Volatility is the spread of your returns, annualised. It answers one narrow question: in a normal year, how far do results wander from their own average. It says nothing about direction. A portfolio that gains 3% every month has low volatility. So does one that loses 3% every month.
saral.money computes volatility from the dense daily portfolio log rather than from rebalance checkpoints, because the standard deviation of a monthly series and a daily series are different numbers. Take the standard deviation of daily returns and scale it:
vol_annual = stdev(r_daily) * sqrt(N)
N is the number of trading sessions in a year on the exchange you traded, close to 250 on the NSE. Worked example: if your daily returns have a standard deviation of 0.9%, annualised volatility is 0.9 * sqrt(250) = 14.2%.
Reading the number
Rough bands, not measurements. Judge your figure against a comparable index, not against this table.
| Annualised volatility | Typical of |
|---|---|
| Under 8% | Debt-heavy or heavily hedged books |
| 8% to 14% | Large-cap indices, low-volatility factor tilts |
| 14% to 22% | Diversified Indian equity portfolios |
| 22% to 35% | Concentrated or mid and small-cap strategies |
| Above 35% | Leveraged, sector-bet, or very few holdings |
Where volatility misleads
It treats upside and downside identically. A strategy that occasionally jumps 8% in a day is punished exactly as hard as one that drops 8%. Sortino Ratio exists because of this.
It understates risk in illiquid names. A small-cap locked at its 5% lower circuit prints a small, tidy daily move while being completely untradeable. Your measured volatility looks calm; your actual exposure does not.
It assumes the shape of the distribution holds. Indian equity returns have fatter tails than a normal curve, so the days that matter most are the ones volatility models least well.
Sampling frequency changes the answer. The same strategy measured on monthly returns will usually report lower volatility than on daily returns, because intra-month swings are invisible to a month-end sample.