Why Most Forex Traders Lose Money: The Data, the Math and the Habits

00The short answer

• Most retail accounts do lose money, and the regulators have the numbers. When the EU restricted CFDs in 2018, its markets authority reported that 74% to 89% of retail accounts typically lose money. Brokers in the EU and the UK must still publish their own figure: on the day we wrote this, 70% and 74.1% at two large brokers.
• The main reason is arithmetic, not bad luck. Every trade pays a spread or a commission. A trader without an edge pays it on every trade and gets nothing back on average. In our simulation on real EUR/USD prices, a trader who picks entries at random with a 15-pip stop, 1% risk and 1 pip of cost ended the year with a loss 98.5% of the time; with a 5-pip stop, 100.0%.
• Frequency multiplies the cost. The scalper paid 97.2% of the starting account in costs in a year, the day trader 41.2%, the swing trader 2.7%.
• Leverage decides how fast. At 10% risk a trade, 99.5% of the day traders with 500:1 leverage lost 90% or more of the account in a year; the EU's 30:1 limit, which capped their size, held that to 65.5%.
• Real traders are not random, and they do better than our random frequent traders: the official figures are 70% to 89%, not 100%. Our simulation shows the baseline every trader has to beat. It cannot tell you how many traders beat it.

01Who we are

RIP Backtest is an independent site run by one retail futures trader. We test the strategies, indicators and claims that traders share, under one fixed protocol: rules written down and dated before any data is run, costs on every trade, a comparison with random entries, and every version reported, including the failures. So far that is 404 strategy variants in 59 tests, 2 of which passed every gate. We are not a broker and we have nothing to sell: no courses, signals, bots or managed accounts, no affiliate links, and no broker pays us. That is why we can write an article like this one, which most sites that earn from new traders have little reason to write. Every number from our own work is reproducible from the files at the end; every figure from someone else is quoted with its source and date. More about the site and our protocol.

02What the official numbers say

The regulators. On 27 March 2018 the European Securities and Markets Authority (ESMA) announced restrictions on contracts for difference (CFDs), the product most retail forex traders actually trade. Its press release cites the national regulators' analyses: 74-89% of retail accounts typically lose money, with average losses per client ranging from €1,600 to €29,000. The response was a set of rules that still apply in the EU and, in similar form, in the UK: leverage limited to 30:1 on major currency pairs, a margin close-out rule, negative-balance protection (in ESMA's words, investors “cannot lose more money than they put in”), and a standard risk warning in which every firm states its own share of losing accounts (ESMA press release, 27 March 2018).

The brokers. That warning is still on every EU and UK broker's website. On 14 October 2026 it read: “70% of retail investor accounts lose money when trading spread bets and CFDs with this provider” at IG in the UK, and “74.1% of retail investor accounts lose money when trading CFDs with this provider” at Pepperstone in the EU (IG UK, Pepperstone EU). We quote them because they are the industry's own figures, not to single them out: the warning is compulsory, and the figures at other regulated brokers are in the same range. Note what they count: accounts trading CFDs of all kinds (currencies, indices, shares, commodities), over a recent period, at one broker.

A study that points the other way. Not every dataset says that currency traders lose. Abbey and Doukas studied 428 accounts at an online currency broker from 2004 to 2009 and found that these traders, on average, earned positive abnormal returns even after transaction costs, with the more active traders doing better (Journal of International Money and Finance, 2015). The same authors found in another study that the use of technical analysis was associated with worse results (working paper). One sample of 428 accounts does not outweigh the regulators' data, but it is a fair reminder that “most lose” is not “everyone loses”, and that some traders do have skill.

