Quantitative Reasoning Interview Prep

Use this as an interview speaking guide, not as a formula sheet.

For almost every problem:

1. Clarify assumptions.
2. Identify the invariant or governing formula.
3. Solve symbolically.
4. Substitute numbers.
5. Sanity-check the result.
6. Explain the intuition.

---

1. Regular Hexagon: Find Side Length From Area

Prompt

Given the area A of a regular hexagon, find side length s.

Key idea

A regular hexagon contains 6 equilateral triangles.

Area of one equilateral triangle:

triangleArea = (sqrt(3) / 4) * s²

Hexagon area:

A = 6 * triangleArea

A = (3 * sqrt(3) / 2) * s²

Solve for s:

s² = 2A / (3 * sqrt(3))

s = sqrt(2A / (3 * sqrt(3)))

JavaScript

function hexagonSideFromArea(area) {
  return Math.sqrt((2 * area) / (3 * Math.sqrt(3)));
}

Example

const area = 54 * Math.sqrt(3);

console.log(hexagonSideFromArea(area));
// 6

Answer

s = sqrt(2A / (3√3))

Sanity check

Area ∝ side²

Therefore:

side ∝ sqrt(area)

That matches the formula.

What to say aloud

A regular hexagon is six equilateral triangles. I write the triangle-area formula, multiply by six, and solve for the side length.

---

2. Why Is Summer Hotter Than Winter?

Short answer

Earth's axial tilt ≈ 23.5°

NOT primarily the Earth-Sun distance.

Two effects matter:

1. Sunlight hits the ground more directly.
2. Summer days are longer.

Incidence angle

Suppose θ is the angle between sunlight and the vertical surface normal.

relativeIntensity = cos(θ)

JavaScript:

function relativeSolarIntensity(degrees) {
  const radians = (degrees * Math.PI) / 180;
  return Math.cos(radians);
}

Examples:

relativeSolarIntensity(0);
// 1.00

relativeSolarIntensity(30);
// 0.866

relativeSolarIntensity(60);
// 0.50

relativeSolarIntensity(75);
// 0.259

So at:

60° from vertical

cos(60°) = 0.5

The same incoming sunlight is spread across approximately twice the ground area.

Why summer gets even hotter

Approximate daily solar energy:

dailyEnergy ≈ solarIntensity × daylightHours

Summer gets:

higher intensity
×
more daylight hours

The effects compound.

Important interview trap

Earth is actually closest to the Sun around early January.

If Earth-Sun distance caused seasons:

Northern Hemisphere
and
Southern Hemisphere

would experience summer together.

They do not.

What to say aloud

Seasons come from Earth's axial tilt. In summer, sunlight arrives closer to perpendicular, so more energy reaches each square meter, and the days are longer. The opposite hemisphere experiences winter at the same time, which is strong evidence that Earth-Sun distance isn't the main cause.

---

3. Two Spheres With the Same Density

Prompt

Two spheres are made from the same material.

The larger sphere has a diameter 50% larger.

The smaller sphere weighs:

8 lb

Find the larger sphere's weight.

Key idea

Diameter scale:

k = 1.5

Volume scales with the cube of linear dimensions:

volumeRatio = k³

Therefore:

massRatio = 1.5³
          = 3.375

Then:

largeMass = 8 × 3.375
          = 27 lb

JavaScript

const smallMass = 8;
const scale = 1.5;

const largeMass = smallMass * scale ** 3;

console.log(largeMass);
// 27

Answer

27 lb

What to say aloud

Same material means same density, so mass scales with volume. Sphere volume scales with diameter cubed. A 1.5× diameter therefore means 1.5³ = 3.375× the mass, giving 27 pounds.

---

4. How Much 1080p30 Video Fits on 1 TB?

Prompt

A 1 TB drive stores:

1920 × 1080 video
30 FPS

How many minutes fit?

First thing to say

The problem is underspecified.

Resolution + FPS

does NOT determine compressed video size.

You need either:

bits per pixel

or

compressed bitrate

Ask:

Should I assume raw RGB video or a compressed bitrate?

