Most investors evaluate a strategy by one number: CAGR. "This strategy returned 42% last year" sounds great. But 42% with a 58% drawdown in between, where your ₹10 lakh fell to ₹4.2 lakh before recovering, is a very different experience than 42% with a maximum 15% drawdown.
Risk metrics give you the complete picture. Before you put real money into any strategy, learn to read the six below. Anyone who shows you only CAGR is leaving out the part that decides whether you can live with it.
how far the Nifty 50 fell, close to close, from January to 23 March 2020
NSE index closes
peak to trough fall of the Nifty200 Momentum 30, two years to 9 Oct 2026
NSE index closes
10 year government bond yield, a common risk free rate in India
Market close, 1 Oct 2026
RBI repo rate after the 7 Oct 2026 hike
RBI monetary policy, 7 Oct 2026
- Max drawdownCan you survive the worst fall?
- CAGRIs the reward worth that fall?
- Sharpe and SortinoReward per unit of pain.
- BetaHow much is just the market?
- AlphaWhat is left that is genuinely extra?
That is the order I read a back test in. Drawdown first, not CAGR. Most sales decks reverse it because CAGR is the cleanest number to market. Drawdown is the one that decides whether you stay invested long enough to collect the CAGR.
The six metrics you need to know
Reading metrics together
| Metric | What it answers | Minimum bar | Excellent |
|---|---|---|---|
| CAGR | How fast did money grow? | > Nifty 500 CAGR | > Nifty 500 + 8% |
| Max Drawdown | What was the worst loss? | < 45% | < 25% |
| Sharpe Ratio | Return per unit volatility? | > 0.8 | > 1.3 |
| Sortino Ratio | Return per unit downside? | > 1.0 | > 1.8 |
| Beta | Market sensitivity? | Understood & intentional | Context-dependent |
| Alpha | Value above market? | > 0% annualised | > 5% annualised |
The Allcap Multi Asset model line grows far faster than the index line, with a fall of 21.3% at its worst; the index line is a smoothed CAGR path.
Model, back tested. Monthly values from lineup.json. The benchmark line is a smoothed CAGR path, not the daily index. Not a promise of future returns.
Drag along that line. The model curve does not go up in a straight line. Its worst fall, 21.3% from December 2024, took about a year to recover. If you had been in it, that year is the only part you would remember. The CAGR is the reward. The drawdown is the price.
A worked example
Real numbers this time. Here are three strategies from the RupeeCase lineup, all model, back tested over the same window, as of 27 April 2026.
| Strategy | CAGR | Max drawdown | Sharpe | Risk band |
|---|---|---|---|---|
| Allcap | 44.70% | −22.72% | 1.68 | Moderately High |
| Allcap Multi Asset | 40.47% | −21.32% | 1.81 | Moderate |
| Energy Basket | 33.12% | −39.48% | 1.12 | High |
| Nifty 50 Equal Weight | 16.35% | −15.71% | 1.13 | Moderate |
Model, back tested. As of 27 Apr 2026, from the RupeeCase lineup data. Not client returns and not a promise of future returns.
Sharpe ratio, model, back tested
- Allcap Multi Asset1.81
- Allcap1.68
- Nifty 50 Equal Weight1.13
- Energy Basket1.12
Model, back tested. As of 27 Apr 2026. Not a promise of future returns.
Read them in my order. Drawdown first: Energy Basket fell almost 40% at its worst, nearly twice the others. CAGR next: it earned less than the two Allcap strategies anyway. Sharpe confirms it: the same reward per unit of risk as a plain equal weight Nifty, with two and a half times the drawdown. Now compare Allcap with Allcap Multi Asset. The multi asset version gave up about 4 points of CAGR for a 10% debt and 10% gold sleeve, and got a slightly smaller drawdown and a higher Sharpe. Neither is "the best". Which one fits depends on the fall you can sit through.
The CAGR trap: Plenty of pitches show only CAGR. A 35% CAGR with a 65% max drawdown means you needed to hold through a 65% loss to get that return. Most investors can't and don't, they panic-sell at the bottom, locking in the loss and missing the recovery. A lower CAGR with lower drawdown may be the far better real-world outcome.
Common mistakes with risk metrics
Comparing Sharpe ratios built on different windows. A Sharpe from a bull market is not comparable with one that includes 2020. Check the dates.
Treating max drawdown as the worst case. It is the worst so far. The next one can be deeper.
Using a monthly series and calling it daily risk. Monthly data hides falls that happened inside the month. Know the frequency behind the number.
Forgetting the risk free rate moves. With the 10 year yield at about 7.2% in October 2026, a strategy needs a higher return than in a 5% world to earn the same Sharpe.
MythA higher CAGR always means the better strategy.
FactOnly if you can hold it. A 40% CAGR with a 60% drawdown needs you to watch ₹100 become ₹40 and not sell. Most people sell, and earn the drawdown without the CAGR.
How this shows up in RupeeCase
Every RupeeCase strategy page leads with the same set of numbers: CAGR, Sharpe and max drawdown, all model, back tested and dated, next to the benchmark, plus a risk band from Moderate to Very High. The strategy pages also list the worst drawdowns with their start and recovery dates. The Allcap Multi Asset page is a good place to practise: find the drawdown first, then decide whether the CAGR is worth it to you.
What is a good Sharpe ratio for an Indian equity strategy?
Above 1 is good for long only equity, above 1.5 is strong. Context matters: a Sharpe from a short or unusually calm window flatters any strategy. Always check the period and the risk free rate used.
What is the Calmar ratio?
CAGR divided by the absolute max drawdown. A strategy with 20% CAGR and a 20% max drawdown has a Calmar of 1. It answers a simple question: how many years of return does the worst fall cost?
Why do back tested drawdowns often look smaller than real ones?
Because real trading adds slippage, delays and your own reactions, and because the future can hold a fall the history did not. Treat any back tested drawdown as a minimum, not a maximum.