: $50 after three consecutive months of inactivity, recurring quarterly, then $100 at twelve months. If you allocate to a provider who goes quiet, the dormancy charge starts eating the balance. Execution averaged 660ms in our testing, which is slow, and slow fills on copied positions mean your entry price drifts from the provider’s. AvaTrade also no longer onboards new UK retail clients — the EU entity sits on the FCA register in supervised run-off, permitting only wind-down of existing UK business.
4. FxPro — copy trading inside a serious platform, at 7.6/10
FxPro takes a different approach: rather than building a proprietary social network, it offers cTrader, which has copy trading built into the platform alongside cAlgo for algorithmic strategies in C#. Our FxPro review found cTrader the best platform in the lineup on technical merit — cleaner interface than MetaTrader, stronger charting, more flexible order management.
The practical consequence is that copying is a feature of a professional trading environment rather than the whole product. You are not corralled into a social feed. You can copy, and you can also run your own orders and your own algorithms in the same window, which makes the transition away from copying much easier than at a proprietary platform.
The cost stack. Which account you copy on matters enormously here. Standard spreads start at 1.2 pips advertised and were the widest of the four brokers we measured directly against each other. Raw+ and cTrader start at 0.0 pips with $3.50 per lot per side, and cTrader is also priced at $35 per $1m of volume. Trading costs were FxPro’s weakest scored category by some distance. There is also a withdrawal fee of roughly 2% to 2.7% by rail if you withdraw without having traded, and an inactivity fee of $15 once then $5 monthly after six months.
Who it suits. Someone who wants to copy now but expects to trade manually or algorithmically later, and who will open a Raw+ or cTrader account rather than Standard.
The drawback. FxPro holds no ASIC licence, so Australian clients are routed offshore to the Bahamas entity with no compensation scheme behind them. The Elite tier got worse since our first testing: commission moved from $2.50 to $3.50 per side, matching Raw+ and replacing the discount with volume rebates. FxPro is private and unlisted, so no financial statements are publicly inspectable — a transparency gap that matters more, not less, when you are handing trading decisions to a third party inside their platform.
5. FXTM — the clearest illustration of leaderboard bias, at 7.1/10
FXTM Invest is a competent copy product attached to a broker whose regulatory position is moving in the wrong direction, and it is on this list partly as a worked example. Our FXTM review examined the strategy manager rankings and found that of the top 75 listed managers, every one showed a positive return.
FXTM describes the performance data as verified, and we have no reason to dispute the individual numbers. That is precisely the point worth understanding. Each figure can be accurate while the display as a whole is misleading, because the ranking shows survivors and the population that produced them is not shown. Anyone forming their first impression of copy trading from a screen where nobody has ever lost is forming it from a filter.
The cost stack. The Advantage account starts from 0.0 pips with a volume-tiered commission; we paid roughly $4 per lot round turn at our tested volume, but the rate varies with account equity and volume, so a small copy account will not get our rate. Minimum deposit is $200 across account types.
Who it suits. On costs alone, an active copier who can reach the higher volume tiers. On regulation, fewer people than it used to.
The drawback. FXTM announced in April 2026 that it is surrendering its UK FCA authorisation to focus on the UAE and Asia, and by our July 2026 check its own licensing page no longer listed the UK entity. Regulation scored 5.0 in our assessment and is the category that pulled the overall rating down. Maximum leverage on the Mauritius entity runs to 1:3000, which we treat as a hazard rather than a feature — leverage that high turns an ordinary provider drawdown into a margin call. The retail loss rate, published as 88% when the FCA disclosure applied, could not be verified at all at our last check.
6. RoboForex — the best provider disclosure, the worst licensing, at 6.5/10
CopyFX has the most informative ranking data we encountered. Each strategy provider’s page discloses return, maximum drawdown, trade count and trading history — more disclosure than several competing copy platforms provide, and the inclusion of maximum drawdown next to return is what separates a useful listing from a promotional one. Our RoboForex review subscribed to two strategies for a fortnight and found copying accurate, with minor deviation on some fill prices.
That fill deviation is worth noting generally: replication is proportional, not identical. Your entry is the price available when the copy order reaches the market, which is not the price the provider got. Over a high-frequency strategy those small differences accumulate against you.
The cost stack. Minimum participation is $100. ECN and Prime accounts start from 0.0 pips, the minimum deposit is $10, and there is no inactivity or dormancy fee on the current published terms. R MobileTrader handles CopyFX monitoring, deposits and withdrawals from the phone and rates 4.5/5 on Google Play across more than a million downloads.
