Why Traders Lose Money: The Truth About the 95% Figure

There is no evidence that exactly 95% of traders lose money. It is a rounded number from course ads and forums, not the result of a study. Verifiable data ranges from 74% to 97% depending on the market, the country, the period and who was counted as a trader. The losses are real, but the question of why traders lose money is answered not by a magic percentage but by understandable mechanisms (fees, leverage, lack of a plan, drawdown math).

Is There Proof That Exactly 95 Percent Lose

Ask where the 95% comes from and you will not get a link to a primary source. The phrase "95 percent of traders lose money" lives in headlines and promo materials, but no regulator and no peer-reviewed paper has ever produced that number for all traders in the world. The figure is convenient because it is frightening enough to sell training "for the remaining 5%" and vague enough that it cannot be disproved.

The second source of confusion is that real studies produce different numbers (89%, 93%, 97%, 74–89%), each with its own market, sample and metric. Retold without context, they blur into "almost everyone loses", and then someone rounds it to 95.

What Counts as a Loss and Over What Period

The same set of trades gives different percentages depending on the definitions. Here is what you need to clarify before trusting any figure.

• Period. The share of losing participants over a month, a year and three years does not match. The longer the window, the more random wins get smoothed out and the higher the share of losers.
• Treatment of costs. The result before fees and after fees differs noticeably. In the Indian SEBI study, transaction costs alone push some accounts from profit into loss.
• Unit of measurement. Are people or accounts being counted? A person with three accounts, two blown up and one in profit, ends up in the statistics in different ways.
• Sample composition. Everyone who opened an account, or only active traders? Only those who lasted a year? Dropping those who quit after their first losses artificially improves the picture.
• Definition of profit. Plus 1 dollar for the year is formally a profit, but it is not an income. That is why some studies separately look at how many participants earned more than the minimum wage or some other threshold.

Until these parameters are stated, the figure carries no information, and that applies to every percentage in this text.

What Specific Studies Show

Below are studies that have an author, a sample and a methodology. The list is not complete, but it shows how strongly the result depends on the conditions.

Four different markets, four different numbers. They have one thing in common. Wherever complete data on active speculators exists, the majority ends up in the red, and only a narrow group earns consistently.

The SEBI Example: India, Equity F&O, FY22–FY24

The study by the Indian regulator SEBI is useful because it is built on real exchange data rather than a survey. In September 2024, SEBI published an updated report on the exchange-traded equity derivatives segment for the three financial years from FY22 to FY24. The main findings.

• 93% of more than 10 million individual traders ended the period with a net loss, on average about 2 lakh rupees (200 thousand) including costs. Only about 1% earned more than 1 lakh (100 thousand) after costs.
• Over three years, individual traders paid about 50,000 crore rupees in transaction costs. Half of that went to brokerage fees, a fifth to exchange charges. In FY24 the average participant spent about 26,000 rupees on costs.
• More than 75% of loss-making traders kept trading despite losses in two consecutive years.
• Proprietary firms and foreign portfolio investors in the same segment posted gross profits in FY24, almost entirely from algorithmic trading.

Subsequent SEBI reports refined the picture. About 91% loss-making in FY25 and 87.7% in FY26. The regulator also noted that about 90% of those who lost money in two consecutive years and kept trading lost again in the third year. Experience by itself did not improve results.

Note the wording. This is 93% of individual participants in the Indian equity derivatives market over three specific years. Not 93% of the world's traders and not 93% of crypto traders.

Why This Is Not a Statistic for All Crypto Traders

No comparable study exists for the cryptocurrency market. Regulators such as SEBI or ESMA see the data of licensed brokers in their own jurisdiction. Crypto exchanges operate in dozens of countries, rarely publish statistics on client results, and one person often holds accounts on several platforms. The US CFTC, in its advisory on the risks of virtual currencies, describes the mechanics of losses (unregulated platforms, manipulation, leverage that can make you lose more than you deposited) but does not name a share of losing traders.

Applying Indian or Brazilian percentages to the crypto market is incorrect for three more reasons.

• A different cost structure. In crypto, costs consist of exchange fees, the funding rate (periodic payments between holders of long and short positions in perpetual contracts) and slippage.
• Different leverage. Retail derivatives in the EU are capped on leverage, while many crypto exchanges offer dozens of x, which speeds up both wins and losses.
• A different sample. All the studies mentioned counted active speculators. A person who bought BTC and has held it for two years will not fall into such a sample, even though formally they also "trade crypto".

There is no figure for crypto traders, but the mechanisms that led to losses in other markets work harder in crypto, not softer.

Five Mechanisms of Capital Loss

Studies show how much is lost. Understanding why traders lose money comes from examining the mechanisms. There are five, and none of them involves "the market being against you". It is these, not a market maker conspiracy, that explain why beginners blow up their deposits.

• Transaction costs multiplied by trade frequency.
• Leverage that turns ordinary volatility into liquidation.
• Lack of a plan and attempts to win back losses.
• Drawdown asymmetry, where recovery requires more percent than was lost.
• Self-deception when evaluating results (a short sample, survivorship bias, memory of wins).

The first two mechanisms can be counted in numbers, so we will look at them in more detail. The rest show up in behavior and are covered further on.

Fees and Trading Too Often

A fee of 0.05% looks negligible. The problem is that it is charged on each side of every trade, and an active trader makes dozens of trades a day. Let us count with a hypothetical example, and take the exact rates from your own exchange's fee schedule.

Example conditions. Deposit 1,000 USDT, each position equal to the deposit (no leverage), taker fee 0.05% on entry and 0.05% on exit, 20 trades a day, 20 trading days a month.

