Ulcer Index
Trading Intelligence
9 min read
ulcerIndex
Quantify downside risk with the root-mean-square of drawdowns — a volatility measure that only penalizes losses.
Loading
9 min read
Quantify downside risk with the root-mean-square of drawdowns — a volatility measure that only penalizes losses.
Standard deviation punishes you for making money. The Ulcer Index only cares about what actually hurts -- drawdowns.
The Ulcer Index (UI) is a volatility metric introduced by Peter Martin and Byron McCann in The Investor's Guide to Fidelity Funds (1989) that measures downside risk exclusively. Unlike standard deviation, which treats upside and downside moves symmetrically, the Ulcer Index only considers the depth and duration of drawdowns from prior peaks.
The name is deliberately vivid: it measures the amount of "ulcer-inducing" stress a portfolio or trading system inflicts on its operator. A smooth, steadily rising equity curve has a low Ulcer Index. A choppy, drawdown-heavy curve has a high one.
This makes it one of the most psychologically honest risk metrics available.
The Ulcer Index is the root mean square (RMS) of percentage drawdowns from the running peak. Here is the step-by-step process:
Track the running peak: At each point in time, record the highest equity value achieved so far.
Calculate percentage drawdown: At each point, compute how far below the running peak the current equity sits.
Drawdown_i = (Peak_i - Equity_i) / Peak_i * 100
When equity is at a new peak, the drawdown is 0%. When equity is below the peak, the drawdown is a positive percentage equal to the gap from peak.
Square all drawdowns: This amplifies larger drawdowns disproportionately, penalizing deep declines more than shallow ones.
Take the mean of the squared drawdowns: Average them over the entire observation period.
Take the square root: This returns the result to the same scale as the original drawdowns.
Step 1: Drawdown_i^2 (square each drawdown)
Step 2: (1/N) * sum(Drawdown_i^2) (mean of squared drawdowns)
Step 3: UI = sqrt( (1/N) * sum(Drawdown_i^2) ) (square root returns to drawdown scale)
The result is always a non-negative number expressed as a percentage. A Ulcer Index of 0.05 (5%) means the average RMS drawdown from peak is 5%.
Worked example: Take an equity curve [100, 110, 105, 100, 95, 102, 108]. The running peaks are [100, 110, 110, 110, 110, 110, 110]. The percentage drawdowns are [0, 0, 4.55, 9.09, 13.64, 7.27, 1.82]. Squared: [0, 0, 20.66, 82.64, 186.05, 52.89, 3.31]. Mean = 49.36. sqrt(49.36) = 7.03. So the Ulcer Index is roughly 7.03 (or 0.0703 in decimal form) — squarely in the "rough" band.
Worked example: 7-period equity curve and running peak. Gap area = drawdown contribution to UI.
Squared drawdowns from the worked example. The squaring step amplifies deeper drawdowns disproportionately — the t4 bar dwarfs the rest.
| Ulcer Index | Assessment | What It Means |
|---|---|---|
0.00 - 0.03 | Excellent | Equity curve is smooth with minimal drawdowns. Very tradable. |
0.03 - 0.05 | Good | Modest drawdowns that resolve quickly. Healthy system. |
0.05 - 0.10 | Moderate | Noticeable drawdowns. Requires discipline to trade through. |
0.10 - 0.15 | Rough | Significant and/or prolonged drawdowns. Psychologically demanding. |
> 0.15 | Severe | Deep, extended drawdowns dominate the equity curve. High stress. |
Context matters. A Ulcer Index of 0.08 might be acceptable for an aggressive momentum strategy but unacceptable for a conservative mean-reversion system. Always compare within the same strategy class.
This is the key insight that separates the Ulcer Index from standard deviation.
Standard deviation measures the dispersion of returns around the mean -- both above and below. A strategy that occasionally produces outsized winners will have high standard deviation, even though those large upside moves are exactly what traders want. Standard deviation treats a +5R winner the same as a -5R loser from a volatility perspective.
The Ulcer Index eliminates this problem entirely:
This means two strategies with identical standard deviations can have wildly different Ulcer Indexes:
Standard deviation says they are equally risky. The Ulcer Index correctly identifies that Strategy B is far more painful to trade — and far more likely to violate the Max Drawdown Rules you set for the account.
UI vs max drawdown: Max drawdown is a single worst-point statistic — it tells you nothing about how the strategy spent its time below peak. Two strategies with -25% max drawdown can have UI of 4% (one quick dip, fast recovery) or 14% (fifteen months underwater). UI integrates the whole experience; max drawdown is a snapshot.
