Guide

Ulcer Index and Pain Index explained

Harbor Capital's allocator committee compared two global macro sleeves side by side in Q1 2025. Both posted 9.1% annualized return over five years. Both reported a maximum drawdown of −18%. On paper they looked interchangeable. Then the risk desk plotted underwater curves: Fund A suffered one sharp −18% event in March 2020 and recovered within four months; Fund B never breached −12% on any single day but spent three years grinding 6–11% below its high-water mark. LP redemption requests on Fund B ran 2.4× higher despite identical headline max drawdown. The desk needed metrics that capture depth and duration of underwater pain, not just the single worst peak-to-trough dip. The Ulcer Index (Peter Martin and Byron McCann, 1989) and the related Pain Index answer that question. Derived ratios — the Martin ratio and Pain ratio — express return per unit of investor discomfort. This guide defines both indices, walks through computation, contrasts them with Calmar, Sterling, and Sharpe, works a Harbor Capital sleeve comparison, provides a metric decision table, lists pitfalls, and ends with an allocator checklist.

Why maximum drawdown is not enough

Maximum drawdown (MDD) records one number: the largest percentage decline from a prior peak to a subsequent trough. It ignores everything else about the equity curve — how long recovery took, how often the fund dipped below peak, and whether underwater periods were shallow and chronic or deep and brief.

Consider two simplified monthly equity paths over 36 months:

Path A (V-shaped): flat for 30 months, one −18% month, full recovery in three months. MDD = −18%. Time underwater = 4 months.
Path B (sawtooth grind): never worse than −12% in any month, but underwater 28 of 36 months at an average depth of −8%. MDD = −12%.

Calmar and many mandate limits key off MDD, so Path B looks safer. An allocator watching LP behavior knows Path B often triggers more redemptions — investors experience duration of pain, not just depth. Ulcer Index and Pain Index quantify that experience numerically.

The underwater curve

Both indices start from the same building block: at each observation date t, compute percentage drawdown from the running high-water mark:

D_t = 100 × (Equity_t / Peak_t − 1)

When the fund is at a new high, D_t = 0. Negative values trace the underwater curve below zero. Plotting D_t over time reveals depth (how far below peak) and duration (how long the curve stays negative) — the visual that convinced Harbor's committee to look beyond MDD alone.

Ulcer Index: root-mean-square drawdown

The Ulcer Index (UI) summarizes the underwater curve as the root-mean-square (RMS) of percentage drawdowns over the measurement window. For N observations:

UI = √( (1/N) × Σ D_t² )

Squaring drawdowns before averaging penalizes deeper dips more than shallow ones — a −10% month contributes four times as much as a −5% month. Taking the square root returns the result to percentage-like units. UI is always non-negative; higher UI means more cumulative underwater stress.

Worked example

Suppose five monthly drawdown observations (percent): 0, −3, −8, −6, 0. Sum of squares = 0 + 9 + 64 + 36 + 0 = 109. UI = √(109/5) = √21.8 ≈ 4.67%. A fund with the same MDD but deeper intermediate dips would score higher UI.

Martin ratio

The Martin ratio (also called UPI — Ulcer Performance Index) divides excess return by Ulcer Index:

Martin ratio = (annualized return − risk-free rate) / UI

It asks: how much excess return does this strategy deliver per unit of RMS underwater pain? Higher is better. Martin ratio is the natural companion to UI when ranking funds for allocator due diligence, similar to how Sharpe ratio pairs with volatility.

Pain Index: average drawdown depth

The Pain Index (sometimes called the drawdown index in retail analytics tools) uses a simpler aggregation: the arithmetic mean of absolute drawdown values over periods where the fund is underwater:

Pain Index = (1/M) × Σ |D_t| for all D_t < 0

Where M is the count of underwater observations. Unlike UI, Pain Index does not square drawdowns — a −10% month counts twice as much as −5%, not four times. Pain Index emphasizes typical underwater depth during time spent below peak, making it closer in spirit to the average drawdown used in the Sterling ratio.

Pain ratio

The Pain ratio divides annualized return (or excess return) by Pain Index:

Pain ratio = annualized return / Pain Index

Definitions vary: some vendors use excess return in the numerator, others use gross return. Always confirm before comparing across databases. Pain ratio rewards strategies that earn strong returns while spending minimal time at moderate depths below peak.

UI vs Pain Index: when they diverge

UI and Pain Index rank funds similarly when drawdowns are uniform. They diverge when one fund has occasional deep spikes and another has chronic shallow grinds:

Deep spike fund: UI rises sharply (squaring punishes depth); Pain Index may stay moderate if underwater periods are brief.
Chronic grind fund: Pain Index rises (many months at −6% to −10%); UI also rises but less dramatically than for an equal MDD spike.

Harbor Capital uses UI when evaluating tail-sensitive macro sleeves (deep spikes matter) and Pain Index when evaluating income-oriented strategies where chronic underwater time drives LP behavior.

Harbor Capital sleeve comparison

Revisiting the two global macro sleeves from the opening scenario with five years of daily data (net of fees, 4.2% annualized T-bill rate):

Metric	Fund A (V-shaped)	Fund B (sawtooth grind)
Annualized return	9.1%	9.1%
Maximum drawdown	−18.0%	−12.0%
Months underwater	4 / 60	28 / 60
Ulcer Index	3.8%	7.2%
Martin ratio	1.29	0.68
Pain Index	4.1%	8.6%
Pain ratio	2.22	1.06
Calmar ratio (36-mo)	0.51	0.76

Calmar favors Fund B (lower MDD). Martin and Pain ratios favor Fund A by wide margins. Harbor's policy: when Calmar and Martin diverge by more than 0.30 on either side, the committee reviews underwater curves and LP redemption history before sizing. Fund B received a 40% smaller allocation despite better Calmar — matching the redemption data the risk desk had flagged.