03Read this first: what our simulation can and cannot tell you

04How we simulated a year of forex trading

We fixed every scenario in writing before running anything (fy_spec.md, frozen 14 October 2026):

• Prices: EUR/USD, the most traded currency pair, from the 1-minute history of a broker's MetaTrader 4 server, every calendar year from 2005 to 2025 (21 years), with the real moves, quiet nights, news spikes and weekend gaps.
• The trader: starts every year with $10,000, picks entry times and directions at random, and closes each trade at a stop or a target the same distance away (1:1). One trade at a time. Three styles: a scalper (5-pip stop and target, up to 20 attempts a day), a day trader (15 pips, up to 5 a day) and a swing trader (50 pips, at most 1 a day).
• Size: 1%, 5% or 10% of the current account at risk on every trade, limited by a maximum leverage of 30:1 (the EU and UK limit for major pairs) or 500:1 (common at offshore brokers).
• Costs: 1 pip per round trip (a low-cost account) or 2 pips (a higher-cost one), plus today's overnight swap on positions held past 17:00 New York. A version with no costs at all shows the result before costs.
• Scale: 1,000 simulated traders for each year and each of the 39 combinations: 819,000 trading years.
• Checks before any result: the trade exits matched a naive minute-by-minute scan, sizing and charges were checked by hand, and a separately written second implementation reproduced the trades and ending balances exactly (self-test result).

Share of simulated trading years that ended with a loss, at 1% risk a trade (30:1 leverage; the no-cost version at 500:1, where the limit never bound at 1%). 21,000 trading years per bar.

05A median year, in dollars

Before the reasons, the result. Three traders open an account with $10,000 on 1 January, risk 1% a trade, pay 1 pip per round trip and trade at random for a year, under the EU's 30:1 limit:

• The day trader takes about 833 trades. The median one ends the year with $5,334. On average the day traders paid $4,125 in spreads and commissions and $283 in overnight swap, and 98.5% of their years ended with a loss.
• The scalper takes about 2,799 trades and pays on average $9,718 in costs, nearly the whole account. The median scalper ends the year with about $50, too little to open even the smallest position.
• The swing trader takes about 143 trades, pays on average $273 in costs and $100 in swap, and the median one ends with $9,548: a small loss, the kind that is easy to blame on bad luck.

None of the three had an edge, and none was reckless: 1% risk, a stop on every trade, a regulated level of leverage. The difference between “lost half” and “lost a little” is almost entirely how often they traded and how small their stops were. The next sections take the reasons one at a time.

06Reason 1: you pay on every trade, win or lose

A trader without an edge wins as often as they lose, on average. The broker's spread or commission does not: it is charged on every trade. What matters is not the cost in pips but the cost as a share of what you risk. One pip on a 50-pip stop is 2% of your risk; on a 15-pip stop it is about 7%; on a 5-pip stop it is 20%. Every trade starts that far behind.

The scalper is the clearest case. With a 5-pip stop, each trade gives away 20% of what it risks, and after about 2,799 trades the account is gone: every one of the 21,000 simulated scalper years ended with a loss, with a median of -99.5%, even at 1% risk a trade. The day trader keeps about half the account and loses money in 98.5% of years; the swing trader, whose stops are ten times wider than the scalper's and who trades a twentieth as often, loses in 65.6% of years and gives up a median of -4.5%. Doubling the cost to 2 pips moves every style toward the scalper.

Work it out for your own account. Take a $10,000 account, 1% risk ($100) and a 15-pip stop on EUR/USD. One pip is worth $10 per standard lot, so the position is 0.66 lot and one pip of cost is $6.60 a trade. That looks harmless, but it is 6.6% of the $100 you risk, and at three trades a day it is about $5,000 a year at the starting size: half the account, paid whether the trades win or lose. Your own numbers are your spread plus commission in pips, your usual stop in pips and the number of trades you take in a year. If the first divided by the second, times your risk and times the third, is a large share of your account, the costs will take it before the market has a say.

07Reason 2: trading often multiplies the cost

The cost of a year of trading is roughly trades × (cost ÷ stop) × risk per trade. For the day trader: about 833 trades × 6.7% × 1% is more than half the account in expected cost drag, and the simulation, in which the account shrinks as it goes, charged 41.2% of the starting $10,000. The scalper's 2,799 trades at 20% of the risk each come to more than the whole account.

