Is Forex Trading Profitable in 2026? Why Most Retail Accounts Lose

Where the "90%" figure comes from

Ask where "90% of forex traders lose money" comes from and you will get plenty of confident answers and no source. The line circulates in trading courses, forum threads, short videos and — pointedly — in the marketing of people selling signals and courses that promise to put you in the other 10%. Sometimes the number is 80%, sometimes 95%. The digits change with the speaker; what stays constant is that no dataset, time period or definition comes attached.

The 90-90-90 rule

A popular version is the "90-90-90 rule": 90% of new traders lose 90% of their money within 90 days. It sticks because it is symmetrical, and the symmetry is itself a clue. Measured outcomes do not usually arrange themselves into three identical round numbers; slogans do.

Why it cannot be checked

A statistic is only useful if you can say exactly what it measures. The 90% claim fails every one of the basic tests:

• Who is counted? "Traders" could mean everyone who opened an account, everyone who funded one, or everyone who placed at least one trade. Each population gives a different answer.
• Over what period? Losing in the first 90 days, over a calendar year and over a lifetime are three different questions with three different answers.
• What counts as losing? An account that finishes the year $5 down and an account that is wiped out are both "losers" under a loose definition. Under a strict one — losing 90% of the money — only the second is.
• Where? Outcomes differ by broker, by product and by leverage regime. A figure with no jurisdiction attached cannot tell you anything about your own situation.

So the honest description of "90% of traders lose" is folklore: a number that may sit somewhere in the right neighbourhood, but that nobody can verify, falsify or update. On a subject where people commit real savings, that is not good enough — particularly when a measured alternative exists and is free to look up.

The number that is actually measured

In 2018 the European Securities and Markets Authority (ESMA) used its product-intervention powers to restrict how contracts for difference (CFDs) are sold to retail clients. For these rules, CFDs include the rolling spot forex contracts through which retail currency trading is done at EU and UK brokers. ESMA cited analyses by national regulators showing that a large majority of retail CFD accounts lost money. The restrictions applied from 1 August 2018; national regulators then made them permanent in their own rulebooks, and the UK's Financial Conduct Authority (FCA) brought its permanent version into force on 1 August 2019.

One of those measures is a standardised risk warning. Each provider must state what percentage of its own retail accounts lost money, recalculated every three months over the previous 12 months, in a sentence of fixed form ending "of retail investor accounts lose money when trading CFDs with this provider." That sentence is the closest thing to a measured answer that a retail trader can actually look up.

The "90% of traders lose" claim compared with the loss disclosure ESMA and FCA require.

Question	"90% of traders lose"	ESMA / FCA loss disclosure
Who produces it?	No traceable source	Each provider, under the regulator's rules
Who is counted?	Unspecified "traders"	That provider's retail CFD accounts
Over what period?	Unspecified, or "90 days"	The previous 12 months
How often updated?	Never	Every 3 months
What counts as losing?	Unspecified	The account lost money over the period
Can you check it?	No	Yes — it is printed on the provider's own website

Where to find it on a broker's website

The warning has to be prominent, so on an EU- or UK-regulated site it is rarely hard to find. Look for:

• a banner or fixed strip at the top or bottom of the homepage, often repeated on every page;
• the risk-warning box on account-opening pages and in the broker's adverts;
• the legal or "risk disclosure" section, which sometimes also gives the date of the latest calculation.

Then check that the figure belongs to the entity you would actually sign with. A group with several licences may show its EU or UK figure on a global website. If you are onboarded by a different entity of the same group, that number describes other people's accounts on a different rulebook, not the book you are joining.

We deliberately do not quote any broker's percentage on this page. The figures are recalculated every quarter, so a number printed here would go stale within months, and the provider's own current page is the authoritative copy.

What the percentage does not tell you

The disclosure is measured, but it is narrow. Read it for what it is:

• It is stated in accounts, not people. The fixed wording is "retail investor accounts", so it describes accounts rather than individual traders, and one person with several accounts may appear in it more than once.
• It covers one rolling 12-month window. It says nothing about whether this year's profitable accounts are profitable next year.
• A small loss and a blown account count the same. The size of losses and of gains is not disclosed.
• It covers all CFD trading with the provider — which may include indices, commodities and shares as well as currency pairs — not forex alone.
• It reflects the clients as much as the broker. A lower figure is not proof of better pricing; it may simply reflect clients who trade less often, use less leverage or have more experience.