Case A: Raw RGB

Assume:

width = 1920
height = 1080
fps = 30
bytesPerPixel = 3
storage = 1 TB = 10¹² bytes

JavaScript

const width = 1920;
const height = 1080;
const fps = 30;
const bytesPerPixel = 3;

const bytesPerFrame = width * height * bytesPerPixel;

const bytesPerSecond = bytesPerFrame * fps;

const driveBytes = 1e12;

const seconds = driveBytes / bytesPerSecond;

const minutes = seconds / 60;

console.log({
  bytesPerFrame,
  bytesPerSecond,
  minutes,
});

Result:

bytes/frame ≈ 6.22 MB

bytes/sec ≈ 186.6 MB/s

duration ≈ 89.3 minutes

Answer for raw RGB

≈ 90 minutes

Case B: Compressed Video

General formula:

durationSeconds =
  storageBits / bitrateBitsPerSecond

For 1 TB:

function videoMinutes(storageTB, bitrateMbps) {
  const storageBits = storageTB * 1e12 * 8;

  const bitrate = bitrateMbps * 1e6;

  return storageBits / bitrate / 60;
}

Examples:

videoMinutes(1, 5);
// ≈ 26667 minutes

videoMinutes(1, 10);
// ≈ 13333 minutes

videoMinutes(1, 20);
// ≈ 6667 minutes

Interview lesson

The most important observation is:

storage depends on encoding / bitrate

not just:

resolution × FPS

---

5. Drop a Dense Stone From a Boat Into Water

Prompt

A dense stone is initially sitting inside a floating boat.

You throw the stone into the water.

The stone sinks.

Does the pool water level:

rise
fall
stay the same

Answer

FALL

Before

While inside the boat, the stone causes the boat to displace water equal to the stone's weight.

displacedVolumeBefore =
  stoneMass / waterDensity

After

When the stone sinks, it displaces only its physical volume.

displacedVolumeAfter =
  stoneMass / stoneDensity

Because:

stoneDensity > waterDensity

therefore:

stoneMass / stoneDensity
<
stoneMass / waterDensity

So:

after displacement
<
before displacement

Therefore:

water level falls

JavaScript representation

const mass = 10;
const waterDensity = 1000;
const stoneDensity = 2500;

const before = mass / waterDensity;

const after = mass / stoneDensity;

console.log(after < before);
// true

What to say aloud

In the boat, the stone causes displacement based on its weight. Once submerged, it displaces only its own volume. Since the stone is denser than water, its volume is smaller than the volume of water having the same weight, so the water level falls.

---

6. $1000 at 100% Interest for 20 Years

Clarification

Ask whether interest is:

simple

or

compounded

Usually assume annual compounding.

Formula

futureValue =
  principal × (1 + rate)^years

Here:

principal = 1000
rate = 1.0
years = 20

Therefore:

futureValue
= 1000 × 2²⁰

We know:

2²⁰ = 1,048,576

Therefore:

$1,048,576,000

JavaScript

const principal = 1000;
const rate = 1;
const years = 20;

const amount = principal * (1 + rate) ** years;

console.log(amount);
// 1048576000

Answer

≈ $1.05 billion

Interview concept

This tests exponential growth.

At 100% annual interest:

money doubles every year

---

7. Randomly Choose a d6 or d8

Setup

Choose randomly between:

D6
D8

So:

P(D6) = 1/2
P(D8) = 1/2

Part A: Probability of Rolling a 3

P(3 | D6) = 1/6

P(3 | D8) = 1/8

Total probability:

P(3)
=
P(D6) × P(3 | D6)
+
P(D8) × P(3 | D8)

Substitute:

= (1/2)(1/6)
+ (1/2)(1/8)

= 1/12 + 1/16

= 4/48 + 3/48

= 7/48

Answer

7/48

Part B: Given a 3, Probability It Was the d6

Bayes:

P(D6 | 3)
=
P(3 | D6) P(D6)
-----------------
       P(3)

Substitute:

=
(1/6 × 1/2)
-------------
    7/48

=
1/12
-----
7/48

=
4/7

Answer

P(D6 | 3) = 4/7

P(D8 | 3) = 3/7

Part C: Expected Value of the Next Roll

Expected d6:

E[D6] = (1 + 6) / 2 = 3.5

Expected d8:

E[D8] = (1 + 8) / 2 = 4.5

Using posterior probabilities:

E[next]
=
(4/7)(3.5)
+
(3/7)(4.5)

JavaScript:

const expected = (4 / 7) * 3.5 + (3 / 7) * 4.5;

console.log(expected);
// 3.928571...