This is why the number of trades matters more than most beginners think. Each trade looks cheap: a pip, a few dollars. But a habit of five or ten trades a day turns a small fee into the largest single item in the account's year, larger than any winning streak a trader without an edge can expect. Year after year it was the same: for the day trader at 1% risk and 1 pip, the share of traders ending the year with a loss ranged from 96.8% to 99.6% in every year from 2005 to 2025, in calm markets and in crises. The market changed; the arithmetic did not.

08Reason 3: risk per trade and leverage decide how fast

Bigger positions do not change the odds of a trade; they change how fast the costs and the losing streaks reach the account. The same traders at 1%, 5% and 10% risk a trade, with the EU limit (30:1) and an offshore one (500:1), 1 pip of cost:

For the day trader, the 30:1 limit cut every trade above 1% risk to roughly 4% of the account, so the 5% and 10% versions behave alike, and 65.5% of them lost 90% or more. At 500:1 nothing stopped the 10% trader, and 99.5% lost 90% or more within the year. That is the limit doing exactly what ESMA designed it to do: it does not make a losing trader win, it slows the loss. A broker that offers 500:1 is offering the faster version.

Even 1% risk, the size most guides recommend, does not save a trader without an edge: it only spreads the loss over more trades. A small size buys time, which is worth a lot to a trader who is still learning, and nothing to one who will never have an edge.

09Reason 4: without an edge, even zero costs lose more often than they win

Take the costs away entirely, and the random traders still ended the year with a loss more often than not:

Three things push the median below zero. Compounding: risking a percentage of a changing account means a loss followed by a win of the same percentage leaves you slightly behind (lose 1% and win 1% and you have 99.99% of what you started with), and over thousands of trades that adds up. Swap: positions held past 17:00 New York pay the overnight financing, which on EUR/USD costs a long position more than a short one earns. Execution: when both the stop and the target fall inside the same minute we assumed the stop, as a real trader would often get. None of these is large on its own. Together they mean that a game without an edge is not a fair coin for your account; it is a slightly unfair one, before the broker has charged a cent.

The spread of outcomes matters too: at no cost, the 10th to 90th percentile of the day trader's year ran from -36.7% to 33.5%. Some random traders had good years. If you had one of them, you would probably believe you had found something. That is how luck recruits new traders.

10Reason 5: strategies that win often and lose big

Many strategies sold to beginners advertise a high win rate. A high win rate is easy to build: take profits quickly and give losing trades a lot of room, or none of the stop at all. The catch is the size of the losses.

• In our test of the classic RSI 30/70 rule on gold, the versions without a stop won 61% to 65% of their trades. Three of the four lost money, and single trades went 47 to 102 times the normal risk against the position before they closed.
• A popular YouTube strategy combining Stochastic, RSI and MACD was shown with a 56% win rate over 100 trades. On 22 years of EUR/USD in our test it won 40% to 42% of its trades, and every version lost money after costs.

The same logic applies to martingale and grid systems, which add to losing positions so that most sequences end with a small profit and a few end with a large loss. A win rate tells you how often you win, not how much. Ask for the average loss, the largest loss while a trade was open, and the result after costs, before you trust any percentage.

11Reason 6: holding costs and gaps

Positions held overnight pay or earn swap, the interest difference between the two currencies plus the broker's margin. On EUR/USD, at the rates we used, a long position pays $7.99 per lot per night and a short earns $1.31, and Wednesday counts three nights. For the swing trader at 1% risk the swap cost 1.0% of the account a year; at 10% risk, 8.6%. Swaps change with interest rates, and in some years they were much smaller or reversed, so treat these as an example, not a forecast.

Prices also gap. Over a weekend or a major news release, EUR/USD can open beyond your stop, and the stop fills at the first available price, not at the level you set. Our simulation used the real gaps in the data, so they are already in the numbers above. With small stops and high leverage, a single gap can take more than the planned risk; negative-balance protection, compulsory in the EU, stops it from taking more than the account.

