MEASURED · 1 SEPTEMBER 2026

Is forex trading profitable?

For a minority. The usual answer is a statistic about how many retail accounts lose money, followed by advice. The more useful question sits underneath it: whether you could tell from your own results that you were profitable. We simulated 1,800,000 trading records to find out, and the answer is uncomfortable for both optimists and sceptics.

How many retail forex accounts lose money

Forex trading is profitable for a small minority. Broker disclosures generally put the share of losing retail accounts around 70–80%. Those numbers are self-reported across different windows and jurisdictions, so they give an order of magnitude, and we will not quote one to a decimal place as if it were a measurement.

What we can measure ourselves is how hard it is to separate a real edge from luck over any realistic stretch of trading.

What we simulated

Take a strategy that risks a fixed amount and targets 2× that amount, which is a common and unremarkable setup. At that reward-to-risk, breaking even requires winning 33.3% of trades. We defined a “good” strategy as one winning 38%: a real but modest edge, the size a working system has, well short of what a course sells.

Then we ran 200,000 independent trading records at each sample size and counted two mistakes. A false positive is a strategy with no edge at all finishing profitable: luck that looks like skill. A false negative is the genuinely good strategy finishing in the red: skill that looks like failure.

How often a no-edge strategy looks profitable and a good strategy looks unprofitable, by sample size
TradesNo edge, but profitableReal edge, but losing
1044.1%43.5%
3041.5%37.4%
5051.2%23.6%
10048.1%17.6%
20050.8%8.2%
50050.4%1.4%

More trades fix one mistake and never the other

Read the middle column down the page. It does not improve: at ten trades a worthless strategy finishes profitable 44.1% of the time, and at five hundred trades it still does, 50.4% of the time.

That comes from the arithmetic. A strategy with zero expectancy has a cumulative profit that wanders up and down with no pull in either direction, so at any point you stop and look, it is roughly a coin flip whether it happens to be above water. More trades will never stop a worthless system from looking good. Only understanding why it should work does that.

The right-hand column responds to sample size, and sharply: from 43.5% at ten trades to 17.6% at a hundred and 1.4% at five hundred. A longer record protects you from throwing away something that was working. It does nothing to stop you being fooled.

Most people quit a good strategy during a normal drawdown and keep a bad one through a lucky run, and the numbers above say both errors are close to a coin toss early on. Reaching 95% confidence takes about 416 trades on this model. The full working is in how many trades it takes to judge a trading bot.

Losing streaks on gold, measured

This measurement comes from real market data instead of a simulation. Across 257 sessions of gold we counted 61 separate losing streaks. The longest ran 9 days. 14 of them ran three days or more.

The worst nine-day streak cost $755.70 and the worst three-day streak cost $738.84: almost the same damage in a third of the time. You do not always feel a losing streak building, because the short ones hurt just as much. “I will stop if it goes badly for a week” is a weaker safeguard than it sounds.

Full data and method: losing streaks on gold and what trailing drawdown really looks like.

How to judge a method instead

Judge a method by whether its reasoning survives inspection. Recent green results are close to uninformative until you have several hundred trades, as the table shows. Ask what market behaviour a strategy is exploiting and whether that behaviour has a reason to persist.

Size positions so the normal losing runs above are survivable, since the difference between a strategy that recovers and one that does not is usually position size more than signal quality. When we simulated a $100 account, a 0.02 lot roughly tripled the chance of emptying it compared with 0.01, for the same typical outcome. Keep a long enough record to reach the point where the right-hand column gets small.

We try to hold ourselves to the same standard. Our research publishes the studies that went against us alongside the ones that did not. In one, we wrote the test down in advance and ran the four strategies in the bot on their own, on gold from 2019 to 2024, years they were not tuned on, and none of the six years was positive. Our gold trading bot is documented with the same numbers instead of a performance screenshot.

Method and limits

The table is a Monte Carlo simulation with no live trading in it. It assumes independent trades of fixed risk at a constant win rate. Real trading does not behave that way: win rates drift with market regime and losses cluster. Both make the real picture worse than the table, so read the figures as an optimistic bound on how quickly you could know anything.

Model parameters are stated above. The full working, including where the 416-trade threshold comes from, is set out in how many trades it takes to judge a trading bot. The losing-streak figures are measured from 70,952 five-minute XAUUSD bars over 2025-07-31 to 2026-08-04, and are specific to gold rather than to forex generally.

Common questions

Is forex trading profitable?

For some participants, and not for most. The harder part is that profitability is very hard to establish from a short record. In our simulation a strategy with no edge at all finished profitable 44.1% of the time after ten trades, and a strategy with a genuine edge looked unprofitable 43.5% of the time over the same span. Both errors are near a coin flip, so a small sample does not answer the question either way.

How long does it take to become profitable in forex?

Long enough that the number of trades matters more than the number of months. On the model we tested, 2:1 reward-to-risk with a 38% win rate, it takes roughly 416 trades before a real edge is distinguishable from luck at 95% confidence. Someone taking two trades a day reaches that in about ten months, and someone taking two a week does not reach it for years.

What percentage of forex traders are profitable?

Published broker disclosures commonly put the share of losing retail accounts between 70% and 80%. Those figures are self-reported over varying windows, so treat them as an order of magnitude rather than a measurement. Our own data adds a reason the number is hard to pin down: over short records, profitable and unprofitable traders are statistically difficult to separate, so any snapshot mixes real skill with luck in both directions.

Can you make a living from forex trading?

Only with an edge, enough capital that a realistic percentage return is a liveable income, and the ability to survive normal losing runs. On gold we measured 61 losing streaks in 257 sessions, the longest running nine days, and the worst three-day run cost almost exactly as much as the nine-day one. Income planning that assumes steady weekly withdrawals runs into that behaviour quickly.