Uncertain outcomes can occur in both trading and gambling. Calling an activity trading does not establish a positive expected return, and a charting method does not prove an advantage. This lesson compares the arithmetic and decision process; it does not provide a legal or clinical classification.

Expected value is a probability-weighted average

Expected value sums possible outcomes weighted by their probabilities. A positive value in a model is not a promise that a particular trade, month or account will finish in profit. In trading, the probabilities and future payoff distribution must be estimated rather than assumed known.

For example, assume a hypothetical trade wins 200 EUR with probability 40% and loses 100 EUR with probability 60%, before costs. The expected result is 0.40 × 200 − 0.60 × 100 = 20 EUR per trade. Four wins and six losses would total 200 EUR before costs, but a sequence of ten need not contain exactly those outcomes.

Include costs in the same calculation

If each trade in that model costs 5 EUR, expectancy falls to 15 EUR. If it costs 25 EUR, expectancy becomes −5 EUR. The cost is neither universally small nor fixed as a fraction of risk: it depends on the contract, size, execution and holding period.

Spread, commission, financing and slippage may all matter. Do not count a cost twice if the recorded net results already include it. Different counterparties and execution arrangements also need assessment; there is no market-wide guarantee that every provider's terms are neutral.

Worked example

Winning every other attempt is not enough

Scroll sideways to see the full diagram

A hypothetical break-even win probability after costsEach attempt pays plus or minus 100 USD before a fixed 4 USD cost. After costs, the outcomes are plus 96 or minus 104 USD. At assumed win probabilities of 48, 50, 52 and 54 percent, the expected results are minus 8, minus 4, zero and plus 4 USD.Costs shift the break-even pointWin after costs+96 USDLoss after costs−104 USDAssumed win probabilityExpected / attempt48%−8 USD50%−4 USD52%0 USD54%+4 USDp × 96 − (1 − p) × 104 = 0 → p = 52%

In this separate model, a 4 USD cost raises the mathematical break-even threshold from 50% to 52%. At an assumed win probability of p = 54%, expectancy is 0.54 × 96 − 0.46 × 104 = 4 USD per attempt. A positive answer in a spreadsheet does not establish that the assumed probability describes a real strategy.

Hypothetical model in US dollars per attempt: fixed payoffs of ±100 USD before costs and a fixed 4 USD cost. This is a separate example from the unequal-payoff example in the article. The results are neither a forecast nor an estimate of market win rates.

A loss rate does not explain every loss

The historical ESMA product-intervention discussion concerns retail CFDs. The dated account-loss range does not measure all FX participants, diagnose their behaviour or establish which trading mistake caused an individual loss.

Likewise, leverage and chasing losses can increase exposure without proving a particular strategy's expected value. Separate evidence about results from a judgment about a person's decisions. For the practical demands of income from trading, see making a living from forex.

An advantage and a risk budget answer different questions

A written reason, planned exit and position size make a trade auditable; they do not create an advantage by themselves. The expectancy calculation measures a different property from the capital available to withstand losses.

Risk of ruin depends on the definition of failure, payoff assumptions and sizing. A fixed percentage of remaining equity also differs from a fixed amount based on starting capital. Neither a one-percent budget nor a positive estimated mean guarantees survival. Read maximum drawdown for the distinction between losses and the recovery required.

A losing sequence does not make a win due

The gambler's fallacy is illustrated by independent trials: after several losses, the assumed probability of the next trial is unchanged. Real market trades need not be independent, so investigate relevant information instead of inferring a reversal solely from your account's losing streak.

Chasing losses by raising the next stake changes the money exposed; it does not automatically improve the probability or quality of the next decision. The psychology section considers these decision problems. Counting a few behaviours is not a clinical diagnosis.

Audit the process and the evidence

Record the reason, instrument, size, planned loss, actual outcome, costs and deviations. Calculate the mean net result over a clearly stated sample and examine how it changes with assumptions. A fixed number of trades or one month of records does not establish statistical reliability.

Choose a risk budget compatible with the funds available and test loss sequences, concurrent positions and execution beyond a stop. Use risk-management basics to connect those checks. The separate diagram below shows how a fixed cost shifts a model's break-even probability; it does not estimate any real strategy's win rate.