Just as the Sharpe Ratio divides excess return by standard deviation, you can divide excess return by the Ulcer Index to get the Ulcer Performance Index (also called Martin Ratio):
UPI = (Return - Risk-Free Rate) / Ulcer Index
This produces a risk-adjusted return metric that only penalizes downside risk. A higher UPI means more return per unit of drawdown stress. Rough heuristics for crypto strategies on daily equity: UPI < 1 = poor, 1–3 = decent, 3–7 = strong, > 7 = exceptional (and verify on out-of-sample data — UPI inflates aggressively on cherry-picked windows).
The UPI is often a better tool for comparing trading strategies than the Sharpe Ratio because:
| Metric | Risk denominator | Penalizes upside vol? | Captures DD duration? | Distribution-free? | Best use case |
|---|---|---|---|---|---|
| Sharpe | Standard deviation of returns (up + down) | Yes | No | No (assumes normality) | Apples-to-apples comparison of normal-return assets |
| Sortino | Downside standard deviation of returns | No | No | No (assumes downside normality) | Comparing strategies with asymmetric return distributions |
| MAR / Calmar | Max drawdown (single worst point) | No | Indirectly (worst point only) | Yes | Quick screen of long-horizon trend systems |
| UPI (return / UI) | RMS of drawdown depths from peak | No | Yes (squared by duration) | Yes | Evaluating real, fat-tailed trading systems |
The Sharpe Ratio assumes returns are normally distributed. Trading returns almost never are -- they exhibit fat tails and skewness. The Ulcer Index makes no distributional assumptions, making it the only honest choice when return distributions exhibit fat tails — which, for crypto and momentum systems, is always.
UI vs Sortino: Sortino divides excess return by the downside standard deviation of returns (the volatility of negative returns only). UI uses the RMS of drawdown depths from running peak — fundamentally different. Sortino sees a -3% day and another -3% day six months apart as identical noise; UI sees the second as worse if equity is still below the prior peak. Sortino punishes volatility of losing days; UI punishes time spent underwater.
One subtle but powerful property of the Ulcer Index is that it naturally incorporates drawdown duration.
Consider two drawdown events, both reaching a maximum depth of 10%:
Drawdown B produces a much higher Ulcer Index, even though the maximum depth was identical. This is correct behavior -- extended drawdowns are psychologically harder and represent greater systemic risk than brief sharp drops that recover quickly.
Calculate the rolling Ulcer Index over your last 50 or 100 trades. Plot it alongside your equity curve. A rising Ulcer Index even while equity is flat or rising slightly is an early warning -- it means drawdowns are deepening or extending.
If you trade multiple setups, calculate the Ulcer Index for each one independently. Some setups may contribute disproportionately to overall portfolio drawdown. Identifying and addressing these can dramatically improve the aggregate Ulcer Index.
Use the Ulcer Index to scale position size inversely. When the rolling Ulcer Index rises above a threshold (e.g., 0.10), reduce risk per trade. When it falls below a comfort level (e.g., 0.04), you have room to increase sizing.
When choosing between two strategies with similar returns, prefer the one with the lower Ulcer Index. You are far more likely to stick with a low-UI strategy through its inevitable rough patches — which is also why low UI is the structural cure for the revenge-trading patterns covered in Behavioral Risk Management.
For most retail strategies, a Ulcer Index below 0.05 is good and below 0.03 is excellent. Readings between 0.05 and 0.10 are moderate, and anything above 0.15 indicates psychologically demanding drawdowns. Always compare within the same strategy class — an aggressive momentum system tolerates higher UI than a conservative mean-reversion one.
Sortino divides excess return by the downside standard deviation of returns — it punishes the volatility of negative days. The Ulcer Index uses the RMS of drawdown depths from the running peak — it punishes time spent below prior highs. Two -3% days six months apart look identical to Sortino but differ to UI if equity is still underwater on the second one.
Peter Martin and Byron McCann introduced the Ulcer Index in The Investor's Guide to Fidelity Funds (1989). The associated risk-adjusted return metric, the Ulcer Performance Index, is also called the Martin Ratio in honor of the same author.
How this fits the module: UI is the capstone of the Risk Management module — it complements Max Drawdown Rules (UI captures the experience, max drawdown captures the worst point), extends Recovery Factor (RF rewards reaching new highs; UPI rewards smooth paths to them), and gives a numerical handle on the psychological cost that Behavioral Risk Management treats qualitatively.