Why your account may not carry one

The disclosure is an EU and UK requirement, and it comes bundled with a set of retail protections. If your account is held by an entity licensed somewhere else — as it is for many traders in Nigeria, Kenya, Ghana, Tanzania and Egypt — none of the rules below necessarily applies, and there may be no published loss figure for your entity at all.

Retail protections that apply at EU- and UK-regulated CFD providers (retail clients only; clients who qualify and opt in as professional give them up).

Measure	Rule for retail clients	What it limits
Leverage, major currency pairs	30:1 maximum (3.33% initial margin) on any pair made of two of USD, EUR, GBP, JPY, CAD and CHF	How large a position a given deposit can open
Leverage, non-major pairs and gold	20:1 maximum (5% initial margin)	Exposure on exotics such as USD/ZAR, and on gold
Leverage, other commodities	10:1 maximum (10% initial margin)	Exposure on oil and other non-gold commodities
Margin close-out	Positions closed when the funds in the account plus unrealised profit and loss fall to 50% of the initial margin required for the open positions	How deep a losing position can run before it is closed
Negative balance protection	Losses capped at the funds in the account	Owing the broker money after a gap
Incentives	No deposit bonuses or trading incentives	Bonuses that encourage larger or more frequent trading
Risk warning	Provider's % of losing retail accounts, 12-month basis, updated every 3 months	Marketing that hides the odds

Offshore entities set their own terms. Some brokers extend negative balance protection beyond the EU and UK — XM, for example, states in our data that it applies on every retail entity — but the leverage ceiling, the close-out level and the rest are whatever that entity's terms say. Before funding, confirm which entity holds your account and check its licence number on the regulator's own register; our guide on how to verify a broker's licence walks through it.

Cost drag: the arithmetic of spread and commission

Before costs, a CFD is a zero-sum contract: whatever one side gains, the other side loses. Add the spread and any commission, plus overnight financing that on most positions is a charge rather than a credit, and the traders' side becomes negative-sum. For a trader with no edge, the expected result of every trade is not zero — it is minus the cost. That is the first structural reason accounts lose, and it operates quietly, one trade at a time.

Start with what one round trip costs. On EUR/USD a pip is 0.0001 and a standard lot is 100,000 units, so a pip is worth $10 per standard lot and $1 per 0.1 lot. You pay the spread once per round trip — you buy at the ask and sell at the bid — and a commission, where there is one, on each side.

• The MetaTrader raw-spread accounts at IC Markets (Raw Spread MT4 and MT5) and Pepperstone (Razor MT4 and MT5) carry a commission of $3.50 per lot per side in our review data; the cTrader versions of both are $3.00. At $3.50, the round trip is $3.50 × 2 = $7 per standard lot, which is 0.7 pip on EUR/USD — before whatever the raw spread is at the moment you trade.
• The Standard accounts at IC Markets, Pepperstone, XM and Vantage list a spread of 1.0 pip with no commission in our data. Treat a listed spread as a starting point: what you actually pay moves with the pair, the hour and market conditions.

The examples below use two round figures: 0.7 pip, the raw-account commission alone, and 1.2 pips, an illustrative all-in cost a little above the listed figures. Both are assumptions for the arithmetic, not measured averages. Substitute what your own trade history shows; our guide to fixed, variable and ECN spreads explains how to work it out.

Monthly trading cost on a $1,000 account trading 0.1 lot EUR/USD ($1 per pip), at two illustrative round-trip costs.

