Risk‑return and maximum drawdown analysis
Don’t only ask “How much was made.” First put the per‑trade risk‑reward, the volatility of historical returns, and the depth and recovery time of equity‑curve drawdowns on the same chart to know whether a result is supported by a stable edge or by a few extreme events.
Returns are not the complete answer
The same cumulative return can come from a smooth upward path or from a long decline followed by a single jump; volatility, skewness and drawdown path change the holder's actual experience.
MDD is not volatility
Volatility incorporates period‑by‑period changes into a statistic; maximum drawdown only cares about the deepest drop from a past peak to a subsequent trough. Both should be shown side‑by‑side, not used as substitutes for one another.
1. First split the two worlds of risk‑reward ratio
“Risk‑reward” has at least two meanings in trading dialogue. The first is for a single trade... R:R (risk‑to‑reward ratio)，describe a not‑yet‑realized scenario using stop‑loss distance and target distance；the second is the strategy's or portfolio'sHistorical risk‑adjusted performance，for example Sharpe, Sortino and Calmar — use a series of realized returns and equity data to answer "how much return was earned in the past and at what cost."
| Question | Single‑trade R:R | Historical performance metrics |
|---|---|---|
| Data type | Entry, stop‑loss, target, direction | Price or return series ordered by time |
| Primary use cases | Before entry, check the maximum loss and reward potential. | Describe volatility, downside risk and drawdown path after the fact |
| Core blind spots | Does not include win rate, slippage, gaps or consecutive losses | Historical samples do not guarantee future results and are affected by frequency and data cleaning |
| Pairs well with | Position size, Kelly sensitivity, trade records | Out‑of‑sample tests, stress scenarios, drawdown contingency plans |
Calculation of single R:R and expectancy
For a long trade, risk distance is entry price minus stop price, reward distance is target price minus entry price; this risk:reward (R:R) can be expressed as “reward distance ÷ risk distance.” If entry 100, stop 95, target 115, risk distance = 5, reward distance = 15, nominal R:R = 3R. This only describes price distances, not that the trade will necessarily reach the target.
Want to check a single scenario first?
First use an independentTrading expectancy calculatorInclude fees and slippage, then use historical series tools to check volatility and maximum drawdown; can also link to position sizing and Kelly risk‑budget tutorials.
2. Convert price series into comparable returns
Before analysis, the most important step is not selecting a pretty metric but confirming data conventions. If you paste a price series, the tool calculates simple returns from old to new; if you paste a return series, the tool compounds the equity curve from starting capital period by period. Do not mix the two silently, and do not treat prices unadjusted for dividends and splits as comparable investment returns.
Simple returns, log returns and annualization
This tool presents simple returns because they can directly reconstruct asset value; log returns have other mathematical and modelling uses, but you should not mix the averages, aggregation methods or annualization formulas of the two. The annualization factor is not absolute: daily data commonly uses 252 trading days, weekly 52, monthly 12; if data contains non‑trading days, time zones or different markets, reconfirm.
| Check items | why it matters | Tool processing |
|---|---|---|
| Order | A drawdown requires a peak before the trough; reversing the order would rewrite the whole path. | Explicitly require old‑to‑new order; do not auto‑sort |
| Price adjustment | Jumps caused by splits or dividends may be misread as returns. | Confirmed by the user at the data source: adjusted／unadjusted |
| Frequency | Volatility and Sharpe annualization change with observation frequency | Select periods per year for day/week/month |
| Fees and cash flow | Contributions, withdrawals and fees change the equity curve — they are not necessarily investment returns. | The tool does not infer cash flows; users must prepare the series themselves. |
3. Maximum drawdown: find the deepest pit along the entire equity path
The CFA Institute describes maximum drawdown as the cumulative loss from a local peak to the subsequent local trough over a specified period; there can be multiple drawdowns in the period but only one maximum drawdown.[3] The time order “peak then trough” is important — you cannot simply subtract the series’ global minimum from its global maximum.
Depth, duration and recovery period should be viewed separately.
Two curves can both have a −20% MDD, yet one returns to prior highs within a few observation periods while the other does not recover for dozens of periods. The psychological strain and capital drag are typically different. Therefore the tool also reports MDD peak, trough, the number of periods required for recovery, and whether the series is still in drawdown at the last sample point. Those fields are descriptive, not promises of "when" recovery will occur.
| Periods | Equity | Running peak | Drawdown |
|---|---|---|---|
| 0 | 100 | 100 | 0% |
| 1 | 110 | 110 | 0% |
| 2 | 104 | 110 | −5.45% |
| 3 | 88 | 110 | −20.00% |
| 4 | 98 | 110 | −10.91% |
| 5 | 112 | 112 | 0% |
4. Sharpe, Sortino, Calmar and recovery factor
The historical view of Sharpe can be understood as “average excess return ÷ historical standard deviation of excess return.” William F. Sharpe also warns the measure depends on the observation period; compounding, serial correlation and other distribution features make simple annualization incomplete.[1] The tool therefore lists frequency and sample size and does not treat fixed Sharpe thresholds as universal performance standards.