Exact:

55/14

Approximate:

3.93

---

8. Probability the 13th Is a Friday

Simple interview assumption

If weekdays are uniformly distributed:

P(Friday) = 1/7

Approximate:

14.29%

Exact Gregorian Calendar

The Gregorian calendar repeats every:

400 years

That contains:

400 × 12
= 4800 months

Friday occurs as the 13th:

688 times

Therefore:

688 / 4800
= 43 / 300
≈ 14.333%

Strong interview answer

Assuming weekday alignment is uniform, the answer is 1/7. If you're asking for the exact Gregorian-calendar frequency, it's 43/300, or about 14.33%.

---

9. 3D Tic-Tac-Toe Winning Lines

Important clarification

For:

3 × 3 × 3

the answer is:

49

For:

4 × 4 × 4

with four cells required:

76

If the expected answer is 49, assume a 3×3×3 board.

Count 3×3×3 Winning Lines

Axis-parallel

Three directions:

x
y
z

Each has:

3 × 3 = 9

So:

3 × 9 = 27

Face diagonals

6 faces
×
2 diagonals
=
12

Running total:

27 + 12 = 39

Middle-plane diagonals

Three central planes:

xy
xz
yz

Each contributes:

2

Therefore:

3 × 2 = 6

Running total:

45

Space diagonals

A cube has:

4

body diagonals.

Total:

27 + 12 + 6 + 4

= 49

Answer

49

---

10. Distance From a Point to a Plane

Plane:

Ax + By + Cz + D = 0

Point:

P = (x0, y0, z0)

Distance:

        |Ax0 + By0 + Cz0 + D|
d = -----------------------------
          sqrt(A² + B² + C²)

JavaScript

function pointToPlaneDistance(point, plane) {
  const { x, y, z } = point;
  const { A, B, C, D } = plane;

  const numerator = Math.abs(A * x + B * y + C * z + D);

  const denominator = Math.sqrt(A ** 2 + B ** 2 + C ** 2);

  return numerator / denominator;
}

Intuition

(A, B, C)

is the plane's normal vector.

The formula measures how far the point extends along that perpendicular direction.

---

11. Estimate Earth-Moon Distance

Useful rough value:

≈ 384,400 km

Interview estimate:

≈ 4 × 10⁵ km

Fermi approach

Earth radius:

≈ 6400 km

Moon distance:

≈ 60 Earth radii

Therefore:

60 × 6400
=
384,000 km

JavaScript

const earthRadius = 6400;
const earthRadiiToMoon = 60;

console.log(earthRadius * earthRadiiToMoon);

// 384000

Light-time sanity check

Speed of light:

≈ 300,000 km/s

Therefore:

384,000 / 300,000
≈ 1.28 seconds

---

12. Expected Number of Times max Changes

Consider:

let max = -Infinity;

for (const value of values) {
  if (value > max) {
    max = value;
  }
}

Suppose values is a random permutation of n distinct values.

How many times does max update on average?

Key observation

At position i, the current element becomes the new maximum if it is the largest among the first i elements.

Each of those i positions is equally likely to contain that maximum.

Therefore:

P(update at position i) = 1/i

Expected updates:

E =
1
+ 1/2
+ 1/3
+ ...
+ 1/n

This is the harmonic number:

Hn

Approximation:

Hn ≈ ln(n) + 0.577

JavaScript

function expectedMaxUpdates(n) {
  let result = 0;

  for (let i = 1; i <= n; i++) {
    result += 1 / i;
  }

  return result;
}

For one million:

expectedMaxUpdates(1_000_000);
// ≈ 14.39

Approximation:

Math.log(1_000_000) + 0.57721;
// ≈ 14.39

Answer

E = Hn ≈ ln(n) + γ

Interesting intuition:

1,000,000 values

but max changes only
≈ 14 times on average.