Trades per month	Cost at 0.7 pip	Share of account	Cost at 1.2 pips	Share of account	12 months at 1.2 pips
20 (about 1 a day)	$14	1.4%	$24	2.4%	$288
60 (about 3 a day)	$42	4.2%	$72	7.2%	$864
100 (about 5 a day)	$70	7.0%	$120	12.0%	$1,440
200 (about 10 a day)	$140	14.0%	$240	24.0%	$2,880
400 (about 20 a day)	$280	28.0%	$480	48.0%	$5,760

The "a day" labels assume roughly 20 trading days in a month. At five trades a day and 1.2 pips, the trader pays $120 a month — 12% of the account — and, holding the same position size for a year, $1,440, which is 144% of the starting balance. Every one of those dollars has to be earned back by skill before the account shows a single dollar of profit. Put the other way, the strategy has to capture 1.2 pips per trade on average just to stand still.

Overtrading multiplies the bill

Cost is charged on every trade; edge, if you have one, exists only on the good ones. Each extra trade taken out of boredom, to "make back" a loss or to be in the market during a move you missed adds a full unit of cost and roughly zero expected gain. That is why overtrading hurts twice: it lowers the average quality of your trades and raises the bill at the same time.

Timing matters too. Spreads are widest when liquidity is thinnest — the hours after the New York close and the late-Friday drain — so the same trade can cost noticeably more in the hour after the New York close than during the London–New York overlap. Your platform's live spread shows the difference directly. Our guide to forex trading sessions sets out those windows in West African, South African, East African and Gulf time.

Short stops pay the most

Cost matters most when it is large relative to what you are risking. The tool for seeing this is the breakeven win rate. If your target is T pips, your stop is S pips and your round-trip cost is C pips, a winning trade nets T − C and a losing trade costs S + C. The strategy breaks even when p × (T − C) = (1 − p) × (S + C), which rearranges to:

Breakeven win rate p = (S + C) ÷ (T + S)

With the target equal to the stop and a 1.2-pip cost, the numbers look like this:

Breakeven win rate for a trade with target equal to stop and an illustrative 1.2-pip round-trip cost.

Stop and target (pips)	Cost as a share of the amount risked	Breakeven win rate	Typical style
5	24.0%	62.0%	Scalping
10	12.0%	56.0%	Fast intraday
20	6.0%	53.0%	Intraday
50	2.4%	51.2%	Day to swing
100	1.2%	50.6%	Swing

A scalper with a 5-pip stop and a 5-pip target must win 62 trades out of 100 just to break even: (5 + 1.2) ÷ (5 + 5) = 0.62. A swing trader using 100-pip stops and targets needs 50.6%. Same market, same broker, same cost per trade — very different hurdles. The fast, many-trades-a-day style can look like the quickest way to learn; it is also the style where the arithmetic is hardest.

Lopsided exits: the stop-and-target trap

The same formula exposes the third structural problem: how exits are managed. Before costs, a strategy whose average win is R times its average loss breaks even at a win rate of 1 ÷ (1 + R). A trader who takes profits at half the distance of the stop (R = 0.5) needs to be right two times in three. A trader who lets winners run to twice the stop (R = 2) can be wrong two times in three and still break even.

Breakeven win rate by reward-to-risk ratio, before costs and after an illustrative 1.2-pip cost on a 20-pip stop.

Reward : risk (R)	Target on a 20-pip stop	Breakeven before costs, 1 ÷ (1 + R)	Breakeven after 1.2-pip cost
0.5 : 1	10 pips	66.7%	70.7%
1 : 1	20 pips	50.0%	53.0%
1.5 : 1	30 pips	40.0%	42.4%
2 : 1	40 pips	33.3%	35.3%
3 : 1	60 pips	25.0%	26.5%

Now watch what happens when a plan meets a nervous trader. On paper the plan is a 20-pip stop and a 40-pip target — 2 : 1, needing about 35% winners after costs. In practice the trader grabs profits at +10 pips because they "might disappear", and moves the stop further away when price approaches it, so losses average 30 pips. The realised trade is now a 10-pip win against a 30-pip loss. The breakeven win rate becomes (30 + 1.2) ÷ (10 + 30) = 78%. Nothing about the market changed; the exits turned a workable plan into one that almost nobody can sustain.

This pattern — small wins, large losses — is a familiar shape in losing trade histories, and it has a name in behavioural finance, which we come to below.