12Reason 7: most strategies do not have an edge after costs

Random entries are a baseline. The real question is whether the strategies people actually use do better. We have tested many of the popular ones, with rules fixed in advance and costs on every trade. In our study of six common indicator strategies on the four major pairs (EUR/USD, GBP/USD, USD/JPY and AUD/USD, 2004 to 2026), 65 of the 96 versions made money before costs and only 30 after; none passed all our tests of whether the result could be told apart from luck. Our separate EUR/USD tests of the golden cross, the Fibonacci 61.8% pullback, the Turtle rules, Stochastic + RSI + MACD and the Ichimoku Cloud found the same: some versions looked good on a chart or in one period, none held up.

That does not prove that no strategy works. It shows that the famous ones, applied as they are usually taught, do not survive costs on their own. A trader who relies on them is, in effect, close to our random trader, with a story attached.

13Reason 8: habits that make it worse

The simulated traders never break their rules. Real traders do, in ways that research on individual investors has documented for decades:

• Holding losers and cutting winners (the disposition effect). Studies of brokerage accounts have found that individual investors sell winning positions more readily than losing ones (Odean, Journal of Finance, 1998). One study of a social trading platform found that access to the network nearly doubled the size of this effect (Heimer, Federal Reserve Bank of Cleveland, 2016). In forex terms: moving the stop further away, or closing a winner early because it might come back.
• Trading too much. Research on stock investors found that the households that traded most earned less than those that traded least, mainly because of costs (Barber and Odean, Journal of Finance, 2000). Section 07 shows why, in forex as well.
• Raising the size after losses, to win it back. In the simulation this is the step from 1% to 5% or 10% risk, and it is the step from a slow loss to a fast one.
• Trusting a good streak. Section 09 showed how often a trader with no edge has a good year by chance. A good month proves even less.

14Losses that have nothing to do with the market

Some money is lost before a single trade: paid signal groups, “guaranteed” trading robots, account managers who promise fixed monthly returns, and brokers that are not regulated where you live. We do not name anyone here, and most of these offers are legal. The checks are simple: is the firm on the register of a financial regulator in your country (for example the FCA register in the UK, or the national regulator in the EU)? Does anyone show a full, verified record, including the losing months, rather than screenshots? Is the promise a fixed return? In trading, nobody can honestly promise one.

15Five things you will hear, checked against the numbers

• “Scalping is safer because the trades are short.” Short trades with small stops are the most expensive kind: every simulated scalper year ended with a loss, because each trade gave away 20% of its risk to costs.
• “High leverage lets a small account grow fast.” It lets it shrink fast. At 10% risk with 500:1, 99.5% of the day traders lost 90% or more in a year.
• “Risk 1% and you will be fine.” 1% slows the loss; it does not stop it. The day trader at 1% still lost money in 98.5% of years, with a median of -46.7%.
• “A 60% win rate means a profitable strategy.” The RSI versions in section 10 won more than 60% of their trades, and three of four lost money.
• “If 70% lose, 30% win, so I just need to be in the 30%.” Being in the winning 30% for one period is not evidence of skill. With 1 pip of cost, the random swing trader, a trader with no edge at all, still ended the year with a profit in 34.4% of our simulated years, and without costs 27% to 44% of the random traders ended in profit. Some of the real winners in any year are simply the lucky ones.