Who it suits. Someone in a market RoboForex actually serves who wants to evaluate providers on drawdown rather than on return alone, and who accepts offshore licensing as the price of that.
The drawback. Licensing is the whole problem. RoboForex Ltd operates under a Belize FSC licence alone, and states it does not target EU, EEA or UK clients and does not work in the USA, Canada, Japan, Australia and other listed territories. For readers in the UK, the EU or Australia there is no retail route into this broker at all. Regulation scored 4.5 in our assessment, there is no statutory compensation scheme behind client funds, and RoboForex’s own current disclosure puts the retail loss rate at 75.85% — up sharply from the 58.42% recorded when the review was first written.
7. FBS — capped protection on a smooth product, at 6.5/10
FBS runs copy trading through a separate CopyTrade app. Our FBS review copied three traders for a week and found the mechanics smooth and one-click, with an automatic 10% stop-loss applied as a floor. Then we read the limits on it.
That stop-loss protection is capped at three positions across three traders. A protective feature with a ceiling is a good illustration of why the marketing copy on copy platforms should always be read down to the terms: the 10% floor is real, and it stops covering you exactly when your exposure grows past a very modest threshold. Search and filtering are also limited next to dedicated platforms such as eToro, which makes evaluating providers harder in the place where evaluation matters most.
The cost stack. FBS advertises 0.5 pips on the Standard account; we measured a 1.3 pip average on EUR/USD, which is a material gap when every copied entry and exit pays it. Account minimums run from $1 on Cent to $1,000 on ECN. There are no deposit, withdrawal or inactivity fees except a $15 charge on bank transfers under $200.
Who it suits. A beginner starting small who values fast fund access — our Skrill payout cleared in three hours with no fee, and live chat answered in two minutes at midnight, both genuinely strong results.
The drawback. FBS holds no FCA authorisation and does not serve UK retail clients. The CySEC and ASIC entities are real, but most clients are onboarded offshore in Belize, and new South African clients now go to the Belize entity rather than the FSCA-licensed one — so the licence list is longer than the protection actually delivered. Regulation and trading costs both scored 5.5. Published loss rates were 66.43% at the EU entity and 76.2% in Australia at our last check.
8. Deriv — copy trading attached to unusual instruments, at 6.4/10
Deriv added cTrader to its platform lineup, bringing copy trading and order book depth alongside Deriv MT5, Deriv Trader, Deriv Bot and Deriv GO. Our Deriv review found the broader offering distinctive, largely because of synthetic indices — instruments generated by Deriv’s own pricing model rather than by an underlying market.
That distinctiveness is a warning in a copy-trading context. If a provider you copy trades synthetics, you are not taking a position in a market that exists independently of your broker. The instrument, the price feed and your counterparty are the same organisation. That is disclosed and legal, and it is a materially different proposition from copying someone trading EUR/USD.
The cost stack. We measured EUR/USD at 0.6 pips, which is competitive, and the minimum deposit is $5. A withdrawal we requested cleared in 29 hours.
Who it suits. Someone specifically interested in synthetic indices who understands what they are, and who is onboarded to the Malta entity rather than the default.
The drawback. Most retail clients outside the EU and Malaysia are onboarded to Deriv (SVG) LLC, which holds no financial services licence at all. Not a weak licence — none. That finding is what moved Deriv’s regulation category to 5.0 and the overall score to 6.4. Support was inconsistent in testing: four minutes to a human on a weekday morning, twelve minutes via bot on a weekday evening, nobody reachable on a Friday evening, and no telephone line at any time. For a product where you may urgently need to stop a copy relationship, that is a real operational risk.
9. LiteFinance — well-built copying, thin protection, at 5.1/10
LiteFinance has the lowest score on this list and it is here because the copy implementation itself is good and the protection behind it is not — a combination readers should be able to recognise. Our LiteFinance review found copy trading genuinely built into the proprietary app rather than bolted on: you browse traders, review win rate, trade count and a risk score, and start or stop copying without leaving the interface. Replication was fast and proportionally accurate in our testing.
The cost stack. The Cent account opens at $10 with a $1,000 contract size instead of $100,000, which is a legitimate way to watch a copy relationship run with money that cannot do much damage. Classic requires $50 and advertises 1.8 pips; we measured 2.0 pips on EUR/USD, among the widest on this list. ECN is $50 from 0.0 pips plus commission from $0.25 per lot. The inactivity fee is $10 per 30 days after just 60 days of no activity — the shortest dormancy trigger here.