Leverage: the same mistake at 1:30 and at 1:1000

Leverage does not change whether a trade is good. It changes how much of your account a price move is worth. The number that matters is not the maximum leverage the broker allows but the effective leverage you actually use: the notional value of your open positions divided by your equity. Our guide to leverage, margin and lot size covers the mechanics; the table shows the consequence.

Effect of a 1% adverse price move on a $1,000 account at different levels of effective leverage.

Effective leverage	Position on $1,000 equity	Loss on a 1% adverse move	Share of the account	Adverse move that equals the whole account
1:1	$1,000	$10	1%	100%
5:1	$5,000	$50	5%	20%
10:1	$10,000	$100	10%	10%
30:1	$30,000	$300	30%	3.33%
100:1	$100,000	$1,000	100%	1%
500:1	$500,000	$5,000	500% — the account is gone after a 0.2% move	0.2%

The last column is simply 1 ÷ leverage. At the 30:1 cap that applies to retail clients of EU and UK providers on major pairs, a fully levered account survives a 3.33% move. At 500:1 it survives 0.2%; at 1:1000, 0.1%; at 1:2000, 0.05%. On EUR/USD at an illustrative price of 1.1000, those are roughly 367 pips, 22 pips, 11 pips and 5.5 pips respectively. In practice the broker's margin-call and stop-out rules usually close positions before equity reaches zero (a price gap can jump past them), but the level at which they do so is set by each broker — within any regulatory minimum, such as the EU/UK 50% close-out rule — and differs between entities. Read it in your own account terms rather than assuming a standard figure.

A worked example in naira

A trader in Lagos funds an account with the naira equivalent of $1,000 — NGN 1,500,000 at an illustrative rate of 1 USD = 1,500 NGN; use today's rate for your own numbers. She opens one standard lot of USD/JPY. Because USD is the base currency, the notional value is exactly $100,000, so her effective leverage is 100:1.

• At 1:500 the margin required is $100,000 ÷ 500 = $200, a fifth of her equity. The platform accepts the order without complaint.
• At the 30:1 retail cap the margin would be $100,000 ÷ 30 = $3,333 — more than three times her balance. An EU- or UK-regulated retail account could not open this position at all.
• A 1% move against her costs about $1,000 (between $990 and $1,010, depending on direction, because the profit and loss on USD/JPY is in yen and converts back to dollars at the new price). That is the whole account. A stop-out closes the position first, but look at where. Stop-out is set as a margin level, meaning equity as a percentage of the margin in use, and the level differs by broker and entity. With only $200 of margin in use, even a stop-out level as high as 100% would close her out only when equity had fallen to $200. By then she has lost $800, or 80% of the NGN 1,500,000, after a single move of a size that major pairs do make.

Now size the trade by risk instead. Risking 1% of equity — $10, or NGN 15,000 at the same illustrative rate — with a 30-pip stop on EUR/USD, where a pip is $10 per standard lot: position = $10 ÷ (30 × $10) = 0.033 lots, rounded down to 0.03. That is €3,000 of EUR/USD, about $3,300 at an illustrative price of 1.1000. Leverage available: 1:500. Leverage used: about 3.3:1. The ceiling was never the problem; the position size was. Our guide to pips, lots and position sizing has the formula for every pair type.

Sizing a USD/ZAR trade in rand

USD/ZAR is priced in rand, so a pip (0.0001) is worth ZAR 10 per standard lot whatever the exchange rate. A trader in Johannesburg with a rand-denominated account of ZAR 20,000 who wants to risk 1% — ZAR 200 — with a 500-pip stop calculates: risk per standard lot = 500 × ZAR 10 = ZAR 5,000; position = ZAR 200 ÷ ZAR 5,000 = 0.04 lots. The stop is wide because USD/ZAR travels many more pips in a session than EUR/USD does.

The same arithmetic prices the spread. At 0.04 lots each pip is worth ZAR 0.40, so a spread of S pips costs 0.4 × S rand, and as a share of the amount risked it is simply S ÷ 500. Exotic spreads are much wider than those on the majors and move with the hour, so read yours off the platform before deciding the trade is worth taking.