---

13. Buffon's Needle

Setup

Parallel lines are separated by:

D

Needle length:

L

Assume:

L <= D

Probability of crossing a line:

P = 2L / (πD)

If:

L = D = 1

then:

P = 2/π
  ≈ 0.637

JavaScript

function buffonProbability(length, spacing) {
  return (2 * length) / (Math.PI * spacing);
}

buffonProbability(1, 1);
// ≈ 0.63662

Important correction

For the standard Buffon's Needle problem:

L = D

P = 2/π

not:

1/π

Short intuition

For an angle θ, crossing occurs when the center is close enough to a line:

distanceToLine
<=
(L / 2) × sin(θ)

Average this over all orientations and positions, giving:

2L / (πD)

---

14. Match a Small 2D Point Set Inside a Larger Set

Problem

Given:

small template S

large target T

Allowed transformation:

rotation
+
translation

No scaling.

Find the best alignment.

Case A: Correspondence Is Known

Suppose:

p[i] ↔ q[i]

We want:

q[i] ≈ R × p[i] + t

where:

R = rotation
t = translation

Algorithm:

1. Compute source centroid.
2. Compute target centroid.
3. Center both point sets.
4. Solve optimal rotation.
5. Compute translation.

Translation:

t = targetCentroid
  - R × sourceCentroid

Rotation can be found with:

SVD
Kabsch algorithm

Case B: Correspondence Is Unknown

This is harder.

Rotation and translation preserve:

distances
angles

Use those properties to generate candidate matches.

Strong architecture

Small Template
      │
      ▼
Compute invariant features
distances / angles
      │
      ▼
Find candidate pairs
inside large target
      │
      ▼
Infer rotation + translation
      │
      ▼
Transform template
      │
      ▼
KD-tree / spatial hash lookup
      │
      ▼
Count matching points
      │
      ▼
RANSAC best candidate
      │
      ▼
Optional ICP refinement

RANSAC

Useful with noise and outliers.

1. Pick candidate anchors.
2. Infer transform.
3. Transform template.
4. Count inliers.
5. Repeat.
6. Keep best transform.

Pros

robust to outliers
easy to parallelize
easy interview explanation

Cons

probabilistic
may require many iterations

ICP

Iterative Closest Point:

initial transform
      ↓
nearest neighbors
      ↓
solve rigid transform
      ↓
repeat

Pros

excellent local refinement

Cons

can converge to local optimum
needs decent initialization

Interview recommendation

Invariant matching
→ RANSAC
→ KD-tree verification
→ ICP refinement

---

15. Seattle Rain + Three Friends

Setup

Probability of rain:

P(R) = 0.25

Probability of no rain:

P(!R) = 0.75

Each friend tells the truth with probability:

2/3

All three independently say:

"It's raining."

Find:

P(rain | all 3 say rain)

If it actually rains

Each tells the truth with probability:

2/3

All three say rain:

(2/3)³
=
8/27

If it does not rain

Each must lie:

1/3

All three say rain:

(1/3)³
=
1/27

Bayes calculation

Compare weighted likelihoods.

Rain:

P(R) × P(YYY | R)

= 1/4 × 8/27

= 8/108

No rain:

P(!R) × P(YYY | !R)

= 3/4 × 1/27

= 3/108

Normalize:

P(R | YYY)

= 8 / (8 + 3)

= 8/11

Answer

8/11 ≈ 72.7%

JavaScript

const rain = 0.25 * (2 / 3) ** 3;

const noRain = 0.75 * (1 / 3) ** 3;

const posterior = rain / (rain + noRain);

console.log(posterior);
// 0.72727...

---

16. Packing Equal Circles on an Infinite Plane

Clarify the problem

If circles may overlap and the question is simply:

Can circles cover the plane?

then:

100%

is possible.

But if the intended question is:

What is the maximum fraction of the plane occupied by equal, non-overlapping circles?

then this is the hexagonal circle-packing problem.

Result

Maximum density:

π / (2√3)

Equivalent:

π / √12

Approximate:

0.9069

or:

90.69%

JavaScript

const density = Math.PI / (2 * Math.sqrt(3));

console.log(density);
// 0.906899...

Intuition

Optimal centers form a triangular / hexagonal lattice:

   ○   ○   ○

     ○   ○

   ○   ○   ○

This packs circles more densely than a square grid.

---

17. 4×4 Grid of Points: How Many Squares?

Assume:

4 × 4 lattice of points

16 total points

Count:

axis-aligned
+
rotated squares

Axis-Aligned Squares

1×1

3 × 3 = 9

2×2

2 × 2 = 4

3×3

1 × 1 = 1

Total:

9 + 4 + 1
= 14

Rotated Squares

There are:

8

valid rotated squares.