Who it suits. Narrowly: someone in a jurisdiction with an EEA route to the CySEC entity who wants cent-sized copy testing and understands what sits behind the account elsewhere.
The drawback. There is no FCA-authorised entity, so UK retail clients have no authorised route, and no ASIC entity, so Australian clients fall offshore. The default for most non-EEA clients is LiteFinance Global LLC in St Vincent and the Grenadines, which holds a company registration rather than a financial services licence. Regulation is the single largest contributor to the 5.1 rating. A card withdrawal took four days in our testing. Twenty years of continuous operation counts for something, but it is not a compensation scheme.
Four brokers we rate highly that are not on this list
Four of our better-scoring brokers are absent, and the reason is worth stating plainly rather than leaving as an implication. Our published testing of these firms did not cover a copy or social trading product. We are not going to describe an offering we did not open, fund and use, because a shortlist that pads itself with untested entries is worth nothing to the person reading it.
- Pepperstone, at 8.5/10 — the highest-scoring broker we have tested that is commonly associated with copy trading. Our review measured Razor pricing, roughly 30ms median execution across 50 orders and a six-hour PayPal withdrawal, but did not assess a copy product.
- Axi, at 8.1/10 — reviewed on its trading offering; no copy-trading assessment in our published testing.
- Eightcap, at 7.9/10 — same position.
- Admirals, at 7.5/10 — same position.
If you are considering copying at one of these firms, read the full review for the broker fundamentals — regulation, costs, withdrawals — and treat the copy product itself as unverified by us. The complete set is at our broker reviews.
The all-in cost of a copied trade
The single most useful thing you can do before copying anyone is to write down every layer of cost that will apply, and then ask what return the provider needs to generate before you break even. The layers are:
- The spread on every entry and exit. Paid by you, at your broker’s rates, on every trade the provider makes. This is the layer people most consistently underestimate, because it scales with the provider’s activity rather than with your decisions.
- Commission, where the account charges it. A raw-spread account with commission is often cheaper all-in than a wide-spread commission-free account for an active provider, and more expensive for an inactive one.
- Overnight financing on positions held past the close. A provider who holds for days is generating swap charges on your account continuously.
- The provider’s share of profits, where one applies. Commonly structured as a percentage of gains, sometimes with a high-water mark and sometimes without. Without a high-water mark, a provider can be paid on the recovery of losses you already absorbed.
- Any markup applied to copied volume specifically. Some designs widen the spread on copy flow relative to manual flow.
- Inactivity and dormancy fees. Copy accounts are unusually exposed to these, because the whole appeal is not having to watch. AvaTrade charges $50 after three months and repeats it quarterly; LiteFinance starts at 60 days; HFM escalates year on year.
- Slippage between the provider’s fill and yours. Not a fee, but a cost. We observed minor fill deviation when copying at RoboForex, and 660ms average execution at AvaTrade means a measurable gap on fast-moving entries.
We could not verify a published performance-fee schedule or a copy-specific spread markup at any of the nine brokers here from our own testing. That is a finding in itself. When the headline cost of a product is easy to find and the total cost is not, the gap between them is where the margin lives. Before you allocate, ask the broker in writing for the complete fee schedule that applies to a copied trade, and keep the answer.
How to read a copy-trading leaderboard
Leaderboards are designed to be scanned by return. Return is the least useful number on the page. Here is the order we would read them in.
Maximum drawdown before return
Maximum drawdown is the largest peak-to-trough fall the account has suffered. It is the closest thing to a preview of what copying will feel like on a bad month, and it is the number you must be able to sit through without closing at the bottom. A provider showing 60% annual return with a 45% maximum drawdown is not a better proposition than one showing 15% with an 8% drawdown; it is a much more leveraged one, and leverage cuts both ways. If a platform does not show drawdown at all, that omission tells you what the ranking is for. CopyFX shows it. eToro shows it. Not everyone does.
Account age and trade count
A five-month record is noise. A trader can produce a spectacular five months by taking concentrated risk that has not yet been punished, and the shorter the record, the more likely you are looking at variance rather than method. Look for a record spanning at least a couple of years and several hundred trades, ideally including a period when the provider’s market went against them. Trade count matters as much as elapsed time: 40 trades over two years tells you almost nothing.
Position sizing and consistency
Look at the shape of the equity curve, not just its endpoint. Steady incremental gains punctuated by a small number of very large losses is the signature of a strategy that sells risk — it will look excellent right up until it does not. Conversely, a provider whose position size jumps around wildly is either changing their mind about conviction or chasing losses, and neither is something you want replicated proportionally into your account.
