Do 90% of Traders Really Lose Money?

Search for trading education and you will find the same statistic on nearly every page: 90% of traders lose money. Sometimes it is 95%. Sometimes 99%. It is stated as a settled fact. Almost nobody links to a primary source.

We went looking for one. What we found is not a single study but several narrower findings, each with qualifiers that tend to disappear when the number gets passed along. Since this article is about what happens when sources go missing, we name and link ours.

What the studies actually measured

The most frequently cited research examined specific populations trading specific products.

Leveraged CFD accounts in Europe. In its 2018 product-intervention decision, the European Securities and Markets Authority reported that analyses by national regulators across EU jurisdictions found 74–89% of retail CFD accounts typically lost money, with average losses per client ranging from €1,600 to €29,000. These are leveraged contracts for difference — not stock portfolios or ETF holdings.

Equity futures and options in India. The Securities and Exchange Board of India reported in September 2024 that 93% of individual traders in equity futures and options incurred net losses between FY2022 and FY2024. Again, leveraged derivatives.

Equity-index futures day traders in Brazil. Chague, De-Losso, and Giovannetti (2020) tracked individuals who began day trading Brazilian equity-index futures between 2013 and 2015. Among those who persisted for more than 300 trading days, 97% finished with net losses after fees. Only about 1% earned more than the Brazilian minimum wage.

Day traders on the Taiwan Stock Exchange. Barber, Lee, Liu, and Odean studied fifteen years of Taiwan Stock Exchange data (1992–2006) and found that fewer than 1% of day traders demonstrated repeatable, benchmark-adjusted skill after fees. This is frequently shortened to "99% lose money," but that is not what it says. It says fewer than 1% showed persistent positive alpha. Some of the other 99% were temporarily profitable; some broke even; some lost money. The study measured skill persistence, not absolute loss.

Three claims that get treated as one

When these findings travel across blogs and social media, three distinct statements collapse into one:

• Negative absolute return. The trader ends with less capital than they started.
• Underperformance. The trader makes money but earns less than a passive benchmark.
• No persistent alpha. The trader may profit in some periods, but the results do not reliably repeat after fees and risk adjustment.

A well-known US study illustrates the gap. Barber and Odean (2000) examined over 66,000 brokerage households during the 1990s. The most active traders earned roughly 11% per year while the broad market returned roughly 18%. Those traders paid a steep activity penalty — but in absolute terms, during a rising market, most of them still made money. They underperformed. They did not go broke.

"Most active traders underperform" has broad academic support. "90–99% of all retail participants lose money outright" does not — at least not without the qualifiers about product type, leverage, and time horizon that usually get dropped.

Why the compressed number persists

A round, frightening number is easy to remember and easy to share. It also serves multiple agendas. Course sellers use it to imply their method is the cure. Content creators repeat it because it generates clicks. None of these parties have a strong incentive to add nuance.

There is one context where the number is real and sourced: regulators require CFD brokers to display the measured loss rate of their own clients, which is why European broker websites carry warnings like "74–89% of retail investor accounts lose money." Notice what that is — a specific product, a specific population, a measured figure. It is the folklore version, the universal "90% of everyone," that has no source.

The result is a statistic that feels authoritative but functions more like folklore. It compresses findings about leveraged futures day traders in São Paulo into a universal claim about anyone who opens a brokerage account.

Why it matters for how tools present themselves

The 90% number is often followed by a pitch: most traders fail because they lack the right indicator, signal, or system — and here is the one that fixes it. That framing treats the indicator as a prediction engine whose job is to put traders on the winning side of a statistic.

We think the framing is wrong.

An indicator that claims to solve the 90% problem is implicitly claiming to predict profitable outcomes. That is a performance claim, and performance claims require performance evidence — evidence that almost no indicator vendor publishes. We've written before about why we publish our methodology; this is the same standard applied to the industry's favorite statistic.

Oisigma's BTM does not claim to solve the 90% problem, because we are not sure the 90% problem, as popularly stated, is real. What is real is that bands and envelopes make statistical claims about containment, and those claims are testable. The BTM's inner band contained the next daily close roughly 71% of the time on the S&P 500 from 1928 to 2024 — past behavior is not a guarantee of future results. The full calibration record, including the misses and the decades where containment dipped, is published on our Proof page.

That is a narrower, less dramatic offering than "the tool that beats the 90% odds." It is also a verifiable one.

The honest position

We do not know what percentage of all retail market participants lose money. Nobody does, because the data does not exist in that form. What exist are studies of specific subsets — day traders in specific futures markets, holders of specific leveraged products — showing high failure rates under specific conditions.

Those findings deserve to be taken seriously on their own terms. They do not deserve to be inflated into a universal scare statistic and then used to sell the next solution.

If you are evaluating any market tool, the useful question is not "will this put me in the winning 10%?" It is "does this tool publish testable claims, and do the published results hold up?" The BTM's calibration methodology and historical results are documented in our working paper, and you can check them against the indicator yourself with a free 30-day trial — $15/month after that, cancel anytime. Have a look at the data and decide for yourself.