The rule everyone dreads

If you have a function made by dividing one function by another, say f(x) divided by g(x), its derivative is given by the quotient rule:

That is the whole rule. It looks harmless on one line. In practice, it is where marks get lost: a sign flipped, a square forgotten, an algebra slip while expanding the top. After many years of tutoring and teaching this, I can tell you that most of the errors come from using the rule when you did not have to.

The good news is that the quotient rule is not a new idea. It is the product rule and a bit of algebra wearing a different coat, and that insight is what lets you skip it so often. Let us start by getting the rule itself solid, then earn the right to avoid it.

First, a trap: you cannot divide the derivatives

When you meet the product rule or the quotient rule for the first time, a natural thought arises: surely the derivative of f divided by g is just f′ divided by g′? It would be so convenient. It is also wrong, and a two-line test shows it.

Take f = x² and g = x. Their quotient is x²/x = x, whose derivative is clearly 1. But f′/g′ is 2x/1 = 2x. At x = 3, the true slope is 1 and the “divide the derivatives” answer is 6. They disagree, so the shortcut is false. Derivatives of fractions need their own rule, and that rule is where the extra terms in the formula come from.

The rule in a picture

The formula has four ingredients: the top f, the bottom g, and their derivatives f′ and g′. A well-known chant helps you keep them in order. Call the top “high” and the bottom “low”:

Say it aloud a few times: low d-high minus high d-low, over low squared. It sounds silly, but it works, and it pins down the part that hurts students most. The first product in the numerator is the derivative of the top times the bottom. Swap the order and the sign of your whole answer flips.

A full worked example

Let us differentiate y = (x² + 1) / (x − 3) with the rule, slowly. I always begin by writing the four ingredients in a small table, because it prevents most slips.

Now fill in the formula:

• Numerator: f′g − fg′ = 2x(x − 3) − (x² + 1)(1).
• Expand: 2x² − 6x − x² − 1.
• Simplify: x² − 6x − 1.
• Denominator: (x − 3)², kept as a square.

So the derivative is (x² − 6x − 1) / (x − 3)². At x = 2, that is (4 − 12 − 1)/1 = −9. I checked it against a numerical slope in the code later in this article, and it matches. Notice that the top of the answer will not factor nicely and the bottom will not cancel. This is a fraction that genuinely needs the rule, and we will use it as our example of “a case you cannot skip.”

Where the rule comes from

Understanding the origin removes the fear. Suppose y = f/g. Multiply both sides by g and you get y · g = f. That is a product of two functions, y and g, and we know how to differentiate a product:

Now solve for y′. Move the y·g′ term across and divide by g:

Put y = f/g back in, so the numerator becomes f′ − (f/g)g′. Multiply the top and bottom of the whole fraction by g to clear the little fraction, and you get exactly (f′g − fg′)/g². That is the quotient rule, derived in four lines from the product rule.

There is a second way to see it. Write f/g as f · g⁻¹, a product with a negative power. The product rule gives f′g⁻¹ + f · (−1)g⁻²g′, and putting that over a common denominator gives the same formula again. The practical meaning is liberating: anything the quotient rule can do, you can also do with the product rule and a negative exponent. If you only ever remember one of the two, the product rule is the safer one to keep.

Why you can often skip it

Here is the heart of this article. The quotient rule is the tool for the hardest case, where a fraction has a complicated top and a complicated bottom that do not simplify. But a large share of the fractions you will meet in homework, exams and real formulas are easier than that. They can be rewritten, and once rewritten they need only the power rule and the constant multiple rule from earlier lessons.

I use a four-question checklist. Ask the questions in this order and stop at the first “yes”.

Shortcut 1: the denominator is just a number

If you are dividing by a constant, you are really multiplying by a constant, and the constant multiple rule says a number out in front simply comes along for the ride.

Shortcut 2: the numerator is a number, or the denominator is a single power of x

Fractions like 12/x² are secretly powers with negative exponents. Rewrite and use the power rule.

The same idea handles a numerator with several terms over a single power. Split it into separate fractions first.

Shortcut 3: split and cancel

When the denominator is a single term, you can divide it into every term of the numerator and cancel as you go.

Shortcut 4: factor the top and cancel

Sometimes the numerator and denominator share a factor, and the fraction collapses to something simple. Always look for this before reaching for the rule.

A small point of honesty here. The original function has a hole at x = 3, because you cannot divide by zero, so the neat answer applies at every x except 3. For most coursework that is fine, but if your question is about behaviour exactly at x = 3, remember the hole.

Shortcut 5: use a negative exponent with the product rule

This is the “quotient rule in disguise” trick. If the denominator is a simple power but the top is complicated, rewrite the fraction as a product with a negative exponent and use the product rule. Many students find it less error-prone than memorising the quotient formula, because the product rule has no minus sign to flip.

A side-by-side comparison

Here is a summary you can keep. Each fraction is paired with the quickest method and the final answer.

Five of the seven rows needed no quotient rule at all. In my experience that ratio is about right for typical textbook sets, and it is one reason I tell students to try simplification first.

Proof by computer: both routes agree

I do not like asking you to trust me, so I tested every row. The program below computes each derivative three ways at x = 2: a numerical slope from two nearby points, the quotient rule applied mechanically, and the shortcut answer from the table.

def slope(f, x, h=1e-6):
    return (f(x + h) - f(x - h)) / (2 * h)          # numeric derivative

def quotient_rule(f, fp, g, gp, x):
    """(f/g)' = (f'g - f g') / g^2"""
    return (fp(x) * g(x) - f(x) * gp(x)) / g(x) ** 2

# name, f, f', g, g', shortcut answer, can we skip the quotient rule?
cases = [
    ("(x^3 + 4x) / 5",   lambda x: x**3 + 4*x,    lambda x: 3*x**2 + 4,       lambda x: 5,      lambda x: 0,     lambda x: (3*x**2 + 4) / 5,  "skip: constant denominator"),
    ("12 / x^2",         lambda x: 12,            lambda x: 0,                lambda x: x**2,   lambda x: 2*x,   lambda x: -24 / x**3,        "skip: write as 12x^-2"),
    ("(4x^2+3x+2) / x^2",lambda x: 4*x**2+3*x+2, lambda x: 8*x+3,            lambda x: x**2,   lambda x: 2*x,   lambda x: -3 / x**2 - 4 / x**3, "skip: split into terms"),
    ("(x^3+5x^2-7x) / x",lambda x: x**3+5*x**2-7*x, lambda x: 3*x**2+10*x-7,  lambda x: x,      lambda x: 1,     lambda x: 2*x + 5,           "skip: split and cancel"),
    ("(x^2 - 9) / (x-3)",lambda x: x**2 - 9,      lambda x: 2*x,              lambda x: x - 3,  lambda x: 1,     lambda x: 1,                 "skip: factor and cancel"),
    ("(x^2 + 1) / (x-3)",lambda x: x**2 + 1,      lambda x: 2*x,              lambda x: x - 3,  lambda x: 1,     lambda x: (x**2 - 6*x - 1) / (x - 3)**2, "use the rule"),
    ("5x / (x^2 + 1)",   lambda x: 5*x,           lambda x: 5,                lambda x: x**2+1, lambda x: 2*x,   lambda x: 5*(1 - x**2) / (x**2 + 1)**2, "use the rule"),
]

x = 2
print(f"At x = {x}")
print(f"{'function':<22}{'numeric':>10}{'quotient':>10}{'answer':>10}   note")
for name, f, fp, g, gp, ans, note in cases:
    num = slope(lambda t: f(t) / g(t), x)
    q = quotient_rule(f, fp, g, gp, x)
    print(f"{name:<22}{num:>10.4f}{q:>10.4f}{ans(x):>10.4f}   {note}")

Output:

At x = 2
function                 numeric  quotient    answer   note
(x^3 + 4x) / 5            3.2000    3.2000    3.2000   skip: constant denominator
12 / x^2                 -3.0000   -3.0000   -3.0000   skip: write as 12x^-2
(4x^2+3x+2) / x^2        -1.2500   -1.2500   -1.2500   skip: split into terms
(x^3+5x^2-7x) / x         9.0000    9.0000    9.0000   skip: split and cancel
(x^2 - 9) / (x-3)         1.0000    1.0000    1.0000   skip: factor and cancel
(x^2 + 1) / (x-3)        -9.0000   -9.0000   -9.0000   use the rule
5x / (x^2 + 1)           -0.6000   -0.6000   -0.6000   use the rule

All three columns agree on every row. That confirms two things at once: the shortcuts are not tricks, they are legitimate, and the quotient rule gives the same value whether it was needed or not. The only difference is how much work it took to get there.