Win rate in context
A 90% win rate is not a good sign on its own; it is a fact that needs a companion number. A strategy that wins nine times out of ten and loses fifteen times its average win on the tenth is a losing strategy with excellent optics. Always pair win rate with average win against average loss. Where a platform shows win rate prominently and the win-loss ratio not at all, treat the ranking as marketing.
What the provider actually trades
Check the instrument concentration. A provider whose entire record comes from one instrument in one market regime has not demonstrated a method, only a fit. And check whether they trade instruments you understand — copying someone trading synthetic indices, exotic crosses or single-name share CFDs commits you to risks you may not have considered.
What can go wrong
These are the specific failure modes, rather than a general caution.
Martingale and grid strategies
A martingale approach doubles down after losses, and a grid approach adds positions as price moves against it. Both produce beautiful equity curves for long stretches, because most adverse moves eventually retrace. Both also carry an unbounded tail: when a move does not retrace, position size has already been escalated into it and the account is destroyed in a single episode. These strategies are heavily represented near the top of copy leaderboards precisely because the curve looks so good until the end. If a provider’s drawdowns are consistently shallow and their recoveries consistently immediate, look at whether position size increases after losses.
Correlated copying
Copying five providers feels like diversification. It usually is not. If four of them are short the dollar, you hold one concentrated position across four relationships, at four times the size you would have chosen deliberately, and you pay four sets of costs for the privilege. Before adding a provider, check what they actually hold against what your existing providers hold. Platforms do not aggregate this for you.
The provider disappearing mid-drawdown
Providers stop. They close their account, hit their own margin call, lose interest, or simply go quiet — and they are under no obligation to tell you first. If that happens while positions are open and underwater, you are left holding trades you did not choose, with no thesis for them and no exit plan, because the exit plan was the other person. Decide in advance what you will do if the provider goes silent with open positions, and know where the button is that closes them.
Leverage mismatch
Replication is proportional to balance, not to risk appetite or margin buffer. A provider running at 1:30 with a large account has room to absorb an adverse move that would margin-call a small account running the same proportional exposure at 1:500. FXTM offers up to 1:3000 on its Mauritius entity and FBS up to 1:3000 on some accounts. High leverage does not make copying more profitable; it makes the same drawdown terminal at a smaller adverse move.
Protections with ceilings
Read the limits on any safety feature. FBS applies an automatic 10% stop-loss on copied positions, capped at three positions across three traders. That is a real protection and a real ceiling, and knowing where the ceiling sits is the difference between being protected and believing you are. Negative balance protection is a related question with a different answer at every entity — our guide to negative balance protection covers who actually provides it.
The broker underneath the platform
A copy relationship is only as safe as the entity holding your money. Several brokers on this list default non-EEA clients to entities with no financial services licence — Deriv to an SVG company, LiteFinance to an SVG company, and RoboForex to a Belize licence with no statutory compensation scheme. If that entity fails, no leaderboard performance matters. Read our guidance on how client funds are actually protected before you fund anything, and on checking a firm is what it claims to be — copy trading is a favourite theme of cloned-firm operations, and several brokers here have live impersonation warnings against their names.
How regulators treat copy trading
There is no single answer, which is itself the point worth taking away. The same product carries different obligations depending on where the entity is licensed and how the arrangement is structured.
Where the client selects each provider and the platform executes, the activity generally sits within ordinary execution-only broking permissions. Where the platform or the provider exercises discretion over the client’s account, regimes may classify it as discretionary portfolio management, which requires additional permissions, brings suitability obligations, and changes what the firm must tell you before you commit. Automated replication sits uncomfortably between the two, and firms structure their terms specifically to determine which side of the line they land on. Those terms are worth reading before the leaderboard is.
Practically, the protections that matter attach to the entity you contract with, not to the brand. The FCA brings FSCS cover to £85,000 and a leverage cap of 1:30 for retail clients. CySEC brings Investor Compensation Fund cover to €20,000 and the same leverage cap. ASIC brings conduct obligations and a leverage cap but no compensation scheme of that kind. Offshore entities in Belize, St Vincent, Vanuatu or the Bahamas typically bring none of it, and several brokers here route the majority of their clients to exactly those entities regardless of how impressive the licence list on the homepage looks.
Before copying at any broker, confirm which legal entity your account will be opened with, what compensation scheme covers it, and what the leverage cap is. Those three answers change the risk of copying far more than the choice of provider does. A directory of the firms we have assessed sits at our brokers section.