Therefore:

14 + 8
= 22

Answer

22 squares

Probability Four Random Points Form a Square

Number of ways to choose 4 points from 16:

C(16, 4)

JavaScript:

function combination(n, k) {
  let result = 1;

  for (let i = 1; i <= k; i++) {
    result *= (n - i + 1) / i;
  }

  return result;
}

console.log(combination(16, 4));
// 1820

There are:

22

sets that form squares.

Therefore:

P = 22 / 1820

Simplify:

11 / 910

Approximate:

1.21%

---

18. Find the Pattern in Number of Edges

The referenced image is missing, so the exact sequence cannot be solved.

But this is the framework to use.

Step 1: Write the sequence

E1, E2, E3, E4, ...

Step 2: First differences

ΔE1 = E2 - E1
ΔE2 = E3 - E2
...

If constant:

linear sequence

En = an + b

Step 3: Second differences

If first differences are not constant:

calculate second differences

Constant second difference often means:

quadratic

En = an² + bn + c

Step 4: Prefer structural counting

Instead of guessing a numeric pattern:

total edges
=
edges introduced
-
shared edges

Example: squares arranged in a row.

First square:

4 edges

Every additional square shares one side and therefore adds:

3 new edges

So:

En
=
4 + 3(n - 1)

=
3n + 1

JavaScript

function edgesForSquaresInRow(n) {
  return 3 * n + 1;
}

---

Fast Review Sheet

┌─────────────────────────────┬─────────────────────────────────────────┐
│ Problem                     │ Key Answer                              │
├─────────────────────────────┼─────────────────────────────────────────┤
│ Regular hexagon             │ s = sqrt(2A / (3√3))                   │
│ Summer vs winter            │ axial tilt + incidence angle + day     │
│ Sphere 50% larger diameter  │ 8 × 1.5³ = 27 lb                      │
│ 1 TB 1080p30                │ need bitrate; raw RGB ≈ 89 min          │
│ Stone from boat             │ water level falls                       │
│ $1000 @ 100% for 20 years   │ $1,048,576,000                          │
│ P(roll 3 with d6/d8)        │ 7/48                                    │
│ P(d6 | rolled 3)            │ 4/7                                     │
│ Expected next die roll      │ 55/14 ≈ 3.93                            │
│ Friday the 13th             │ ~1/7; exact Gregorian = 43/300          │
│ 3×3×3 tic-tac-toe           │ 49 winning lines                        │
│ Point-plane distance        │ |Ax+By+Cz+D| / √(A²+B²+C²)            │
│ Earth-Moon distance         │ ≈ 384,400 km                            │
│ Expected max updates        │ Hn ≈ ln(n) + γ                          │
│ Buffon's Needle L=D         │ 2/π ≈ 0.637                             │
│ Point-set matching          │ RANSAC + KD-tree + ICP                   │
│ Seattle rain                │ 8/11 ≈ 72.7%                            │
│ Hex circle packing          │ π/(2√3) ≈ 90.69%                       │
│ 4×4 lattice squares         │ 22                                      │
│ P(4 points form square)     │ 11/910 ≈ 1.21%                          │
└─────────────────────────────┴─────────────────────────────────────────┘

---

Interview Strategy

For a quantitative problem, say your reasoning in this order:

Clarify
   ↓
Find invariant / formula
   ↓
Solve symbolically
   ↓
Substitute values
   ↓
Check units
   ↓
Check magnitude
   ↓
Explain intuition

Strong Clarification Examples

Video storage

Resolution and frame rate don't determine compressed storage by themselves. Should I assume raw RGB or a specific bitrate?

Circle problem

Do you mean covering, where circles may overlap, or packing equal non-overlapping circles?

Tic-tac-toe

I'll assume this is a 3×3×3 board where three aligned cells win, since that corresponds to 49 winning lines.

Interest

Should I assume the 100% interest compounds annually?

---

High-Value Concepts

1. Scaling Laws

Length:

L → kL


Area:

A → k²A


Volume:

V → k³V


Mass at constant density:

m → k³m

Memorize:

50% bigger diameter

does NOT mean

50% heavier.

