Learn the Sharpe ratio

HOW do we derive it from first principles?

Formula Build it up, don't memorize it

Step 1 — Reward. You have a choice. Option A: risk-free, get $r_f$ . Option B: risky portfolio, get $R_p$ . Why this step? The only reason to accept risk is the chance of beating $r_f$ . So the meaningful reward is the difference: $\text{Reward} = R_p - r_f$

Step 2 — Cost (risk). Uncertainty of returns is captured by variance $\sigma_p^2$ . But variance is in "squared %" units and can't be compared to a return. Take the square root to get $\sigma_p$ , which lives in the same units as return (%). Why this step? We want reward and risk in comparable units so their ratio is dimensionless. $\text{Risk} = \sigma_p$

Step 3 — Reward per unit risk. Divide: $S = \frac{R_p - r_f}{\sigma_p}$ Why this step? A ratio lets us rank portfolios of any size. Doubling your position doubles both $R_p-r_f$ and $\sigma_p$ (leverage scales both), so $S$ stays constant — Sharpe measures the quality of the risk, not the quantity.

Annualizing (the sneaky detail)

Worked Examples

Common Mistakes (Steel-manned)

Limitations (know when it lies)

Active Recall

Recall Test yourself (hide the answers)

What does the numerator of the Sharpe ratio represent, and why subtract $r_f$ ?
Why standard deviation, not variance, in the denominator?
Two funds return 12%. How can one still be "better"?
How do you annualize a monthly Sharpe ratio? Why the square root?
Geometrically, what is the Sharpe ratio on a return-vs-risk plot?

Recall Feynman: explain to a 12-year-old

Imagine two ice-cream sellers both make ₹100 profit a day. One sells calmly and always makes about ₹100. The other some days makes ₹300 and some days loses ₹100 — super jumpy. Both average ₹100, but the calm one is smarter because you can trust it. The Sharpe ratio is a score: take the extra money you made above what a totally safe piggy bank gives, then divide by how "jumpy" your earnings were. Bigger score = you earned good money without the scary rollercoaster.

Connections

Capital Market Line — the steepest Sharpe line from $r_f$ .
Tangency Portfolio — the portfolio with maximum Sharpe ratio.
Efficient Frontier — Sharpe picks the best point on it.
Standard Deviation as Risk — the denominator's foundation.
Risk-Free Rate — the baseline the excess return is measured from.
Sortino Ratio — Sharpe's downside-only cousin.
Capital Asset Pricing Model — where excess-return thinking generalizes.

Sharpe ratio formula

S = (R_p - r_f)/\sigma_p

— excess return over standard deviation.

Why subtract the risk-free rate?

Only return above

r_f

is a genuine reward for taking risk;

r_f

is available risk-free.

Why divide by std dev, not variance?

Std dev shares the same units as return, so the ratio is dimensionless (reward per unit risk).

Two funds, same 12% return — how can one be better?

The one with lower

\sigma_p

has a higher Sharpe ratio (more reward per unit risk).

How to annualize a monthly Sharpe ratio?

Multiply by

\sqrt{12}

(generally

\sqrt{n}

for

n

periods/year).

Why

\sqrt{n}

and not

n

when annualizing?

Return scales with

n

, std dev with

\sqrt{n}

; ratio scales with

n/\sqrt{n}=\sqrt{n}

Geometric meaning of Sharpe ratio?

Slope of the line from

(0, r_f)

(\sigma_p, R_p)

on a return-vs-risk plot.

What portfolio maximizes the Sharpe ratio?

The tangency portfolio (touches the efficient frontier, defines the Capital Market Line).

A Sharpe-ratio alternative for skewed returns?

The Sortino ratio, which divides by downside deviation only.

Main assumption/weakness of Sharpe?

Assumes symmetric (normal) returns; penalizes upside volatility and can be gamed by tail-risk selling.

Concept Map

Hinglish (regional understanding)

Intuition Hinglish mein samjho

Dekho, Sharpe ratio ek simple sawaal ka jawaab hai: "Jitna risk main le raha hoon, uske badle mujhe kitna extra return mil raha hai?" Sirf return dekhna dhoka hai — agar ek fund 20% deta hai par roz upar-neeche uchhalta hai, aur doosra 15% deta hai bilkul shaant tareeke se, to shaant wala often behtar hota hai. Isliye hum return me se risk-free rate ( $r_f$ , jaise FD ya T-bill ka return) minus karte hain, kyunki utna to bina risk ke bhi mil jaata. Yeh "excess return" hi asli reward hai.

Formula hai $S = (R_p - r_f)/\sigma_p$ . Upar excess return, neeche standard deviation ( $\sigma_p$ ) jo batati hai returns kitna wobble karte hain. Std dev use karte hain, variance nahi, kyunki std dev return ke same units (%) me hoti hai — tabhi ratio ka matlab banta hai "har ek unit risk ke badle kitna reward". Zyada Sharpe = better deal. Graph pe socho: risk-free point se portfolio tak jo line jaati hai, uski slope hi Sharpe ratio hai — jitni steep line, utna accha.

Ek important trick: agar aap monthly data se Sharpe nikaalte ho, to annual banane ke liye $\sqrt{12}$ se multiply karo, seedha 12 se nahi. Kyunki return to 12 guna badhta hai par risk sirf $\sqrt{12}$ guna, to ratio $\sqrt{12}$ guna badhta hai. Yeh chhoti si galti bahut log karte hain.

Yaad rakhna: Sharpe maan leta hai ki returns symmetric (normal-ish) hain, aur woh upar ke volatility ko bhi risk maanta hai. Jahan returns skewed hon (jaise crash-prone strategies), wahan Sortino ratio behtar hai. Par basic portfolio comparison ke liye Sharpe hi sabse pehla aur sabse zaroori tool hai — 80/20 me yeh must-know hai.

WHY does this metric exist?

WHAT is it — the definition

HOW do we derive it from first principles?

Annualizing (the sneaky detail)

Worked Examples

Common Mistakes (Steel-manned)

Limitations (know when it lies)

Active Recall

Connections

Concept Map

Hinglish (regional understanding)

Test yourself — Portfolio Theory

Connections