When you really cannot skip it

Now the fair half of the story. Some fractions have a complicated top and a complicated bottom that share nothing, and rewriting them does not help. In that case, use the rule, and use it with confidence. Here are the signs:

• The denominator has more than one term and does not cancel. For instance, x² + 1 or x − 3 sitting underneath something that is not a multiple of it.
• The top is not a constant, and nothing factors. You are stuck with two live functions.
• Trigonometric or exponential parts are involved. The derivative of (sin x)/x is (x cos x − sin x)/x², and there is no simpler route.

Here is my method for these cases, which has saved me from countless slips:

• Write the four-row table of f, f′, g and g′ before doing anything else.
• Put brackets around everything when you substitute, especially around g and f in the subtracted term, since a missing bracket is how minus signs go wrong.
• Expand the numerator only and collect like terms. Leave the denominator as a square, because a factored denominator is usually easier to read and to work with later.
• Check one value numerically. Pick an easy x, evaluate your formula, and compare with a rough slope between two nearby points.

The quotient rule in real life

Fractions of changing quantities appear everywhere: cost per unit, speed, electrical current, pressure and the concentration of a medicine. Click a tab below to see five examples, and notice which ones can skip the rule. This is the whole theme of the article in action.

Checking the real examples with numbers

I checked the five examples from the tabs with the same approach: compute the derivative with the shortcut or the rule, then compare with a purely numerical slope.

def slope(f, x, h=1e-6):
    return (f(x + h) - f(x - h)) / (2 * h)

print("Average cost per unit  AC = (500 + 2q) / q  at q = 50")
print("  skip method  -500/q^2          =", round(-500 / 50**2, 4), "dollars per extra unit")
print("  numeric slope                  =", round(slope(lambda q: (500 + 2*q) / q, 50), 4))
print("Average speed  (40t + 5t^2) / t  at t = 2 h")
print("  skip method  derivative of 40+5t =", 5, "km/h per hour")
print("  numeric slope                  =", round(slope(lambda t: (40*t + 5*t**2) / t, 2), 4))
print("Ohm's law  I = 12 / R  at R = 4 ohms")
print("  skip method  -12/R^2           =", round(-12 / 4**2, 4), "A per ohm")
print("  numeric slope                  =", round(slope(lambda r: 12 / r, 4), 4))
print("Boyle's law  P = 100 / V  at V = 5 L")
print("  skip method  -100/V^2          =", round(-100 / 5**2, 4), "kPa per litre")
print("  numeric slope                  =", round(slope(lambda v: 100 / v, 5), 4))
C = lambda t: 5 * t / (t**2 + 1)
Cp = lambda t: 5 * (1 - t**2) / (t**2 + 1) ** 2
print("Drug concentration  C = 5t / (t^2 + 1)  (quotient rule needed)")
for t in (0.5, 1, 3):
    print(f"  t = {t:<3}  C = {C(t):.3f}   rule C' = {Cp(t):+.4f}   numeric = {slope(C, t):+.4f}")

Output:

Average cost per unit  AC = (500 + 2q) / q  at q = 50
  skip method  -500/q^2          = -0.2 dollars per extra unit
  numeric slope                  = -0.2
Average speed  (40t + 5t^2) / t  at t = 2 h
  skip method  derivative of 40+5t = 5 km/h per hour
  numeric slope                  = 5.0
Ohm's law  I = 12 / R  at R = 4 ohms
  skip method  -12/R^2           = -0.75 A per ohm
  numeric slope                  = -0.75
Boyle's law  P = 100 / V  at V = 5 L
  skip method  -100/V^2          = -4.0 kPa per litre
  numeric slope                  = -4.0
Drug concentration  C = 5t / (t^2 + 1)  (quotient rule needed)
  t = 0.5  C = 2.000   rule C' = +2.4000   numeric = +2.4000
  t = 1    C = 2.500   rule C' = +0.0000   numeric = -0.0000
  t = 3    C = 1.500   rule C' = -0.4000   numeric = -0.4000

The first four were handled by shortcuts and the last needed the rule. The two columns agree each time. The drug example is worth a second look, because it shows what a derivative can tell you that a table of values cannot: the slope is positive at t = 0.5, exactly zero at t = 1, and negative at t = 3. A slope of zero between a rising and a falling stretch marks the peak.

And here is how fast average cost slows down as production grows in the first example. The slope is −500/q², so it flattens quickly:

Common mistakes and how to avoid them

• Dividing the derivatives. (f/g)′ is not f′/g′. We showed this with x²/x earlier.
• Swapping the order in the numerator. It is f′g − fg′, derivative of the top first. Reversed, your whole answer has the wrong sign.
• Forgetting to square the denominator. The bottom of the answer is g², not g.
• Squaring only part of the denominator. If g = x − 3, the denominator is (x − 3)², which is not x² − 9.
• Losing a minus sign while expanding. In f′g − fg′, the minus applies to the whole product fg′. Use brackets.
• Using the rule when a shortcut was available. It is not wrong, just slower and riskier. Run the four-question checklist first.
• Ignoring the hole after cancelling. After simplifying (x² − 9)/(x − 3) to x + 3, remember x = 3 is still excluded from the original function.
• Not simplifying the final answer. Leaving an expanded numerator uncollected costs marks and hides useful structure, such as where the slope is zero.

A quick word on logarithmic differentiation

For fractions that are really long chains of multiplications and divisions, such as a product of five factors over three others, there is a third strategy: take logarithms of both sides first. That turns products and quotients into sums and differences, which are far easier to differentiate. It needs the rules for logarithms and the chain rule, so it belongs to a later lesson, but it is worth knowing it exists. It is the polite way of saying that the quotient rule is not the only tool in the box.

Practice problems with answers

Try each one on paper first, ideally running the checklist before you decide on a method. Then tap to check.

Test yourself: five quick questions

Tap a question to reveal the answer. The correct option is marked.

Frequently asked questions

The post The Quotient Rule, and Why You Can Often Skip It appeared first on Learn With Examples.

The rule in one line

Here is the whole thing. If you have a function f(x) = xⁿ, then its derivative is:

That is the power rule. It turns x² into 2x, x³ into 3x², and x¹⁰ into 10x⁹. It is the first rule most students use to differentiate anything, and it handles an enormous share of the problems you will meet: polynomials, roots, reciprocals and many formulas from physics and economics.

I have taught this to students for many years, and the pattern I see is always the same. The ones who only memorise the rule are fine until the first problem that looks slightly different. The ones who understand where the “n” and the “minus one” come from never get stuck, because they can rebuild the rule on the spot.

First, what is a derivative?

Imagine you are driving and glance at the speedometer. It says 60 km/h. That number is a derivative. It is the rate at which your position is changing, right now, at this instant, not your average over the whole trip.

Graphically, the derivative is the slope of a curve at a single point. For a straight line, the slope is the same everywhere: rise over run. For a curve, such as y = x², the steepness changes as you move along it. Near x = 0 it is almost flat. Further right it climbs more steeply. The derivative gives you the exact steepness at any chosen point.

So when we say “the derivative of x² is 2x”, we mean: at any x, the slope of the curve is twice that x. At x = 3 the slope is 6. At x = 10 it is 20. The curve gets steeper the further you go, and the rule tells you exactly how steep.

Using the rule: three steps

Before we get to why, let us make sure the mechanics are comfortable. There are only three steps.

• Step 1. Identify the exponent n.
• Step 2. Multiply by n.
• Step 3. Subtract 1 from the exponent.

Why does it work? Four ways to see it

Now the real question. I will show you four separate explanations. They all lead to the same place, and different people find different ones convincing. Read them all if you can, and treat the one that clicks as your own.

Explanation 1: just try it with numbers

The honest definition of the derivative is: take two points on the curve that are very close together, draw the line through them, and work out its slope. Then slide the second point closer and closer until it practically sits on top of the first. The slope the lines settle towards is the derivative.