Offshore entities and headline leverage

Brokers that serve Africa and the Gulf commonly do so through more than one entity, and the headline leverage figures that appear in their marketing come from outside the EU and UK rulebooks. These are the figures in FxReviewLab's own review data:

Headline maximum leverage and lower-tier licences for the six brokers we review, compared with the 30:1 EU/UK retail cap on major pairs.

Broker	Headline maximum leverage	Multiple of the 30:1 cap	Move that equals the whole account at full leverage	Tier-2 and Tier-3 licences in our data
Exness	Unlimited (Pro entity)	—	—	FSCA FSP 51024; FSA Seychelles SD025
HFM	1:2000	66.7×	0.05%	FSCA FSP 46632; FSA Seychelles SD015
IC Markets	1:1000	33.3×	0.1%	FSA Seychelles SD018
XM	1:1000	33.3×	0.1%	DFSA F003484; IFSC Belize 60/354/TS/19
Vantage	1:500	16.7×	0.2%	FSCA FSP 51268; CIMA 1383491; VFSC 700271
Pepperstone	1:500	16.7×	0.2%	DFSA F004356

Licence details verified 14 June 2026 against each regulator’s own public register — the same check recorded on our broker reviews.

A headline maximum is the most a group offers anywhere; the leverage on your own account depends on the entity, the instrument and, at many brokers, your account equity. South African traders have the option of FSCA-licensed entities at three of these brokers, and two hold DFSA licences in the Dubai International Financial Centre, which is relevant to traders in the UAE. Elsewhere, the entity that opens your account is often an offshore one. Our reviews of HFM and Exness list the entities and licence numbers we track. The regulator's own register is the final word on which of them holds your account.

High available leverage is not in itself why accounts lose. It removes the guard rail that would otherwise stop an oversized position from being opened — and oversized positions, combined with the habits below, are how small mistakes become blown accounts.

The behavioural traps

The structural problems are arithmetic. The behavioural ones explain why traders who understand the arithmetic still break it. Each of the four below is a well-established concept in behavioural finance, described in academic work over several decades. We describe them rather than attach numbers to them: the effects are well documented, but their size varies from study to study and from person to person, and no honest figure applies to you.

Loss aversion

Daniel Kahneman and Amos Tversky's prospect theory, published in 1979, described how people weigh losses more heavily than gains of the same size. For a trader this shows up as reluctance to accept a loss at all: moving a stop further away, removing it "just for now", or adding to a losing position to improve the average price. Each of these converts a small, planned loss into a larger, unplanned one — the exact behaviour the lopsided-exits example above turns into arithmetic.

System fix. Place the stop before you enter, based on where the trade idea is proved wrong, and never widen it. Moving a stop is allowed in one direction only: toward profit, and only after price has travelled at least 1R in your favour.

The disposition effect

Hersh Shefrin and Meir Statman named the disposition effect in 1985: the tendency to sell winners too early and ride losers too long. Locking in a small gain feels like success; closing a loss feels like admitting a mistake. The result is the trade history of small wins and large losses that no realistic win rate can rescue.

System fix. Set a take-profit and a stop-loss on every trade at entry, and let one of them close it. If you want to trail a winner, write the trailing rule down before the trade, not during it.

Revenge trading

Richard Thaler and Eric Johnson described a "break-even effect" in 1990: people who are behind become more willing to take risks that offer a chance of getting back to even. In trading it is called revenge trading — the next trade after a loss is taken bigger, faster and with less analysis, to "make it back". The logical endpoint is doubling after every loss, and the arithmetic of doubling is brutal:

Doubling the stake after each loss, starting at 1% risk, with the chance of such a streak appearing within 100 trades at a 50% win rate.

Consecutive losses	Risk on that trade (% of starting balance)	Cumulative loss	Chance of a streak this long or longer in 100 trades
3	4%	7%	Close to 100%
4	8%	15%	97.3%
5	16%	31%	81.0%
6	32%	63%	54.6%
7	64%, but only 37% is left	127% — the account is gone	31.8%

Cumulative loss is 2^N − 1 per cent after N losses. After six losses 63% is gone, so the seventh doubled stake of 64% cannot even be placed in full; the next loss ends the account either way. At a coin-flip win rate, a trader who doubles after every loss over 100 trades has better-than-even odds of meeting a six-loss streak that takes 63% of the account, and close to a one-in-three chance of a seven-loss streak that takes all of it. The probabilities come from the losing-streak formula in the next section.

