How the equity curve is calculated
For each trade, current equity E is multiplied by 1 + f × b after a win or 1 − f after a loss. Here f is the risk fraction and b is reward divided by risk. Each outcome is drawn independently at the same assumed win probability p. Equity includes only completed simulated trades, with no intratrade mark-to-market movement.
For example, $10,000 at 2% risk and reward/risk 2 becomes $10,400 after a win. A following loss risks 2% of $10,400 and leaves $10,192. The same number of wins and losses produces the same final equity in this fixed-fraction model regardless of order, though drawdowns can differ.
Arithmetic expectancy and compound growth
Expected payoff in risk units is p × b − (1 − p). At p = 0.5 and b = 2 it is 0.5R before costs. This is an input-based mathematical expectation, not a measured trading edge. Compound growth also depends on the size of f: a positive arithmetic expectation can coexist with negative expected logarithmic growth.
For 0 ≤ f < 1, expected log growth per modeled trade is p × ln(1 + f × b) + (1 − p) × ln(1 − f). At 100% risk, a single modeled loss reduces equity to zero and that path stops. This simulator does not estimate practical ruin thresholds, margin calls or losses beyond the chosen risk amount.
Interpret the displayed statistics
Maximum drawdown is the greatest percentage decline from a preceding simulated equity peak. Profit factor divides gross winning dollar amounts by gross losing dollar amounts; infinity means no losing dollars in that sample, not a proven strategy. With zero risk both totals and the displayed profit factor are zero.
The summary averages the generated paths. Its range and profitable-run fraction describe only those one to ten paths and are not confidence limits or a market success probability. No annualized Sharpe ratio is calculated because trade count does not supply a time interval or risk-free return.
Use the model within its limits
Enter finite values within the displayed bounds, then run the simulation. Each run draws new random outcomes; there is no saved seed. Download CSV to preserve that run. Editing any assumption clears its results so old paths cannot be read against new inputs.
The model excludes costs, gaps, taxes, changing payoff sizes, correlated outcomes and deposits. Its inputs and calculations stay in the simulator; normal site analytics are separate. Historical strategy testing requires a separate dataset and execution assumptions. See Investor.gov on investment risk for the distinction between return expectations and potential losses.