Overview

Risk metrics are compressed descriptions of history. Useful decisions combine them with scenario analysis, liquidity and an explicit definition of failure.

A risk metric is a lens with blind spots. The right question is not which number looks sophisticated, but which loss mechanism the number can observe and which one remains outside the dataset.

AssetsNest research desk

The Owl view

Evidence checked · 7 August 2026

Risk metrics compress history; they do not discover the future. A robust dashboard pairs volatility and drawdown with leverage, liquidity, concentration, valuation frequency and a scenario stated in rupees.

ConfirmedFY22–FY24
93%

individual F&O traders lost money

A strategy can show many small positive days and still have negative full-period economics.

Open source ↗
ConfirmedDecember 2025
2025

IPEV valuation standard refreshed

Private marks require process and disclosure precisely because frequent market prices do not exist.

Open source ↗

Case file

SEBI FY22–FY24 study

F&O losses expose the weakness of win-rate metrics

A trader can win often by collecting small option premiums and still lose after rare large moves, costs and leverage. SEBI's aggregate outcome data is therefore a better starting point than screenshots of percentage winners.

Measure expectancy, tail loss and capital at risk—not just volatility or the fraction of profitable trades.

What the market often misses

  • Sharpe ratios can be inflated by smoothed marks or hidden tail exposure.
  • Beta depends on the chosen benchmark and sample period.
  • Tracking error says how differently a portfolio moved, not whether the difference added value.

Questions before acting

  1. How does each metric change with a different sample, benchmark or valuation frequency?
  2. What loss is absent because an asset has not been repriced?
  3. Which scenario matters economically even if it has never appeared in the historical sample?

Topic 1 of 5

Volatility & drawdown

Volatility measures dispersion; drawdown measures peak-to-trough loss. Neither alone captures every form of risk.

The part that changes the answer

Review downside frequency, maximum loss, recovery time and valuation frequency. Illiquid assets can report low volatility simply because prices are estimated infrequently.

The underwriting question

Translate percentages into rupee loss and years to recovery.

Work the numbers

An asset can have moderate annual volatility and still suffer a 45% peak-to-trough loss if negative returns cluster; volatility and drawdown answer different questions.

What the underwriter checks

Measure frequency-consistent volatility, maximum drawdown, duration, recovery and liquidity. Use unsmoothed proxies for assets with appraisal marks.

Where the argument breaks

Quarterly marks suppress volatility while economic leverage grows, or a short history excludes the only regime that matters.

Real-world caseIPEV 2025: a private-company mark is a documented judgement, not a market priceRead the complete case study →

Topic 2 of 5

Sharpe & Sortino

Sharpe compares excess return with total volatility; Sortino focuses on downside deviation.

The part that changes the answer

Both depend on the return sample, benchmark rate and distribution assumptions. They can reward smoothed marks, leverage or rare-tail strategies until a loss arrives.

The underwriting question

Use ratios to compare similar strategies, never as proof of safety.

Work the numbers

A 10% return, 4% cash rate and 8% volatility gives 0.75 Sharpe. If rare losses are absent from the sample, the ratio can look excellent until the first tail event.

What the underwriter checks

Use net returns, relevant risk-free rate, sufficient history, downside deviation and drawdown. Inspect skew, serial correlation and valuation smoothing.

Where the argument breaks

Option selling earns a high Sharpe from frequent small gains while hidden short-gamma risk creates one ruinous loss.

Real-world caseSEBI's F&O study: the missing denominator behind trading success storiesRead the complete case study →

Topic 3 of 5

Alpha & beta

Beta estimates sensitivity to a benchmark; alpha is the residual return after the chosen model's exposures.

The part that changes the answer

Results change with benchmark, time window and factors. Apparent alpha may be compensation for size, value, illiquidity, leverage or hidden tail risk.

The underwriting question

Ask whether reported skill survives fees and a more complete risk model.

Work the numbers

A fund earning 12% with beta 1.2 when the market earns 10% and cash earns 4% has simple CAPM alpha of 0.8%: 12% − [4% + 1.2×6%]. Fees and added factors can reverse it.

What the underwriter checks

Choose the right benchmark, estimate stability across windows, add size, value, momentum and sector factors, and use net investable returns.

Where the argument breaks

Market or factor exposure is relabelled alpha, while the beta estimate is unstable and selected after the result.

Real-world caseSPIVA India 2025: benchmark failure and survivorship belong in the same denominatorRead the complete case study →

Topic 4 of 5

Tracking error

Tracking error measures how much active returns vary around a benchmark.

The part that changes the answer

A low number can still coexist with benchmark losses, while a high number may be intentional for a concentrated strategy. Read it with active share, factor exposure and mandate limits.

The underwriting question

Benchmark-relative control is different from protection of capital.

Work the numbers

A 3% tracking error means active return commonly varies by several points; it does not say whether the average active return is positive.

What the underwriter checks

Measure ex-post and forecast tracking error, active share, factor bets, concentration and benchmark fit. Pair it with information ratio and downside periods.

Where the argument breaks

Closet indexing charges active fees for low deviation, or high tracking error comes from unintended sector concentration rather than researched selection.

Real-world caseSPIVA India 2025: benchmark failure and survivorship belong in the same denominatorRead the complete case study →

Topic 5 of 5

Correlation

Correlation measures co-movement but is not causation and is not stable across regimes.

The part that changes the answer

Estimate multiple windows and stressed periods. Use economic scenarios to reveal common exposure to growth, rates, credit spreads, currency and liquidity.

The underwriting question

Diversification should be evaluated in the scenario that threatens the goal.

Work the numbers

At −0.2 correlation, a 60/40 mix may diversify in ordinary periods; if correlation jumps to +0.6 in stress, expected protection falls sharply.

What the underwriter checks

Use rolling, conditional and factor correlation; test non-linear payoffs and valuation lags. Seek an economic reason for diversification.

Where the argument breaks

Historical correlation is low only because one asset is marked infrequently or because the sample excludes a common funding shock.

Real-world caseFranklin Templeton's six debt schemes: when daily access met hard-to-sell creditRead the complete case study →

India lens

What Indian readers should test

Indian investors should translate every risk percentage into a rupee drawdown and recovery time. Compare metrics only across products with compatible valuation frequency, leverage and liquidity.

Primary sources & further reading

Dated facts are linked to their source. Hypothetical calculations are labelled illustrative.

SEBI — Equity F&O profit-and-loss study, FY22–FY24IPEV — 2025 private-capital valuation guidelinesHow AssetsNest researches and labels evidence
Important information

AssetsNest Investor Services — ARN 318691. This guide is educational and informational only. It is not personalised investment, legal or tax advice, an offer, recommendation or solicitation. Rules, products and taxation can change; verify current official documents before acting.