| Indicator | Tool definitions | What does it answer? | How not to use it |
|---|---|---|---|
| Annualized volatility | 樣本標準差 × √periods/year | Dispersion of returns across periods | Not equal to maximum possible loss |
| Sharpe | Excess mean return ÷ sample standard deviation, then annualize by frequency | Historical excess return per unit of total volatility | Does not imply tail risk has been captured |
| Sortino | Excess mean return ÷ MAR downside deviation, then annualize by frequency | Downside risk per unit below MAR | Do not output fabricated infinite values when there are no downside observations. |
| Calmar | Annualized return ÷ |MDD| | Annualized growth relative to deepest historical drawdown | Different periods, strategies and MDD definitions should not be compared blindly |
| Recovery factor | cumulative net profit ÷ MDD amount (explicit definition used by this tool) | net profit relative to maximum drawdown | Do not name it the same as or interchange it with Calmar |
Why Sharpe and Sortino may give different signals
Sharpe uses standard deviation for deviations both up and down; Sortino focuses on shortfalls below the minimum acceptable return (MAR). Charles Schwab describes Sortino as a downside‑risk‑focused variant of Sharpe and cautions that it is still built on historical data and should be a starting point for further analysis.[2] For right‑skewed return sequences — a few large gains and many small declines — the two metrics can diverge substantially; that does not make one metric inherently “correct,” it means they illuminate different risk questions.
Paste your own sequence into the analyzer.
Supports both price and return modes, frequency, MAR, starting capital, drawdown chart and CSV download; does not automatically fetch market data.
5. Practical analysis workflow: organize data first, then interpret indicators
Good risk control is not about collecting as many ratios as possible, but about ensuring each number can be recomputed. Save the original source and download date, confirm price‑adjustment methods, then build series at the same frequency; look first at cumulative returns and equity charts, then at MDD, recovery period and drawdown distribution, and only afterward interpret Sharpe, Sortino or Calmar within your strategy assumptions.
Step 1: data contract
Specify clearly whether the series is prices or returns, ordered old→new, observation frequency, sample period, currency, whether it's adjusted close, and whether it includes dividends, fees, taxes and cash flows.
Step 2: examine the trajectory first
Check equity, running peak, drawdown, low‑water mark and recovery status; don’t only capture terminal returns—path matters for capital and psychological stress.
Step 3: conduct stress testing
Ask: if historical MDD widens, recovery time lengthens or trading costs rise, can positions, liquidity and cash buffers still be sustained?
Step 4: avoid over‑interpretation
Small samples, high autocorrelation, survivorship bias, data misalignment, strategy revisions and omitted costs all erode the comparability of attractive metrics.
Link results to existing tools.
If you are evaluating a single trade, start with ourPosition sizing and risk budgetingCheck max loss, then useKelly risk budgetCheck sensitivity of win rate and payoff‑ratio estimates; at the portfolio level, useAsset allocation and dynamic rebalancingCheck asset weight deviations. The three inputs and their issues differ and cannot be replaced by a single number for the full process.
6. Data sources, assumptions and model limitations.
This article uses William F. Sharpe's original Sharpe Ratio exposition, Charles Schwab's Sortino tutorial, and the CFA Institute's educational article on maximum drawdown; they support the indicators' definitions and limitations but do not provide any personalized asset allocation or guarantees of future returns. [1] [2] [3]
[1] William F. Sharpe, The Sharpe Ratio, Stanford University
[2] Charles Schwab, How to Use the Sortino Ratio
[3] CFA Institute, Maximum Drawdown and Optimal Portfolio Strategy
[4] Choi et al., Maximum Drawdown, Recovery, and Momentum
Basis：Simple returns, sample standard deviation, downside deviation relative to a specified MAR, running-peak peak-to-trough MDD; recovery factor explicitly defined as “cumulative net profit ÷ MDD amount.”Time：Does not use real‑time quotes; the time base is entirely determined by the user‑provided series and the selected frequency.Assumptions:Default annualization factors for daily/weekly/monthly are 252/52/12; starting capital and MAR are user inputs.Sources & Confidence：Definitions follow the public educational/research sources above; users should verify the quality of historical samples, price adjustments and trading costs.Compliance：This is research and analysis only, not personalized financial advice.
7. Full breakdown of a fictional sequence in a research report
Assume the researcher gets six time‑ordered price observations: 100, 110, 104, 88, 98, 112. Convert to period simple returns and rebuild equity with starting capital 100,000; this example only illustrates the formula and does not represent historical performance of any market or asset. The equity path goes 100,000 → 110,000 → 104,000 → 88,000 → 98,000 → 112,000, so the first peak is at 110,000 and the deepest trough is the subsequent 88,000.
If the researcher only looks at start and end, they get +12% cumulative return; if they only report the final Sharpe they may ignore interim peaks and troughs of −20% drawdown. A full report should list at least data frequency, sample size, cumulative/annualized returns, volatility, MDD, recovery time, cost conventions and whether still in drawdown. That lets readers break “good results” into checkable evidence instead of being misled by a single summary metric.