Instead:

1.5 ** 3;
// 3.375

2. Bayes' Theorem

Conceptually:

posterior
∝
likelihood × prior

Full formula:

P(A | B)
=
P(B | A) × P(A)
----------------
      P(B)

Useful mental model:

Prior
  ↓
Evidence likelihood
  ↓
Reweight possibilities
  ↓
Normalize
  ↓
Posterior

3. Expected Value

For discrete values:

E[X]
=
Σ value × probability

Example:

function expectedValue(outcomes) {
  return outcomes.reduce((sum, { value, probability }) => sum + value * probability, 0);
}

4. Linearity of Expectation

Extremely useful:

E[X + Y]

=

E[X] + E[Y]

Independence is not required.

This is why the expected-max-update problem is easy:

E[updates]

=
P(update at 1)
+
P(update at 2)
+
...
+
P(update at n)

5. Harmonic Numbers

Hn
=
1 + 1/2 + 1/3 + ... + 1/n

Approximation:

Hn ≈ ln(n) + 0.577

Examples:

Math.log(1_000) + 0.577;
// ≈ 7.48

Math.log(1_000_000) + 0.577;
// ≈ 14.39

Math.log(1_000_000_000) + 0.577;
// ≈ 21.30

Huge input sizes can therefore still produce surprisingly small harmonic expectations.

6. Geometry Formulas Worth Knowing

Equilateral triangle

A = √3 / 4 × s²

Regular hexagon

A = 3√3 / 2 × s²

Sphere

V = 4/3 × πr³

Point-to-plane

        |Ax0 + By0 + Cz0 + D|
d = -----------------------------
          √(A² + B² + C²)

7. Circle Packing

Maximum equal-circle packing density:

π / (2√3)

≈ 0.9069

≈ 90.69%

Think:

hexagonal / triangular lattice

not:

square grid

8. Buffon's Needle

For:

L <= D

probability of crossing:

P = 2L / (πD)

Special case:

L = D

P = 2/π

≈ 63.7%

---

Common Interview Traps

Trap 1: Linear vs Cubic Scaling

Wrong:

diameter +50%
→
mass +50%

Correct:

diameter ×1.5
→
mass ×1.5³
→
mass ×3.375

Trap 2: Calculating Before Clarifying

Bad:

1080p30
→ immediately calculate disk usage

Better:

"What bitrate or encoding should I assume?"

Interviewers often intentionally leave information out.

Trap 3: Weight vs Volume Displacement

Stone inside boat:

displacement based on weight

Stone underwater:

displacement based on physical volume

That distinction determines the answer.

Trap 4: Forgetting Bayesian Updating

Before observing a 3:

P(D6) = 1/2
P(D8) = 1/2

After observing a 3:

P(D6 | 3) = 4/7
P(D8 | 3) = 3/7

The probabilities changed because a 3 is more likely on the smaller die.

Trap 5: Trusting the Supplied Answer

If an interview prompt says:

Buffon's Needle answer = 1/π

do not blindly accept it.

Standard assumptions give:

2/π

State your assumptions and derive the result.

---

60-Second Mental Math Sheet

√3 ≈ 1.732

π ≈ 3.14

2/π ≈ 0.637

1/π ≈ 0.318

ln(10) ≈ 2.303

2¹⁰ = 1024 ≈ 10³

2²⁰ = 1,048,576 ≈ 10⁶

Useful scaling:

1.5² = 2.25

1.5³ = 3.375

2³ = 8

10³ = 1000

Useful physical values:

Earth radius
≈ 6400 km

Earth → Moon
≈ 60 Earth radii
≈ 384,000 km

Speed of light
≈ 300,000 km/s

Earth → Moon light time
≈ 1.28 sec

---

Final Interview Cheat Pattern

When you see a math or physics question, think:

┌─────────────────────┐
│ What is unspecified?│
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ What stays invariant?│
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ What scaling law or │
│ formula applies?    │
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ Solve symbolically  │
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ Plug in numbers     │
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ Check units and     │
│ order of magnitude  │
└──────────┬──────────┘
           ↓
┌─────────────────────┐
│ Explain intuition   │
│ in one sentence     │
└─────────────────────┘

The goal is not just:

get the answer

The goal is to demonstrate:

clarification
+
modeling
+
reasoning
+
sanity checking
+
clear communication