That slope, for two points a small distance h apart, is called the difference quotient: (f(x + h) − f(x)) / h. We do not need any theory to compute it. Here is a program that tries it at x = 3, for the powers 2, 3, 4 and 5, with the gap h shrinking each time:

def diff_quotient(f, x, h):
    """Slope of the line through (x, f(x)) and (x+h, f(x+h))."""
    return (f(x + h) - f(x)) / h

x = 3
print(f"{'h':>8} | " + " | ".join(f"x^{n}" .rjust(9) for n in (2, 3, 4, 5)))
for h in (1, 0.1, 0.01, 0.001, 0.0001):
    row = [diff_quotient(lambda t, n=n: t**n, x, h) for n in (2, 3, 4, 5)]
    print(f"{h:>8} | " + " | ".join(f"{v:9.4f}" for v in row))
print(f"{'power rule':>8} | " + " | ".join(f"{n * x**(n-1):9.4f}" for n in (2, 3, 4, 5)))

This is the output I got:

       h |       x^2 |       x^3 |       x^4 |       x^5
       1 |    7.0000 |   37.0000 |  175.0000 |  781.0000
     0.1 |    6.1000 |   27.9100 |  113.5210 |  432.9151
    0.01 |    6.0100 |   27.0901 |  108.5412 |  407.7090
   0.001 |    6.0010 |   27.0090 |  108.0540 |  405.2701
  0.0001 |    6.0001 |   27.0009 |  108.0054 |  405.0270
power rule |    6.0000 |   27.0000 |  108.0000 |  405.0000

Read down any column. For x³, the slope of the line starts at 37 when h is 1, drops to 27.91, then 27.09, then 27.009, closing in on 27. The bottom row is the power rule’s prediction, 3 × 3² = 27. The same story holds for every column: 6, 27, 108 and 405 are exactly what n · 3ⁿ⁻¹ gives. No proof needed, just a calculator.

This is a real piece of evidence. But it does not say why the answer is n · xⁿ⁻¹ and not something else that happens to match a few numbers. For that, we need the pictures and the algebra.

Explanation 2: the growing square

Picture a square whose sides are each x units long. Its area is x². Now let the sides grow by a tiny amount, call it h, and ask: how much extra area did that add?

Look at the new area as three new pieces. A strip along the right side, which is x long and h wide, so its area is x·h. A strip along the top, with the same area, x·h. And a tiny square in the corner, h by h, with area h².

The total added area is 2xh + h². Divide by h to get the growth per unit of nudge: 2x + h. Now let h shrink towards zero. The leftover h disappears, and we are left with 2x. That is the derivative of x².

Notice why the corner does not matter. If h is 0.01, the strips have area 0.01x each, but the corner is only 0.0001. As h gets smaller, the corner shrinks much faster than the strips. In the limit it vanishes, and only the strips remain. The “2” in 2x is simply the number of strips. There were two edges that could grow, so we got two strips.

Now try a cube with side x. Its volume is x³. Grow each side by h, and the cube sprouts new material on three faces: the right, the top and the front. Each is a flat slab of area x·x = x² and thickness h, so each slab adds x²h. Three slabs make 3x²h. Everything else (the thin edges and the corner cube) involves h² or h³ and vanishes in the limit. So the derivative of x³ is 3x².

Here is the pattern emerging: for a shape with n dimensions each of length x, there are n faces that can grow. Each new slab has size xⁿ⁻¹ (every dimension except the one that grew) times the thickness h. So n slabs of size xⁿ⁻¹h, or n·xⁿ⁻¹ per unit of h. The “n” counts the faces. The “n − 1” counts the dimensions that stayed the same size.

Explanation 3: the algebra (binomial expansion)

Now the same idea in symbols. We want the difference quotient of xⁿ:

The binomial theorem tells us that when you multiply out (x + h)ⁿ, you get:

• xⁿ (the original term),
• plus n · xⁿ⁻¹ · h,
• plus a pile of terms that each contain h² or a higher power of h.

Subtract xⁿ and the original term disappears. Divide by h and every remaining term loses one factor of h:

• The middle term becomes n · xⁿ⁻¹ with no h left.
• Every other term still has at least one h in it.

Now shrink h to zero. All the terms that still have an h vanish. What remains is exactly n · xⁿ⁻¹. That is a complete proof of the power rule for whole-number exponents, and it is only five lines long.

For x² the expansion is x² + 2xh + h², the same three pieces as the square picture. For x³ it is x³ + 3x²h + 3xh² + h³. The coefficient in front of the h term is always n.

Explanation 4: “n ways to choose”

This one is my favourite because it needs no formula at all. Write xⁿ as n copies of x multiplied together:

Now replace x by x + h in every factor. You are multiplying out (x + h)(x + h)(x + h)… and, from each bracket, you must pick either the x or the h. Picking all x gives back xⁿ. Picking exactly one h and the rest x gives a term with a single h and (n − 1) x’s: xⁿ⁻¹h. How many ways can you do that? You can choose the h from the first bracket, or the second, or the third, up to the n-th. That is n ways.

So there are n terms of the form xⁿ⁻¹h. Everything that picks two or more h’s has h² or more, and vanishes in the limit. The change is n · xⁿ⁻¹ · h, and the rate is n · xⁿ⁻¹.

The rule for more than a single power

Real functions are rarely a lone xⁿ. Luckily, three small companion rules extend the power rule to cover almost every polynomial you will see.

The constant rule fits the power rule too. A constant like 9 is 9x⁰, because x⁰ equals 1. Apply the rule: 9 · 0 · x⁻¹ = 0. Nothing changes, so the rate of change is zero, just as you would expect from a flat horizontal line.

Here is a small program that applies the rule to a whole polynomial, term by term, with each term stored as a (coefficient, exponent) pair:

def derivative_power(coef, n):
    """d/dx of coef * x**n  ->  (coef * n) * x**(n - 1)"""
    return coef * n, n - 1

# 3x^4 - 5x^2 + 7x - 9, written as (coefficient, exponent) pairs
poly = [(3, 4), (-5, 2), (7, 1), (-9, 0)]
print("f(x)  =", " ".join(f"{c:+d}x^{n}" for c, n in poly))
d = [derivative_power(c, n) for c, n in poly if n != 0]   # constants vanish
print("f'(x) =", " ".join(f"{c:+d}x^{n}" for c, n in d))

Output:

f(x)  = +3x^4 -5x^2 +7x^1 -9x^0
f'(x) = +12x^3 -10x^1 +7x^0

Reading the printout: x¹ is just x and x⁰ is just 1, so this says the derivative of 3x⁴ − 5x² + 7x − 9 is 12x³ − 10x + 7. The 3 × 4 = 12 and 5 × 2 = 10 come straight from the rule, the 7 survives because the derivative of 7x is 7, and the lone −9 has vanished.

Negative and fractional exponents

This is where the power rule really earns its reputation, because it keeps working when the exponent is not a positive whole number. The trick is to rewrite first, then apply the rule.

Look at the first row. After bringing down the exponent −1, the new exponent is −1 − 1, which is −2, not 0. Subtracting one from a negative number makes it more negative. This is the single most common slip in this topic, and we will return to it in the mistakes section.

The square root case is worth pausing on. The derivative of √x is 1/(2√x). That says the square root curve is very steep near zero and flattens as x grows. It matches what you see on a graph: √x shoots up quickly at first and then slows down. The slope of √x at x = 1 is 1/2, and at x = 100 it is only 1/20.

A note on rigour: the algebra proof above works for whole-number exponents. The rule also holds for negative and fractional exponents, but proving that takes extra tools, namely the quotient rule for negative powers and implicit differentiation or logarithms for fractions. You do not need those to use the rule confidently, and the numeric check from Explanation 1 works on any of these cases if you want reassurance.

The power rule in real life

The rule is not just classroom exercise. Here are five places it turns up. Click a tab to explore each one.

Checking the numbers: five real formulas

I took the five examples from the tabs and checked each one two ways: with the power rule, and with a purely numerical slope (the same two-close-points idea as before, using points on both sides of the target). If the power rule is right, the two answers should agree.

import math

def slope(f, x, h=1e-6):
    return (f(x + h) - f(x - h)) / (2 * h)       # centred difference

g = 9.8
print("Falling object  s = 4.9 t^2, at t = 3 s")
print("  power rule  v = 9.8 t       =", round(g * 3, 4), "m/s")
print("  numeric slope               =", round(slope(lambda t: 4.9 * t**2, 3), 4))
print("Circle  A = pi r^2, at r = 5")
print("  power rule  2 pi r          =", round(2 * math.pi * 5, 4))
print("  numeric slope               =", round(slope(lambda r: math.pi * r**2, 5), 4))
print("Sphere  V = 4/3 pi r^3, at r = 5")
print("  power rule  4 pi r^2        =", round(4 * math.pi * 25, 4))
print("  numeric slope               =", round(slope(lambda r: 4/3 * math.pi * r**3, 5), 4))
print("Kinetic energy  E = 1/2 m v^2, m = 1000 kg, at v = 20 m/s")
print("  power rule  m v             =", 1000 * 20, "J per m/s")
print("  numeric slope               =", round(slope(lambda v: 0.5 * 1000 * v**2, 20), 4))
mu = 0.7
d = lambda v: v**2 / (2 * mu * g)
print("Braking distance  d = v^2 / (2 mu g), mu = 0.7")
print("  at 20 m/s =", round(d(20), 1), "m,  at 40 m/s =", round(d(40), 1), "m")
print("  power rule  v / (mu g) at 20 =", round(20 / (mu * g), 3), "m per m/s")
print("  numeric slope                =", round(slope(d, 20), 3))