System fix. A hard daily loss limit — for example 3R, three times your normal risk per trade. When it is hit, the platform closes for the day. No exceptions, and no "one more to get back to flat".

Overconfidence and fear of missing out

Brad Barber and Terrance Odean's studies of individual brokerage accounts in the United States linked frequent trading with worse net returns, and connected that frequent trading with overconfidence. Their data was on shares, not currencies, but the mechanism carries straight across: confident traders trade more, and every extra trade pays the cost in the table above. Its twin is fear of missing out — chasing a pair after a large move because the move is visible, when the setup that produced it has already paid out.

System fix. Cap risk per trade at 1–2% however good a setup looks — better setups justify patience, not bigger stakes. Write entry levels down before the session opens; if price never reaches your level, there is no trade.

Trading without a record

All four traps share one feature: they are invisible without records. A trader without a journal can repeat the same mistake hundreds of times without noticing it, because each instance feels like a special case. A one-line entry per trade — setup, entry, stop, target, result, one sentence on what happened — reviewed once a week turns the habits above from feelings into data you can act on.

Losing streaks and risk of ruin

Even a trader with a genuine edge, low costs and perfect discipline will have long losing streaks, and the streaks are longer than intuition expects. Knowing how long is what lets you choose a risk per trade that survives them.

The formula

If each trade loses with probability q, independently of the others, the chance that a particular run of N trades are all losers is:

P(N losses in a row) = q^N

At a 50% win rate that is 0.5⁶ = 1.56% for six in a row, which sounds rare. But you do not trade six times; you trade hundreds of times, and a streak can start at any point. The chance of at least one run of N or more losses somewhere in n trades is found with a short recursion. Let S(n) be the probability of no such run in n trades, with win probability p = 1 − q. Then S(n) = 1 for n < N, S(N) = 1 − q^N, and for n > N:

S(n) = S(n − 1) − p × q^N × S(n − N − 1)

The chance of meeting the streak is 1 − S(n). The table applies both formulas at three win rates.

Chance of consecutive losses: for one specific run at a 50% win rate, and for at least one such streak anywhere in 100 trades at 60%, 50% and 40% win rates.

Losing streak (N)	One specific run, 50% win rate (0.5N)	At least once in 100 trades, 60% win rate	At least once in 100 trades, 50% win rate	At least once in 100 trades, 40% win rate
4	6.25%	80.1%	97.3%	99.9%
5	3.13%	45.9%	81.0%	97.6%
6	1.56%	21.2%	54.6%	87.3%
7	0.78%	8.9%	31.8%	68.9%
8	0.39%	3.6%	17.0%	49.0%
10	0.10%	0.6%	4.4%	20.5%

Read the 40% column carefully. A 40% win rate sounds poor, but for a strategy that targets 2 : 1 it clears the 35.3% breakeven in the reward-to-risk table after costs, so it is profitable. Yet over 100 trades such a strategy has close to even odds of an eight-loss streak. Over the same 100 trades the expected longest losing streak is about 4.6 trades at a 60% win rate, 6.0 at 50% and 7.9 at 40%; over 250 trades those rise to about 5.6, 7.3 and 9.7. Plan for the streak, not the average.

What a streak does to the account

With a fixed percentage risked per trade, the balance after N straight losses is (1 − r)^N of where it started. The arithmetic is in percentages, so it works in any currency. The last column shows it for an account of KES 130,000.

Share of the account remaining after consecutive losses at a fixed risk per trade, with the balance of a KES 130,000 account after 10 losses.

Risk per trade (r)	After 5 losses	After 10 losses	After 20 losses	KES 130,000 account after 10 losses
1%	95.1%	90.4%	81.8%	KES 117,570
2%	90.4%	81.7%	66.8%	KES 106,219
5%	77.4%	59.9%	35.8%	KES 77,836
10%	59.0%	34.9%	12.2%	KES 45,328

A trader in Nairobi risking 1% per trade comes out of a ten-loss streak with about KES 117,570 and an intact plan. The same streak at 10% per trade leaves KES 45,328 — and the gain needed to recover is not symmetrical. After a drawdown D, the gain needed to get back to the start is D ÷ (1 − D):

Gain needed to recover from a drawdown, from the formula D ÷ (1 − D).