The risk, stated plainly
Most retail clients lose money trading leveraged products. That is not a disclaimer we are obliged to reprint; it is the measured outcome the brokers themselves publish. Among the firms on this list, the disclosed retail loss rates we recorded were 51% at eToro, 66.43% at the FBS EU entity and 76.2% at FBS Australia, 71% at AvaTrade’s EU entity, 76% at FxPro and 75.85% at RoboForex. FXTM’s figure, once published at 88%, could not be verified at our last check.
Copying someone else does not move you out of that population. You are in it, using the same instruments, the same leverage and the same broker, with an additional layer of cost and an additional party whose decisions you do not control. There is a reasonable argument that copying amplifies the problem: it removes the friction that would otherwise make a beginner hesitate, it encourages position sizes chosen by someone with a different balance, and it makes it psychologically easier to keep an underwater position open because the decision to hold was not yours to defend.
Nothing in this article is investment advice, a recommendation to copy any trader, or a suggestion that any outcome is likely. Past performance of a strategy provider does not indicate future results, and no ranking, risk score or verified badge changes that. We assess brokers. We do not forecast what will happen to your money, and neither can anyone selling you a copy product.
If you do copy, treat the first allocation as the cost of learning how the mechanism behaves rather than as an investment, size it at an amount whose total loss would not affect you, and read the drawdown column before the return column every single time.
Frequently asked questions
Is copy trading a good way for a beginner to start?
It is a common way, which is not the same thing. Copying lets you take leveraged positions before you understand leverage, which is the specific risk. It removes the requirement to form a view but not the requirement to survive a drawdown, and surviving drawdowns is the part beginners find hardest. If you copy as a beginner, the useful version is a small allocation treated as an education in how positions and margin actually behave, alongside reading the broker’s terms rather than the leaderboard.
How much do I need to start copying?
The documented minimums among the brokers we tested were $25 to follow a strategy at HFM, $100 at RoboForex CopyFX and $200 per trader copied at eToro. Minimums at the other platforms on this list were not documented in our testing. The more useful question is not the minimum but the amount at which the position sizes being replicated into your account are proportionate to a balance that can absorb a bad month.
What does copy trading actually cost?
More than trading yourself. You pay the broker’s normal spread and commission on every trade the provider makes, plus overnight financing on anything held past the close, plus any share of profits the provider takes, plus any markup applied to copied volume, plus inactivity fees if the account goes quiet. We were unable to verify published performance-fee schedules at any of the nine brokers here, so request the full schedule from the broker in writing before you allocate.
Is a trader with a 90% win rate safe to copy?
Not on that number alone. A high win rate is fully compatible with a losing strategy if the occasional losses are far larger than the frequent wins, and strategies built that way are common near the top of copy rankings because they look excellent until the tail arrives. Pair win rate with average win against average loss, and with maximum drawdown. If the platform shows the first and not the others, that is information about the platform.
Can I lose more than I deposit?
It depends on the entity holding your account, not on the fact that you were copying. Retail clients of FCA, CySEC and ASIC-regulated entities generally have negative balance protection, so the account floors at zero. Clients routed to offshore entities frequently do not, and several brokers on this list default most clients offshore. Confirm the position for your specific entity before funding rather than assuming it applies.
What happens if the trader I am copying stops trading?
You keep whatever positions are open, and they remain open until you close them or they hit a stop. Providers are not obliged to warn you before stopping, and the scenario that hurts is a provider going quiet while positions are underwater — you inherit trades with no thesis and no planned exit. Know in advance how to close a copy relationship and its open positions on your platform, and check that support is reachable at the hours you trade. At Deriv, for example, we found nobody reachable on a Friday evening and no telephone line at all.
Do I keep control of my own account?
Yes. You can stop copying and close positions at any time, and on most platforms you can cap the loss on a copy relationship — eToro allows a stop-loss across the whole relationship, and FBS applies an automatic 10% floor, though that one is capped at three positions across three traders. Retaining control also means retaining responsibility: you are the client of record, the margin is yours, and no platform makes you whole for a provider’s losses.
Is copy trading regulated?
The broker is regulated by whichever authority licenses the entity you contract with; the copy product’s treatment varies. Some regimes view automated replication as discretionary portfolio management requiring permissions beyond standard broking, others treat it as execution-only. The practical consequence is that two brokers can offer a similar-looking product under quite different obligations. Check the entity, the compensation scheme and the leverage cap rather than the brand.
Reader reviews
Traded with this broker? Tell others what actually happened.