Output:

Falling object  s = 4.9 t^2, at t = 3 s
  power rule  v = 9.8 t       = 29.4 m/s
  numeric slope               = 29.4
Circle  A = pi r^2, at r = 5
  power rule  2 pi r          = 31.4159
  numeric slope               = 31.4159
Sphere  V = 4/3 pi r^3, at r = 5
  power rule  4 pi r^2        = 314.1593
  numeric slope               = 314.1593
Kinetic energy  E = 1/2 m v^2, m = 1000 kg, at v = 20 m/s
  power rule  m v             = 20000 J per m/s
  numeric slope               = 20000.0
Braking distance  d = v^2 / (2 mu g), mu = 0.7
  at 20 m/s = 29.2 m,  at 40 m/s = 116.6 m
  power rule  v / (mu g) at 20 = 2.915 m per m/s
  numeric slope                = 2.915

Every pair agrees. That is the power rule doing its job on a falling stone, a circle, a sphere, a moving car and a braking distance. One caution on the last one: the braking formula is an idealised model that ignores the driver’s reaction time and assumes a steady friction value of 0.7, so treat the numbers as an illustration, not as a safety guide.

Common mistakes and how to avoid them

• Forgetting to subtract one. The derivative of x⁵ is 5x⁴, not 5x⁵. After you bring the exponent down, the power must come down too.
• Getting the sign wrong with negative exponents. The derivative of x⁻² is −2x⁻³. Subtracting one from −2 gives −3, not −1.
• Mixing up xⁿ and n^x. The power rule needs the variable in the base and a fixed number in the exponent. For 2^x, where the variable is in the exponent, the power rule does not apply. That is an exponential function with a different rule.
• Applying it to something bigger inside. (x² + 1)³ is not 3(x² + 1)². The inside also changes with x, and you need the chain rule to account for it. The power rule alone only works when the base is plain x.
• Treating a constant like a variable. The derivative of 5 is 0, not 5 · x⁻¹. And in a formula like 6x, the 6 is a multiplier, not something you differentiate.
• Not rewriting roots and fractions first. Convert √x, 1/x and 1/x³ into powers of x, and the rule applies directly.
• Dropping the multiplier on the “answer”. The derivative of 7x³ is 21x². You multiply the 7 by the 3, not discard it.

The power rule in reverse

If differentiating lowers the exponent, then undoing it should raise it. That is integration, the reverse process, and its power rule runs backwards: add one to the exponent, then divide by the new exponent.

Check it by differentiating the answer. The derivative of x⁴/4 is (4 · x³)/4 = x³, which is what we integrated. The 4 on top cancelled the 4 underneath, so the two rules are perfect mirror images of one another. The exception n = −1 would mean dividing by zero, and that case leads to the natural logarithm, a story for another article.

Practice problems with answers

Try each one on paper first, then tap to check yourself.

Test yourself: five quick questions

Tap a question to reveal the answer. The correct option is marked.

Frequently asked questions

The post The Power Rule Explained: Why the Exponent Drops Down appeared first on Learn With Examples.

A story I tell every new analyst

Years ago I worked with a team that monitored server performance. They had an alert that fired whenever response time crossed a threshold. In the first month it fired forty times, and thirty-nine were nothing. By the second month, people stopped reading the messages. In the third month the one alert that mattered arrived, and it sat unread for two hours while a real outage grew.

That team had made both classic mistakes, one after the other. First they set the alarm too jumpy, which produced false alarms. Then, by training everyone to ignore it, they made real signals get missed. If you understand why that happened, you already understand the core of this article.

Here is the promise: by the end, you will be able to name each error, explain the trade-off between them, use the words alpha, beta and power without flinching, and decide which error matters more in a given situation. No advanced maths required. When the numbers do appear, I have computed them so you can follow along.

The idea of a null hypothesis in plain words

Before we can talk about errors, we need one setup idea. Almost every statistical test begins with a boring default assumption called the null hypothesis, written H0. It says “nothing is going on.” The new drug does nothing. The new web page converts no better than the old one. The coin is fair. The defendant is innocent.

Then you look at evidence and ask a single question: is this evidence surprising enough, under the assumption that nothing is going on, that I should stop believing it? If yes, you reject the null. If not, you fail to reject it, which is different from proving it true. It just means you did not find enough evidence.

The four possible outcomes

Any test ends in one of four situations, depending on what is true in reality and what the test says. Two are correct decisions and two are errors. The grid below is worth memorising.

Look at how symmetrical the grid is, and how differently the errors feel. A false alarm is loud and visible: someone complains, someone investigates, someone wastes an afternoon. A missed signal is silent by nature. Nobody notices what did not happen. That imbalance in visibility is why people routinely underweight Type II errors.

The memory trick that actually works

Students confuse the two constantly, and honestly so did I in the beginning. The trick that finally fixed it for me is the boy who cried wolf.

• The first time, he shouts “Wolf!” and there is no wolf. That is a false alarm: Type I. Type one, the first mistake in the story.
• The second time, a real wolf shows up and the villagers ignore him. That is a missed signal: Type II. Type two, the second mistake.

The order in the fable matches the numbering. If you remember nothing else from this article, remember the wolf.

Alpha and beta: the two probabilities behind the errors

Each error has a probability with a Greek letter attached.

• α (alpha) is the probability of a Type I error. It is also called the significance level, and you choose it before you look at the data. The most common choice is 0.05, meaning you accept a 5% false-alarm rate when nothing is really happening.
• β (beta) is the probability of a Type II error, the chance you miss a real effect.
• Power = 1 − β is the probability you catch a real effect. Researchers often aim for at least 80%.

Think of alpha as how easily you are fooled by noise, and power as how well you can hear a real signal. A good study keeps the first low and the second high.

Seeing the trade-off with two overlapping curves

Here is the picture I draw on whiteboards. There are two bell curves. The left one shows what your test statistic looks like when nothing is going on. The right one shows what it looks like when there is a real effect. You choose a cut-off: any result to the right of the line triggers the alarm.

The red tail is the false-alarm zone. Even when nothing is going on, a small share of results land beyond the line, and with this cut-off that share is 5.0%. The blue tail is the missed-signal zone. When the effect is real, some results still fall short of the line, and here that is 19.6%.

Now imagine sliding the cut-off. Move it to the left, and you catch more real effects but raise more false alarms. Move it to the right, and false alarms shrink while you miss more real effects. You cannot push both down by sliding the line. This is the fundamental trade-off of hypothesis testing. Here are actual numbers for four cut-off positions on this same pair of curves.

Read down the columns. As the cut-off rises, the false alarms fall from a wild figure to a tiny one, but the missed signals climb steadily. There is no free lunch. The only question is which mistake you can better afford.

Which error is worse? It depends on the stakes

This is the part that separates textbook knowledge from professional judgment. There is no universal answer; the answer depends on what each mistake costs. Tap through five situations below.

Notice the pattern. Courts guard hard against Type I errors (convicting the innocent). Smoke alarms guard hard against Type II errors (missing a fire). Spam filters lean cautious about Type I errors (losing real mail). Screening programs accept lots of false alarms to avoid misses. In every case the threshold is a moral and practical choice wearing the costume of a number.

A closer look at medical screening

Medical tests give the clearest illustration of why false alarms are common even when a test is good. Suppose a condition affects 2% of a population. A screening test correctly flags 90% of people who have it (sensitivity) and correctly clears 95% of people who do not (specificity). Sounds excellent. Now screen 10,000 people.

• 200 people have the condition. The test catches 180 and misses 20 (Type II errors).
• 9,800 people are healthy. The test wrongly flags 490 of them (Type I errors) and correctly clears 9310.

That gives 670 positive results, of which only 180 are real. If you get a positive result, the chance you actually have the condition is about 27%. This surprises almost everyone. It does not mean the test is bad. It means the condition is rare, so even a small false-alarm rate produces a big pile of false alarms in absolute terms. This is why doctors follow up screening with a second, more specific test, and why a single positive result is a reason to investigate rather than a diagnosis.

How to reduce both errors: more data

If moving the cut-off only swaps one error for another, how do you get better on both? The answer is to sharpen the picture itself. A larger sample makes the two bell curves narrower, so they overlap less. With less overlap, you can hold false alarms at 5% and still catch far more real effects.

Here is a concrete illustration. Suppose a real effect exists that is about 0.3 standard deviations in size, which is modest. With a one-sided test at α = 5%, power grows with sample size like this.