For path-dependent survival analysis during extended crashes, pair UI with sequence-of-returns risk when evaluating retiree or endowment portfolios making concurrent withdrawals.

Ulcer and Pain vs Calmar, Sterling, Sharpe, and Sortino

Metric	What it measures	Denominator / risk unit	Best when
Ulcer Index	RMS underwater depth	Standalone risk (lower = better)	Comparing equity curve shape; input to Martin ratio
Martin ratio	Excess return per UI	Ulcer Index	Ranking funds where depth-weighted pain matters
Pain Index	Mean underwater depth	Standalone risk (lower = better)	Chronic grind detection; LP behavior prediction
Pain ratio	Return per Pain Index	Pain Index	Income strategies with long underwater periods
Calmar	Return per worst drawdown	Maximum drawdown	Mandate max-loss limits; crisis tail focus
Sterling	Excess return per avg drawdown	Average drawdown	CTA evaluation; chronic underwater without squaring
Sharpe	Excess return per volatility	Standard deviation	Symmetric return distributions; liquid equity
Sortino	Excess return per downside vol	Downside deviation	Asymmetric return profiles; penalizes bad months only

No single metric captures investor experience completely. Sharpe treats upside and downside volatility equally. Sortino ignores upside vol but still uses monthly returns, not path shape. Calmar keys off one worst event. Sterling averages drawdowns linearly. UI squares depth; Pain Index averages absolute depth over underwater time. A robust allocator dashboard shows at least three: Sharpe or Sortino (return variability), Calmar (tail), and Martin or Pain ratio (path shape).

Computing UI and Pain Index in practice

Data frequency

Daily data captures intra-month dips that monthly data smooths away. A fund that spikes −15% mid-month and recovers before month-end shows 0% monthly drawdown but material daily UI. For hedge fund due diligence, daily NAV is preferred when available. Mutual fund and ETF analysis typically uses daily adjusted close prices from public market data.

Lookback window

Common windows: 36 months (aligns with Calmar convention), 60 months (full market cycle), or since-inception for younger funds. Rolling 36-month UI charts reveal whether underwater pain is improving or deteriorating — a declining UI trend with flat return often signals better risk control after a regime change.

Implementation sketch (Python/pandas)

Given a series of portfolio values equity:

peak = equity.cummax()
drawdown_pct = 100 * (equity / peak - 1)
UI = sqrt((drawdown_pct ** 2).mean())
pain = drawdown_pct[drawdown_pct < 0].abs().mean()

Annualize return separately (CAGR over the window). Martin ratio = (CAGR − Rf) / UI. Verify your database squares drawdowns for UI — some retail screeners mislabel average absolute drawdown as “Ulcer Index.”

Metric decision table

Question you are asking	Use this metric	Watch out for
What is the single worst peak-to-trough loss?	Maximum drawdown / Calmar	Ignores duration and frequency of dips
How deep is typical underwater pain, depth-weighted?	Ulcer Index / Martin ratio	Sensitive to one deep spike; confirm daily data
How deep is the fund on average while below peak?	Pain Index / Pain ratio	Linear weighting; less punishing than UI for spikes
Does the fund live in chronic shallow drawdowns?	Pain Index + months underwater	Calmar may look fine while LP redemptions rise
Return per typical drawdown episode (CTA style)?	Sterling ratio	Different from UI; uses average not RMS
Return per unit of downside volatility?	Sortino ratio	Monthly returns, not path shape
Will withdrawals during a grind cause plan failure?	UI + sequence-of-returns analysis	Static metrics miss concurrent cash flows

Common pitfalls

Confusing UI with average drawdown: UI squares before averaging; average absolute drawdown is Pain Index, not Ulcer Index.
Monthly data on volatile sleeves: intra-month spikes invisible on monthly NAV; UI understated.
Different lookback windows: comparing 36-month UI on one fund to since-inception UI on another mis-ranks strategies.
Ignoring fees: gross UI on a fund with 2/20 fees overstates investor experience; use net returns.
Survivorship bias: databases that drop liquidated funds inflate average Martin ratios in peer groups.
Calmar-Martin divergence without context: a V-shaped recovery fund may show great Martin and mediocre Calmar — both are informative, not contradictory errors.
Using UI for options-heavy books without mark caveats: stale or model marks smooth drawdowns; UI on marked-to-model NAV can lie.

Practitioner checklist

Plot underwater curves alongside headline return and MDD for every manager review.
Compute UI and Pain Index on daily data when NAV frequency allows.
Report Martin ratio and Pain ratio alongside Calmar and Sharpe in allocator memos.
Flag funds where Calmar > 0.60 but Martin < 0.50 for underwater-curve review.
Use consistent 36-month (or full-cycle) windows across peer comparisons.
Count months underwater as a simple LP-behavior proxy alongside Pain Index.
Verify vendor definitions — confirm squaring for UI, not mislabeled averages.
Net all returns of fees before computing ratios.
Pair UI analysis with behavioral finance insights on loss aversion and redemption timing.
Recompute rolling 36-month UI quarterly; trend matters as much as point estimate.

Key takeaways

Ulcer Index is the RMS of percentage drawdowns — it penalizes depth more than duration alone.
Pain Index is the average absolute drawdown while underwater — it captures chronic grind.
Two funds with identical return and max drawdown can have vastly different UI and investor redemption behavior.
Martin ratio and Pain ratio express return per unit of discomfort — use them when Calmar hides path risk.
Harbor-style allocators treat Calmar-Martin divergence as a signal to inspect underwater curves, not as a data error.