16What you can actually do

1. Know your cost as a share of your risk. Spread plus commission, divided by your stop in pips. Above 5% is a heavy handicap; at 20%, as for the 5-pip scalper, no plausible edge survives it.
2. Trade less. Fewer, better-defined trades with wider stops give an edge, if you have one, room to show, and cost less while you find out whether you do.
3. Risk 1% or less a trade, and never raise it to recover a loss. Small size will not create an edge, but it buys the time to learn.
4. Use a regulated broker with negative-balance protection, and treat very high leverage as a warning sign, not a feature.
5. Measure your edge before you trust it. Write down your rules, backtest them with costs, then trade them on a demo account for a few hundred trades. If the result after costs is not clearly positive, do not add money.
6. Keep a journal of every trade with the planned stop and the actual exit; it shows the habits in section 13 better than memory does.
7. Count swap if you hold overnight, and check what it costs on your pair today.
8. Be suspicious of win rates without average losses and results after costs.
9. Consider not trading. For money you cannot afford to lose, a low-cost index fund has beaten most retail traders. That is not advice for anyone in particular; it is what the numbers above imply.

17The limits of this article

• The simulated traders enter at random. They show what happens without an edge; they cannot show how many traders have one. Real traders did better than our random frequent traders in the official figures.
• One pair, one broker's price history, fixed 1:1 stops and targets, constant costs and today's swap rates. Real spreads widen at news and at the daily rollover, which our fixed costs do not capture.
• The official figures cover CFDs of all kinds at particular brokers and periods; they are not a measure of forex alone.
• Nothing here is advice to trade or not to trade. Trading leveraged products can lose you all of your money.

FAQCommon questions

What percentage of forex traders lose money?

Most retail accounts do. The European Securities and Markets Authority reported in March 2018 that 74% to 89% of retail CFD accounts typically lose money. EU and UK brokers must publish their own figure; on 14 October 2026 two large brokers showed 70% and 74.1%. These figures cover CFDs of all kinds, not only currencies.

Why do most forex traders lose money?

Costs are charged on every trade, most traders trade often, leverage speeds up the losses, and few strategies have an edge after costs. In our simulation on real EUR/USD prices, a day trader with no edge, 1% risk and 1 pip of cost lost money in 98.5% of years, and costs took 41.2% of the starting account.

Do scalpers lose more than swing traders?

With the same cost per trade, yes. In our simulation every scalper year (5-pip stops, 1 pip of cost) ended with a loss, with a median of -99.5%, while swing traders (50-pip stops) lost money in 65.6% of years with a median of -4.5%. One pip is 20% of a 5-pip stop and 2% of a 50-pip one.

How much do spreads cost a forex day trader in a year?

Roughly the number of trades times the cost as a share of the stop times the risk per trade. A day trader with 15-pip stops, 1 pip of cost and 1% risk making about 833 trades a year paid 41.2% of the starting account in our simulation.

Does high leverage make you lose money faster?

Yes. Day traders in our simulation who wanted to risk 10% a trade lost 90% or more of the account in 99.5% of years with 500:1 leverage, and in 65.5% with the EU's 30:1 limit, which cut their position size.

Can anyone make money trading forex?

Some traders do: one academic study of 428 accounts from 2004 to 2009 found positive returns after costs on average. But a profitable year is weak evidence: in our simulation a swing trader with no edge still ended the year in profit in 34.4% of years. Of 96 common indicator strategies we tested on four major pairs, 30 were profitable after costs and none passed our tests for luck.

Is risking 1% per trade safe in forex?

It slows losses but does not prevent them. A day trader with no edge risking 1% a trade still lost money in 98.5% of our simulated years, with a median result of -46.7%. Small size buys time to learn; it does not create an edge.

Is forex trading gambling?

For a trader with no edge after costs, the arithmetic is the same as a game with a house fee: the costs are certain and the wins average out. With a measured edge, a sensible size and low costs it is not. The difference is whether you can show the edge with rules fixed in advance and costs included.

18Check our arithmetic

Code provenance: our own code. Official figures are quoted from the sources linked in section 02 and are theirs, not ours.

19Corrections

None since publication.

+More tests and guides

• Forex strategy backtest
• Stochastic rsi macd strategy
• Rsi strategy backtest
• How to pass a prop firm challenge
• Free prop firm challenge calculator
• What all the backtests found so far
• The backtesting protocol
• How to backtest a trading strategy without fooling yourself
• Backtesting glossary