A study with 10 subjects has a small chance of detecting this effect at all, and its “no significant difference” result would be nearly meaningless. It tells you the study was too small, not that the effect is absent. This is one of the most common misreadings of research: treating a failure to find something as proof that nothing is there.

Other ways to raise power include reducing noise in the measurements, using a more precise instrument, comparing matched pairs instead of independent groups, and looking for larger effects. Statisticians call planning this before a study “a power analysis,” and it is one of the cheapest ways to avoid wasting a research budget.

A worked example: the A/B test

Let me show you the numbers in a familiar business setting. You redesign your checkout page and compare it against the old one across 100 head-to-head comparisons. Under the null hypothesis, the new page is no better, so each comparison is like a fair coin flip and you would expect about 50 wins.

You decide in advance that you want no more than a 5% false-alarm rate. Working through the exact binomial probabilities, that means you declare the new page better only if it wins at least 59 of 100. Under a fair coin, the chance of hitting that bar by luck is 4.4%. That is your effective Type I error rate.

Now suppose the new page really is better and wins 60% of the time. How often would this test detect that? Only about 62% of the time. The other 38% of the time you would return “no significant difference” and shelve a page that really was better. That is a Type II error, and it is expensive because nobody ever finds out about the money that was left on the table.

The fix is not to loosen alpha. It is to collect more comparisons. If you had 400 comparisons instead of 100, the same 60% effect would be detected far more often while the false-alarm rate stays at 5%.

The multiple-testing trap: false alarms pile up

Here is a hazard that catches even experienced people. An alpha of 5% sounds low. But it applies to each test separately. If you run many tests on data where nothing is truly going on, the chance of at least one false alarm grows fast.

Run 20 tests and you have a 64% chance of at least one “significant” result by pure luck. Run 50 and it is 92%. This is the mathematical heart of “p-hacking”: slicing the data many ways until something looks significant. It is also why a marketing dashboard with forty metrics will always show a few “winners.” The remedy is to decide your hypotheses in advance, limit the number of comparisons, or adjust the threshold using methods such as the Bonferroni correction, which divides alpha by the number of tests.

What the p-value has to do with all this

The p-value is the probability of seeing data at least as extreme as yours, assuming the null hypothesis is true. You compare it with your pre-chosen alpha: if p is below alpha, you reject the null. The p-value itself is not the probability that the null is true, and it is not the probability that you made an error on this particular test. The false-alarm rate belongs to the procedure across many uses, and alpha sets it.

A quick way to hold both ideas: alpha is the standard you set before the race; the p-value is the time you actually ran.

Real-world places these errors hide

One more subtlety worth knowing: which error is called “Type I” depends on how you phrase the null hypothesis. If you flip the null and alternative, the labels flip. In practice, the null is the cautious default, and you should always state it out loud before you decide what each error means.

Six mistakes I keep seeing (tap to open)

Quick quiz: test yourself

Tap each question to reveal the answer and its reasoning.

Frequently asked questions

The post Type I vs Type II Errors: False Alarms and Missed Signals appeared first on Learn With Examples.

The problem with “the average”

I have reviewed reports for a long time, and the sentence that worries me most is: “the average delivery time is 27 minutes.” It sounds precise. But is every delivery close to 27, or do half arrive in 15 and the rest take 40? The average cannot tell you. It describes the centre of your data and says nothing about how spread out the values are.

Spread matters in daily life more than most people realise. A commute that averages 35 minutes but sometimes takes 75 makes you leave early. A salary band with an average of ₹67k that is really one founder and eight employees makes the average meaningless. A medicine that lowers blood pressure by 10 points on average but by 40 for some people and zero for others deserves a very different conversation.

So we describe data with two questions: where is the middle, and how far apart are the values? This article is about the second question, with the simplest measure first (range), the smarter measure next (IQR), and the quartiles that connect them.

Range: the quick, blunt measure

The range is the easiest statistic in the book. Sort the data, subtract the smallest from the largest, done. If eleven students score 52, 58, 61, 64, 67, 70, 72, 75, 78, 84 and 95, the range is 95 − 52 = 43 marks.

The range has real virtues. It is instant to compute, easy to explain to anyone, and good for sanity checks: a thermometer reading range of 200 degrees in one day tells you a sensor is broken. Many everyday questions are really range questions: “What is the cheapest and the most expensive flight?” “What are the coldest and hottest days this week?”

But the range has one serious flaw. It uses only two numbers, and they are the two most extreme ones. Everything in between is ignored. Change one value in the middle and the range does not budge. Change one extreme value and the range can explode. That makes it fragile.

Picture ten homes in a neighbourhood priced between ₹38 lakh and ₹64 lakh. Now a farmhouse worth ₹410 lakh sells at the edge of town. The range leaps from about 26 lakh to 372 lakh, even though nothing changed for the other nine families. That is exactly the situation where we need something sturdier.

Quartiles: cutting your data into four equal groups

You already know the idea from the median: line the data up in order and split it in half. Quartiles take that one step further and split the ordered data into four groups with (about) the same number of values in each.

• Q1 (first quartile, the 25th percentile): a quarter of the data sits at or below it.
• Q2 (second quartile): this is simply the median, with half the data on either side.
• Q3 (third quartile, the 75th percentile): three quarters of the data sits at or below it.

Here is a small example you can see rather than imagine. Twelve colleagues report their commute times in minutes. The dots are sorted, and colours mark the four groups of three.

Notice that quartiles are not fancy. They are just three cut-points that make four equal piles. Q1 is 28.5, the median is 36.5, and Q3 is 47.5 minutes. A quarter of people commute 28.5 minutes or less, half commute 36.5 or less, and three quarters commute 47.5 or less.

How to find quartiles by hand, step by step

Let me walk through the method most textbooks teach, using the eleven exam scores. This is the one I recommend learning first because you can do it on paper without any software.

Step 1: Sort the data

52, 58, 61, 64, 67, 70, 72, 75, 78, 84, 95. Always sort first. Skipping this step is the most common way to get a wrong answer.

Step 2: Find the median (Q2)

There are 11 values, so the middle one is the 6th: 70. Five values sit on each side.

Step 3: Find the median of the lower half (Q1)

The lower half is 52, 58, 61, 64, 67. Its middle value is 61. That is Q1.

Step 4: Find the median of the upper half (Q3)

The upper half is 72, 75, 78, 84, 95. Its middle value is 78. That is Q3.

Step 5: Subtract to get the IQR

IQR = Q3 − Q1 = 78 − 61 = 17. So the middle half of the class scored within a 17-mark window, even though the full range is 43.

What about an even number of values?

With 12 commute times, there is no single middle value, so the median is the average of the 6th and 7th values. Then you split the data cleanly into two halves of six and find the median of each. The lower half is 22, 25, 27, 30, 32, 35, and its median is (27 + 30) / 2 = 28.5. The upper half is 38, 41, 45, 50, 58, 75, and its median is (45 + 50) / 2 = 47.5. That is where Q1 = 28.5 and Q3 = 47.5 came from. No mystery.

Why your calculator and Excel sometimes disagree

Here is something nobody warns you about. If you compute quartiles in Excel and compare with the textbook, you may get a slightly different answer. That does not mean anyone made a mistake. There are several accepted methods for defining quartiles, and they agree on large data but can differ on small datasets. Here is the same exam data run through three common methods.

And for the twelve commute times:

The answers differ by a point or two. In real analysis with hundreds of records, the difference is negligible. In an exam, follow the method your teacher uses. In a report, mention which method your software uses, or just stay consistent. I once saw a two-hour argument between two analysts who were both correct, using different quartile definitions. Do not be those two.

The five-number summary

Once you have quartiles, you can describe any dataset with just five numbers: minimum, Q1, median, Q3 and maximum. It is called the five-number summary, and it is the backbone of the box plot.

Five numbers, and you already know where the centre is, how wide the middle half is, and how far the extremes reach. That is a lot of information in one row.

Same range, very different classes

Now the payoff. Here are two classes of ten students. Both have a lowest score of 45 and a highest of 95. Both have a range of 50. Any teacher looking only at the range would say the classes look identical.

Class X has an IQR of only 6. Most students scored between 62 and 68, with two students far away at the edges. Class Y has an IQR of 25, with scores spread evenly from 55 to 80 in the middle half. Same range, wildly different teaching challenges. In Class X you teach to the pack and support two outliers. In Class Y you need differentiated instruction across the whole room.

This is why I say the IQR is often the honest sibling of the range. It answers “how spread out are the typical values?” rather than “how far apart are the two weirdest values?”