• Cost of one trade 1,000 × 0.1% = 1 USDT.
• Per day 20 × 1 = 20 USDT, that is 2% of the deposit.
• Per month 400 USDT, that is 40% of the deposit.

Just to break even, such a trader needs to earn 40% a month from trading before fees. That is a demand on the strategy that few can meet. Add slippage (the difference between the expected and the actual execution price) and the funding rate on perpetual contracts, and the bar rises further.

Now switch on 5x leverage with the same deposit. The position notional becomes 5,000 USDT, and the fee is calculated on the notional, not on the margin. The cost of one trade rises to 5 USDT, per day to 100 USDT, that is 10% of the deposit. Over a month, fees alone would eat twice what was in the account. For the same reason, in the SEBI data costs over three years came to about 50,000 crore rupees, even though each individual trade looked cheap.

The number of trades is an expense item in itself. Before looking for "the best entry point", calculate what your trading frequency costs as a percentage of your deposit per month.

Leverage, No Plan and Trying to Win It Back

Leverage does not change the probability of guessing the direction. It changes the distance to liquidation. At 10x leverage, a price move of about 10% against the position wipes out the margin (the exact level depends on the exchange's maintenance margin and fees). For BTC such a move happens within a day, for altcoins it can happen within an hour. A trader can be right about the direction over a week and still lose the position within a day.

Lack of a plan shows up when position size, stop-loss level and the reason for exiting are decided after entry, under the influence of the current PnL. Such a trader locks in a small profit because they are afraid of losing it, and holds a loss because they hope for a return. As a result, the average loss ends up larger than the average profit, and even a high win rate does not save them (we do the math in the FAQ).

Trying to win it back adds a third layer. After a series of losses, the trader increases the position size to "get it all back in one trade". Drawdown asymmetry works against them. If the deposit has fallen by 20%, returning to the starting level requires earning not 20% but 25%. Going from the remaining 800 USDT to 1,000 USDT is a gain of a quarter. The deeper the drawdown, the steeper the requirement.

Increasing risk after losses speeds up the slide down this table. The state in which decisions are made to recover what was lost rather than according to the plan is called tilt. The article Tilt in Trading: How to Recognize and Stop It covers its signs and ways out.

How to Evaluate Your Trading Without Self-Deception

The fifth mechanism is more dangerous than the others because it hides the first four. A few rules help you look at your results truthfully.

Count the result only after all costs. Fees, funding and slippage must be included in the PnL of every trade. A trading journal that takes execution data directly from the exchange is more reliable than a manual spreadsheet. In Secret Terminal, the Trading Journal module automatically collects trades from connected API keys and shows a calendar of daily results, a cumulative profit curve and a card for each position with volume, leverage and fees.

Separate win rate from expectancy. The share of profitable trades says nothing about the result until you know the average size of a win and a loss. Expectancy per trade equals win rate × average win minus (1 − win rate) × average loss. If the number is negative, trading frequency only speeds up the losses.

Do not draw conclusions from a short series. Ten winning trades in a row happen even with negative expectancy. A series of 50–100 trades under the same rules gives at least some idea of the system, while a smaller sample mostly shows luck.

Remember survivorship bias. Those who made money are the ones visible in public, writing posts and showing screenshots. The thousands who left with losses stay silent, so the feeling that "everyone around me is succeeding" does not reflect the real proportions.

A separate word about demo accounts. A demo is useful for learning the interface, the order book and the tape, and for checking that stops and take-profits are placed as intended. But the simulation does not reproduce the main thing. It has no real emotions from losing your own money, nor the slippage and execution delays that occur when the order book has little liquidity. It cannot be claimed that everyone necessarily blows up a deposit after demo, that is unproven. The sensible next step is modest. A small real amount, for example 10 dollars, that you are prepared to lose entirely, taking into account the exchange's minimum order size and fees. Leverage as a way to "enlarge" a small deposit is not an option here, it enlarges not the capital but the speed of losing it.

A systematic review of typical missteps, from trading without a stop-loss to averaging down a losing position, is collected in Beginner Mistakes in Crypto Trading: The Complete Checklist.

FAQ

• Can 95% be treated as a universal statistic?

  No. The 95% figure has no primary source. Real studies give between 74% and 97% of losing participants, and each number is tied to a specific market, period and counting method. It is correct to say "the majority of active speculators in the markets studied lost money", not to name one universal percentage.

• Why does a high win rate not guarantee profit?

  Because the result depends on the ratio of average win to average loss, not just on the share of wins. Example. Win rate 70%, average win 10 USDT, average loss 30 USDT. Expectancy per trade equals 0.7 × 10 − 0.3 × 30 = 7 − 9 = −2 USDT, and that is before fees. Seven wins out of ten, and the account shrinks.

• How much does trying to win it back cost?

  More than the original loss, because of drawdown asymmetry. Hypothetical example. Deposit 1,000 USDT, the plan allows 2% risk per trade, that is 20 USDT. After five losing trades in a row, 900 USDT remains and +11.1% is needed to recover. The trader decides to win it back and risks 10% (90 USDT). If the trade loses, the deposit falls to 810 USDT, and now +23.5% is needed. One violation of the plan more than doubled the required future return.

• Can a system be evaluated on a few trades?

  No. Five or ten trades mostly show randomness, not the quality of the rules. The minimum benchmark for first conclusions is 50–100 trades under unchanged rules, with all costs included and preferably across different market conditions. If the rules changed along the way, the sample has to start over.

Instead of searching for "the right percentage", it is more useful to do two things. Write a trading plan with position size, stop-loss and maximum daily loss, and then review your trades regularly in a journal. Secret Terminal, with its order book, tape and Trading Journal module, helps you see the market and your own statistics in one window, but the plan and the discipline remain with the trader.