Drawdown	Gain needed to recover
10%	11.1%
20%	25.0%
30%	42.9%
50%	100%
75%	300%
90%	900%

Risk of ruin

The classic gambler's-ruin result gives the probability of eventually losing a whole bankroll. Assume each trade risks a fixed amount, wins or loses the same amount after costs, and wins with probability p greater than 50%. Starting with U units of risk, the chance of ever losing all of them is:

Risk of ruin = (q ÷ p)^U, where q = 1 − p

If p is 50% or less — no edge after costs — ruin is certain given enough trades. The table shows how quickly a small edge and a small stake change that.

Probability of eventually losing the whole starting stake, betting a fixed amount at even payoff after costs, by net win rate and stake size.

Net win rate (p)	10% of stake per trade (U = 10)	5% per trade (U = 20)	2% per trade (U = 50)	1% per trade (U = 100)
50% or less	100%	100%	100%	100%
52%	44.9%	20.2%	1.8%	0.03%
55%	13.4%	1.8%	under 0.01%	under 0.01%
60%	1.7%	0.03%	under 0.01%	under 0.01%

Two lessons follow. First, the edge that counts is the one after costs, which is why the cost-drag section matters. A trader who wins 52% of 1 : 1 trades on a 20-pip stop looks like they have an edge, but the 1.2-pip cost raises the breakeven to 53%, so after costs that trader has none. Second, stake size is the variable you control. A 52% edge risking 10% of the starting stake per trade has close to a 45% chance of eventually going bust; the same edge at 1% per trade has about 0.03%. The model is a simplification — real trades are not all the same size or perfectly independent, and traders who risk a percentage of the current balance shrink their stakes as they lose — but the direction of the result holds.

So, is forex trading profitable?

Yes, it can be — the disclosures themselves show a minority of retail accounts that did not lose money over the last 12 months. But three things are true at the same time:

1. For most retail accounts, over the period regulators measure, it is not profitable. When ESMA reviewed national regulators' data in 2018, 74–89% of retail CFD accounts typically lost money, with average losses per client of €1,600 to €29,000.
2. The structure is against you by default. The contract is zero-sum before costs and negative-sum after them, and leverage magnifies every mistake. A trader with no edge loses slowly through costs; a trader with no edge and high leverage loses quickly.
3. A good year is weaker evidence than it feels. The disclosure does not say how many of this year's profitable accounts stay profitable, and the streak tables show how long luck can run in either direction over 100 trades.

So the useful question is not "is forex trading profitable?" but "do I have a tested edge that is larger than my costs, and am I sized so that the losing streaks I will certainly have cannot end me?" If you cannot yet answer the first half with your own records, the answer to the original question, for you, is not yet.

What the traders who last do differently

None of the following guarantees a profit. Each one removes a specific, measurable way of losing:

• They know their cost per trade. They calculate spread plus commission on their own account and choose a trading style where the cost is a small share of the amount risked — the difference between a 62% and a 50.6% breakeven in the table above.
• They size by risk, not by margin. A fixed 1–2% of equity per trade keeps effective leverage low whatever ceiling the broker offers; our guide to the 2% rule explains the method.
• They define both exits before entry and never widen a stop, so the realised reward-to-risk matches the planned one.
• They cap the day, not just the trade. A daily loss limit such as 3R stops a bad morning becoming a revenge-trading afternoon.
• They trade less and at better hours — the liquid sessions, where spreads are tightest — rather than filling quiet hours with low-quality trades.
• They keep a journal and compute their real win rate and average R every month, then compare them with the breakeven formula instead of with how the month felt.
• They choose the entity, not just the brand. They know which company holds their money, what its licence is, and whether it offers negative balance protection.
• They expect the streak. They have looked at the losing-streak table, sized so that ten losses in a row is survivable, and decided in advance what they will do when it happens.

Frequently asked questions