Outliers and the 1.5 × IQR rule

The IQR does a second job that I use constantly: it gives you a fair way to flag unusual values. The statistician John Tukey proposed a simple rule that is now standard in box plots.

For the exam scores, the fences are 35.5 and 103.5, so nobody is unusual. Now look at the startup where nine people earn between 28 and 45 (thousand rupees a month) and the founder takes 320.

The upper fence is 62.25, so 320 is flagged. Two things are worth noticing. The mean pay is 67.2, higher than eight of the nine people, so the average paints a false picture. The median is 36, which is far more representative. And the IQR of 12.5 describes the spread among typical employees, uninfluenced by the founder. Whenever a dataset has one or two giant values, the median and the IQR should be your default pair.

A word of caution I give every junior analyst: an outlier flag is a prompt to investigate, not a permission slip to delete. It might be an error (someone typed 410 instead of 41), or it might be the most important data point you have. Look before you remove.

Range versus IQR on the same datasets

To see the difference at a glance, here are four datasets used in this article, each with its range (orange) and IQR (navy).

For exam scores and commute times the two measures are in the same neighbourhood. For house prices and startup pay, the range towers over the IQR because of one extreme value. That gap is itself a diagnostic: when the range is many times bigger than the IQR, you almost certainly have outliers or heavy skew.

Five real cases to explore

Tap a tab below. Each example uses actual numbers I ran through the method, with the sorted data shown so you can check my work by hand.

In every tab, the same routine applies: sort, find the median, find the halves’ medians, subtract, check the fences. Learn the routine once and you can use it on any list of numbers, from test marks to server response times.

Reading a box plot like a professional

A box plot squeezes the five-number summary into a picture. Once you know the anatomy, you can read one at a glance.

• The box spans Q1 to Q3, so its width is the IQR. A wide box means a wide spread among typical values.
• The line inside is the median. If it is off-centre in the box, the data is skewed.
• The whiskers reach the smallest and largest values that are still inside the fences.
• The dots beyond the whiskers are the flagged outliers.

Two extra reading tips. If the median sits close to Q1 and the upper whisker is long, the data is skewed to the right, as with incomes or house prices. And when you compare several box plots, look at the boxes first and the whiskers second. The boxes tell you about typical performance, and the whiskers tell you about extremes.

Where these measures show up in real life

In finance, the middle 50% of returns or prices is often more informative than the extremes. In education, admissions offices publish the middle 50% of test scores for admitted students. In sports analytics, a player’s IQR of scores describes consistency: a wide IQR means unpredictable, a narrow one means reliable.

Range, IQR or standard deviation?

You will sometimes be asked which measure of spread to use. Here is my rule of thumb after years of picking wrongly and correcting myself.

• Use the range for quick checks, small datasets and situations where the extremes themselves matter, such as safety limits.
• Use the IQR with the median when the data is skewed or contains outliers: incomes, prices, response times, hospital stays.
• Use the standard deviation with the mean when the data is roughly symmetric and you plan further statistical work.

None of these is universally best. Each answers a slightly different question, and reporting two of them together, such as the median and the IQR, usually gives a fair picture.

Seven mistakes I see all the time (tap to open)

Quick quiz: test yourself

Tap each question to reveal the answer and the working.

Frequently asked questions

The post Range, IQR and Quartiles Explained appeared first on Learn With Examples.

The puzzle, exactly as it is usually told

The French mathematician Joseph Bertrand published this puzzle in 1889, and more than a century later it still catches smart people out. I have used it in workshops for years, with analysts, teachers and engineers, and the pattern is remarkably stable: about four in five people answer 50% within seconds and defend it with real conviction.

Here is the setup. There are three identical boxes, each with two drawers. Inside them:

You choose a box at random, open one drawer at random, and see a gold coin. What is the probability that the other drawer in the same box also holds gold?

Before I explain, I want you to feel why 50/50 is so seductive, because if you understand the temptation you will spot the same mistake in a medical report or a court case.

The argument that feels airtight

Here is the reasoning almost everybody uses. “I drew a gold coin, so this cannot be the silver-silver box. That leaves two boxes: gold-gold and gold-silver. They were equally likely at the start, so it is a coin toss whether I am holding the gold-gold box or the mixed one. Half the time the other coin is gold.”

Every sentence in that paragraph sounds fine. The first step is correct: silver-silver is out. The second step is where it quietly goes wrong. The two remaining boxes were equally likely before you saw a coin. But you did not just learn that the box is not silver-silver. You learned something more specific: a gold coin came out. And the two boxes are not equally good at producing gold coins.

• The gold-gold box can only ever give you gold.
• The gold-silver box gives you gold only half the time.

So the observation of gold is stronger evidence for the box that always produces it. That is the whole secret. Evidence does not just eliminate possibilities; it reweights the ones that survive.

The clean way to solve it: count coins, not boxes

When I teach this, I tell people to stop thinking about boxes altogether. Boxes are a distraction. The thing that is randomly chosen, in effect, is one of six coins, each equally likely to be the coin you touch. Label them and list their situations.

Now apply the evidence. You saw gold, so throw away every case where you drew silver. Three outlined cases remain:

• Gold coin A from the gold-gold box. The other coin is gold.
• Gold coin B from the gold-gold box. The other coin is gold.
• The gold coin from the mixed box. The other coin is silver.

Three equally likely possibilities, and in two of them the other coin is gold. That gives 2/3. No formulas, no jargon, just counting the things that really are equally likely.

The reason I love this puzzle as a teaching tool is that the correct method is a habit rather than a trick: list the equally likely basic outcomes, cross out the ones that contradict what you saw, and count what is left. Once you adopt that habit, entire families of confusing probability questions become easy.

Do not trust me: run the experiment

Whenever a probability answer feels wrong, do not argue about it, simulate it. I wrote a short program that repeats the experiment: pick a box at random, open a random drawer, keep the run only if the coin is gold, and record whether the other coin was also gold. Here are the results, at increasing numbers of gold-first draws.

At 10 draws the number can bounce around, because small samples are noisy. At 1,000 draws it sits close to two thirds, and by 100,000 it is unmistakable. This is the law of large numbers doing what it does. If you want to convince a stubborn colleague, a simulation is more persuasive than any argument, because it removes the feeling that you are just trying to trick them with words.

You can even do a physical version at home. Put two gold and one silver stickers on a few index cards, or use red and white cards, and repeat it about fifty times. You will see about two thirds emerge, not exactly, but clearly.

The same idea, five different costumes

Bertrand’s puzzle is not really about coins. It is about how new information changes probabilities when the information arrives through a process that favours some cases over others. Once you have the pattern, you will start noticing it everywhere. Tap through the tabs below. Each one is a real, well-known version of the same reasoning, with the numbers worked out.

Notice how each tab has the same three moves: list the equally likely cases, remove the ones contradicted by the evidence, and count what remains. The medical and spam examples add a fourth idea, that rare things stay rare even after a positive signal, which brings us to Bayes.

The Bayes view: updating your beliefs

Statisticians phrase all of this in terms of Bayes’ theorem. Do not let the name scare you. It is a rule for updating a belief when you get new evidence. Start with what you believed before (the prior), ask how likely the evidence is under each possibility (the likelihood), and rescale.

For the boxes:

• Prior: each box is chosen with probability 1/3.
• Likelihood of drawing gold: gold-gold gives 1, gold-silver gives 1/2, silver-silver gives 0.
• Multiply: 1/3 × 1 = 1/3 for gold-gold, 1/3 × 1/2 = 1/6 for gold-silver, 0 for silver-silver.
• Rescale so the total is 1: gold-gold gets (1/3) / (1/2) = 2/3, gold-silver gets 1/3.

The two thirds falls right out. The gold-gold box started at one third and rose to two thirds because it was better at explaining the evidence. Same answer, different language. Some people find the coin-counting version more natural, others the Bayes version. I suggest learning both, because when the numbers get bigger, Bayes becomes the safer bookkeeping.

Where this goes wrong in real life

Medical screening

Consider a test for a condition that affects 1% of people. The test catches 90% of true cases and wrongly flags 9% of healthy people. You test positive. Intuition shouts that you are 90% likely to be sick. Let us count in a population of 10,000.

Of the 100 sick people, 90 test positive. Of the 9,900 healthy people, 891 also test positive. So among 981 positives, only 90 are truly sick: about 9.2%. The test is not bad. The condition is just rare, so false alarms outnumber true detections. Doctors in famous studies have made this exact error, which is why good clinics follow a positive screening with a confirmatory test rather than reacting to the first result.

Fraud alerts and spam filters

Banks send you a “suspicious transaction” text. Most of the time it is a false alarm, because fraud is rare among millions of transactions. This does not mean the fraud system is broken. It means its precision is limited by the base rate, and it is a deliberate trade: a few annoying alerts for a lot of caught fraud.

Evidence in court

Lawyers call the mistaken version the prosecutor’s fallacy: taking “the chance of this evidence if the person were innocent is one in a million” and hearing it as “the chance the person is innocent is one in a million.” Those are two different conditional probabilities. Confusing them has contributed to real miscarriages of justice, and it is the same logical slip as reading “gold came out of the gold-gold box” as if it were “the box is gold-gold.”

Hiring and screening

A company designs a screening test that 95% of great candidates pass. Then it assumes anyone who passes is 95% likely to be great. If only a small percentage of applicants are great, most passers are not. Again the rate at which the thing occurs in the population is the piece people forget.

Why our brains keep choosing 50/50

Psychologists have a few explanations, and I find them all helpful when I am teaching this.

• We count the visible options, not the weights. After the silver box is excluded, two boxes are visible, so we say half and half. The weights are invisible unless you deliberately look for them.
• We confuse the box with the coin. The question is about boxes in our heads, but the random selection was really about coins.
• We love symmetry. Two options often feel symmetrical even when they are not. The mind reaches for 50/50 when it is unsure.
• We ignore how evidence was generated. A gold coin is more likely to appear from a box with more gold in it. Whenever a signal is easier to get from one hypothesis than another, seeing the signal shifts the odds.

Once you know these four traps, you can build a checklist. It is what I use before I trust any probability I have just worked out in my head.

Variations that test your understanding

A good way to make sure you really understand a puzzle is to change it slightly and predict what happens. Try these.

What if you are told only that the box is not silver-silver?

Then the two remaining boxes really are equally likely, and the chance that the box is gold-gold is exactly 1/2. The difference between this and the original is that you did not see a coin. The coin observation is what carries the extra weight. This variation is a fantastic reminder that how you learned something can matter as much as what you learned.

What if a coin is chosen at random from the gold coins?

Suppose someone gathers all three gold coins and hands you one at random, then asks whether its box-mate is gold. You would get the same 2/3, because you are effectively picking among the same three cases.

What if there are four boxes?

Add a second gold-silver box. Now there are four gold coins in total, two in the gold-gold box and two in the mixed boxes. Draw gold, and the chance the other coin is gold is 2 out of 4, or 1/2. The count changes, so the answer changes. This shows the method is flexible: nothing magical about two thirds, it is simply the result of the count.

What if the coin-drawing is not random?

If somebody peeked and deliberately showed you a gold coin whenever they could, the mechanism changes and the probabilities can change again. This is the same subtlety that makes the Monty Hall problem sensitive to the host’s rules. Always ask how did this evidence reach me?

A short history and why the name “paradox” is fair

Strictly speaking it is not a contradiction; the mathematics is completely consistent. It is usually called a veridical paradox: the answer is true, but it clashes with a strong intuition. The name “paradox” was attached to the puzzle later, and it is not something I would claim Bertrand himself chose. What he did give us is a lovely argument for why 1/2 cannot be right, and it needs no counting at all.

The symmetry argument against 1/2

Choose a box. Before you look at anything, the chance that it holds two coins of the same kind is 2/3, because two of the three boxes are matched. Now point to a drawer. You will see gold or silver. If seeing gold moved the chance of a matching box down to 1/2, then by symmetry seeing silver would move it to 1/2 as well. But if it moves to 1/2 whichever coin you see, you did not need to look: the answer was already 1/2 before opening the drawer, which contradicts the 2/3 we started with. The probability cannot change by looking when every possible look changes it the same way. So it stays at 2/3, and the matching box is just as likely to contain gold as silver. That is the cleanest way I know to see why the tempting answer fails.

The same family includes the Monty Hall problem and the two-child problem. If you enjoy Bertrand’s boxes, those are the natural next puzzles. They share a structure: some information is revealed, and the trap is to treat the remaining cases as equally likely when they are not. One caution on the two-child family: small changes in wording can really change the answer. “At least one is a boy” gives 1/3 for two boys, while “at least one is a boy born on a Tuesday” gives 13/27. The day is not irrelevant, because it changes which families qualify, and the symmetry argument above does not carry over: a family can have boys born on several different days, so the possible Tuesday-style announcements overlap instead of splitting the families into clean groups. The answer also depends on how you learned the fact. Meeting a random child who turns out to be a boy born on a Tuesday leads back to 1/2 for the other child.

The Sleeping Beauty problem is a different kind of puzzle, and I would not file it here. It asks about a self-locating belief after memory erasure, and thoughtful people still disagree. “Thirders” say 1/3 and “halfers” say 1/2, depending on how they model what waking up tells her. There is no settled consensus, so treat anyone who calls it closed with some suspicion.

Five mistakes people make with conditional probability

Quick quiz: test yourself

Tap a question to reveal the answer and its reasoning.

Frequently asked questions

The post Bertrand’s Box Paradox: Why “It’s Obviously 50/50” Is Wrong appeared first on Learn With Examples.

Why I still start every probability class with a coin

I have spent a lot of years explaining probability to engineers, analysts, nurses, marketers and one very patient group of warehouse supervisors. Every time, I start with a coin, and every time somebody looks slightly insulted. A coin? Really? But a coin is the cleanest possible laboratory. Two outcomes, no hidden moving parts, a probability everyone already believes. Once the coin makes sense, you can swap the word “heads” for “defective charger” or “customer clicked” or “seed sprouted” and the maths does not change at all.

That swap is the whole story of the binomial distribution. It is the tool for counting successes when you repeat the same yes-or-no event a fixed number of times. It answers questions like these:

• If I flip a coin 10 times, what is the chance of exactly 4 heads?
• If 5% of chargers are faulty and I test 20, how likely is it that I find none?
• If a shooter makes 80% of free throws, how likely is a perfect night?
• If I email 200 people and 5% usually click, is 15 clicks luck or a real improvement?

By the end of this article you will be able to answer all four, by hand if you want, and you will know when the answer can be trusted and when it cannot. We will move from coin flips to quality control, then into free throws, exam guessing, email campaigns and airline overbooking. There is a small interactive panel, a few graphics, a quiz and an FAQ at the end.

Is my situation actually binomial? The four checks

Before you use any formula, check that the situation qualifies. Skipping this step is the number one source of bad statistics I have seen in reports over the years. The formula will happily give you a number even when the setup is wrong. A wrong setup just gives you a confident wrong number.

A quick habit that helps: say the four conditions out loud about your problem. “I have 20 chargers, each is defective or not, one charger does not affect another, and the defect rate is 5% for all of them.” If any sentence makes you hesitate, stop and think before calculating.

The coin flip: building the idea from scratch

Let us take the smallest interesting case. Flip a fair coin 4 times and ask for exactly 2 heads. Each flip is independent and heads has probability 0.5, so any particular sequence of four flips has probability 0.5 × 0.5 × 0.5 × 0.5 = 1/16.

Now the key question: how many different sequences contain exactly two heads? Here they are all.

There are six. Each has probability 1/16, and they cannot happen together, so we add them: 6 × 1/16 = 6/16, which is 37.50%. That is the entire logic of the binomial distribution. Count the ways, then multiply by the probability of each way.

Listing sequences works for four flips. For twenty flips you would need over a million lines. So mathematicians invented a shortcut for the counting part, and it has a friendly name: “n choose k.”

Counting the ways with Pascal’s triangle

The number of ways to choose k successes among n trials is written C(n, k). You can compute it with factorials, but there is a prettier way. Each number in Pascal’s triangle is the sum of the two numbers above it, and row n, position k gives C(n, k). The highlighted circle below is C(4, 2) = 6, our six coin sequences.

Notice how the numbers rise toward the middle of each row. There are far more ways to get a balanced result than an extreme one. Only one sequence gives 4 heads out of 4 (HHHH), but six give 2 heads out of 4. That simple fact is why the middle of the binomial chart is always the tallest for a fair coin.

The formula, one piece at a time

People stare at this and feel intimidated. Do not. It is three pieces you already understand:

• C(n, k) counts how many different orders produce exactly k successes.
• p^k is the probability that k particular trials all succeed.
• (1 − p)^{n − k} is the probability that all the remaining trials fail.

Worked example: exactly 5 heads in 10 fair flips

Here n = 10, k = 5, p = 0.5. Then C(10, 5) = 252. Each specific sequence has probability 0.5¹⁰ = 1/1024. So the probability is 252/1024, which is 24.6%. Not quite one in four. Most people guess it is closer to half, because “five is the average.” The average is five, but the exact value of five is only one of eleven possible outcomes competing for probability.

Look at the tails of that chart. Zero heads or ten heads each has a probability of 0.10%, about one in a thousand. Meanwhile, getting between 4 and 6 heads happens 65.6% of the time, and 8 or more heads happens 5.5% of the time. A run of 8 heads out of 10 is unusual, but it is not a miracle. I tell people that if they never see it once in a while, the coin is the strange one.

Quality control: the same maths on a factory floor

Now change the story. A company buys phone chargers from a supplier who says the defect rate is 5%. The receiving team pulls 20 chargers from a shipment and tests them. Is this binomial? Fixed n (20), two outcomes (defective or fine), independent units, and a constant defect rate. Yes, with the usual caveat that we treat a large shipment as if each pick is independent.

Here, “success” means “defective”. That trips people up. Success is just the thing we count. So n = 20, p = 0.05, and we can produce the whole table of outcomes.

Read that chart carefully because it corrects two common instincts. First, a defect-free sample happens only 35.8% of the time, even though the supplier really is at 5%. Finding zero defects in 20 does not prove the supplier is perfect. Second, finding one defect (37.7%) is just as likely as finding none. Three or more defects has probability 7.55%, so if that happens you have a real reason to call the supplier.

The “at least one” shortcut every analyst should know

Suppose your manager asks, “What is the chance we see at least one defective charger?” Do not add up 20 terms. Use the complement: P(at least one) = 1 − P(none). Here, that is 1 − 0.95²⁰ = 64.2%. Nearly two in three samples of 20 contain at least one bad unit, even at a healthy 5% rate. It is one of the most useful tricks in the whole subject, and it works for any “at least one” question.

The wrong shortcut goes: 20 chargers × 5% each = 100%, so we are sure to find one. Nope. Percentages of different events do not simply add up unless the events cannot overlap, and here they can.

Mean and standard deviation: what to expect and how far off you can be

The binomial distribution has two beautifully simple summary numbers.

For 10 fair coin flips, the mean is 5 and the standard deviation is √(10 × 0.5 × 0.5) = 1.58. For 20 chargers at 5% defect, the mean is 1 and the standard deviation is 0.97. So you expect about one defect, plus or minus one. That explains the chart above, where 0, 1 and 2 defects are all perfectly normal.

When I coach new analysts, I ask them to memorise this rule: the mean tells you what to expect, the standard deviation tells you how surprised to be. A result within about two standard deviations of the mean is ordinary luck. A result far beyond that deserves an investigation.

Five real situations, one formula

Below is a small panel. Tap a tab to switch situations. It runs on plain HTML and CSS, so it works anywhere the article does. In each case, I ran the exact numbers so you can see the formula do real work.

Notice the pattern. In every tab the mean tells a comforting story (8 baskets, 10 clicks, 10.8 sprouts), but the interesting decisions live in the details of the spread. The exact result is rarely the average result. That is not a flaw of the model. It is the model working as intended.

How the shape changes with p

A fair coin gives a symmetric, hill-shaped chart. But change p and the hill slides sideways. When p is small, successes are rare and the pile of probability sits near zero. When p is large, it sits near n. Only p = 0.5 gives perfect symmetry. The three charts below all use n = 10.

Here is a practical reading of those pictures. A left-leaning chart (p = 0.1) tells you that “zero” and “one” are the typical answers, and that seeing four or five is a real signal. A right-leaning chart (p = 0.9) is the mirror image. If you understand one, you understand the other by swapping the words “success” and “failure”.

Acceptance sampling: how factories really use it

Let me show you the most valuable use of this distribution in industry. Testing every unit is expensive, and sometimes destructive (you cannot crash-test every car). So companies test a sample and use a rule. A classic example: test 20 units; accept the lot if you find at most 1 defective, otherwise reject.

The natural question is how good this rule is. It depends on the true defect rate of the lot, which nobody knows. But the binomial lets us compute the acceptance probability for each possible defect rate.

This is a real piece of quality engineering, called an operating characteristic curve. Read it like a report card on your inspection rule. A 1% defective lot is accepted 98% of the time, which is good for the supplier. A 5% lot is accepted 74% of the time, which may be too lenient if 5% is unacceptable to you. A 10% lot is still accepted 39% of the time. If that bothers you, you do not change the maths, you change the plan: test more units, or accept only when zero defects appear.

What I love about this example is that it turns an argument (“is this sample big enough?”) into a number. Instead of saying “20 feels low,” you can say, “with 20 units we still let a 10% bad lot through about two times in five.” That sentence changes meetings.

Cumulative probability: “at most” and “at least”

Real questions are rarely about exactly k. They are about ranges. “At most 2 defects.” “At least 8 baskets.” “Between 40 and 60 heads.” The rule is simple: add the individual probabilities in the range.

• At most k: add P(0) up to P(k). This is the cumulative column in the charger table.
• At least k: use 1 minus P(at most k − 1). Careful with that minus one, it is where most slips happen.
• Between a and b: add P(a) through P(b), or subtract two cumulative values.

As an example, in 100 fair flips the chance of exactly 50 heads is only 8.0%, but the chance of landing anywhere from 40 to 60 heads is 96.5%. The exact value is stingy, while the range is generous. In real work you almost always want the range.

When n gets big: the normal approximation

Computing C(200, 15) by hand is unpleasant. Before computers, statisticians noticed something lovely: as n grows, the binomial chart looks more and more like the smooth bell curve, centred at np with width equal to the standard deviation. That gave them a shortcut. Treat the count as roughly normal with mean np and standard deviation √(np(1 − p)).

A common rule of thumb says the approximation is decent when both np and n(1 − p) are at least 10. Our email campaign (n = 200, p = 0.05) has np = 10, right at the edge, so it works but not perfectly. Today, software gives exact binomial answers instantly, so the approximation matters more for understanding than for calculation. Still, the ideas are the same: a big sample makes the outcome more predictable in proportion, even though the raw count can wobble more.

That last point deserves emphasis. With 10 flips, the share of heads can easily land at 30% or 70%. With 1,000 flips it rarely strays beyond 47% to 53%. Bigger samples do not make luck disappear. They make luck small relative to the total.

Six mistakes I keep seeing (tap to open)

Where else you will meet this distribution

Any time your data is “how many out of how many,” the binomial is probably lurking underneath.

Quick quiz: test yourself

Tap a question to reveal the answer with the reasoning behind it.

Frequently asked questions

The post Binomial Distribution Explained with Coin Flips and Quality Control appeared first on Learn With Examples.

The post What Is a Probability Distribution? appeared first on Learn With Examples.

The post Coefficient vs Constant: What’s the Difference? appeared first on Learn With Examples.

price    = 1200
quantity = 3
discount = 0.10   # 10% off
gst      = 0.18   # 18% tax

subtotal = price * quantity                # 3600
after    = subtotal * (1 - discount)       # 3240
total    = after * (1 + gst)            # 3823.2

let downloaded = 340;   // MB so far
let totalSize  = 850;   // MB in the file

let percent = (downloaded / totalSize) * 100;   // 40
let remaining = totalSize - downloaded;          // 510 MB

import math

results  = 47
per_page = 10

pages = math.ceil(results / per_page)   # 5

# which results appear on page 3?
start = (3 - 1) * per_page + 1         # 21
end   = min(3 * per_page, results)     # 30

def to_fahrenheit(c):
    return c * 9 / 5 + 32

def to_celsius(f):
    return (f - 32) * 5 / 9     # same equation, solved for c

to_fahrenheit(30)   # 86.0
to_celsius(86)      # 30.0

for row in range(1, 7):
    if row % 2 == 0:          # remainder 0 → even
        colour = "grey"
    else:
        colour = "white"

/* 3 cards + 2 gaps = full width. Solve for one card: */
.card {
  width: calc((100% - 2 * 20px) / 3);
}

x, y   = 1, 1        # starting position
vx, vy = 1, 0.5      # velocity: moves 1 right, 0.5 up per frame

for frame in range(8):
    x = x + vx         # x = x + 1
    y = y + vy         # y = y + 0.5

# algebra: f(x) = 2x + 3
def f(x):
    return 2 * x + 3

f(4)    # 11
f(10)   # 23

# real-world function: delivery fee = ₹40 base + ₹8 per km
def delivery_fee(km):
    return 40 + 8 * km

delivery_fee(5)   # 80

if cart_total >= 499:
    shipping = 0              # free shipping
else:
    shipping = 49

if len(password) < 8:
    print("Password too short")

if age >= 18 and has_id:
    print("Access granted")

The post How Algebra Is Used in Computer Programming appeared first on Learn With Examples.

The post How to Solve Systems of Equations by Substitution appeared first on Learn With Examples.