Part A · the single most important distinction — % vs percentage points
See the difference — drag the slider and watch both outputs
40%
+10
If the change is in PERCENTAGE POINTS
If the change is in PERCENT (%)
Rule: "Percentage points" = raw arithmetic subtraction (40% → 50% = +10 pp). "Percent change" = relative to the original (40% → 44% = +10%). Politicians and media almost always use whichever sounds bigger — which is usually % when rates are low, and pp when rates are high.
Part B · the 5 tricks used to mislead with percentages
Trick 1 · the moving baselinevery common
What they say
"Prices fell 20% then rose 25% — we're back to normal!"
What's actually true
You are NOT back to normal.
Start at 100. −20% = 80. Then +25% of 80 = +20 = 100? No: 25% of 80 = 20. 80 + 20 = 100. Wait — that works here. Let's try: −20% then +20%: 100 → 80 → 96. Still 4% below start.
The rule: Percentage changes are not reversible with the same number. A 50% drop needs a 100% rise to recover. A 20% drop needs a 25% rise. The base changes each time.
Trick 2 · cherry-picking the timeframevery common
Politician A says
"Crime is down 15% since we took office 2 years ago."
True — but starts from a 20-year high.
Politician B says
"Crime is up 40% over the last decade."
Also true — measured over a longer window that includes the spike.
The rule: Always ask "compared to when?" A trend can go up or down depending entirely on the start point chosen. Ask for the longest available timeframe.
Trick 3 · the average that hides everythingeverywhere
Mean (average)
5 people earn: €10k, €20k, €30k, €40k, €400k → mean = €100k
The mean is €100k, but 4 out of 5 people earn below it. One outlier destroyed the average.
Median (middle value)
Same 5 people → median = €30k
The median is far more representative of "typical." For income, house prices, and wealth, always prefer the median.
The rule: When the topic involves wealth, income, house prices, or anything with outliers — ask for the median. The mean is almost always higher and almost always misleading. "Average salary" is typically 20–40% higher than "typical salary."
Trick 4 · percent of what?sneaky
What they say
"We increased the education budget by 5%."
What you need to know
5% of the education budget — or 5% of GDP? Or 5% of last year's budget? Each gives wildly different numbers.
5% of a €1B budget = €50M. 5% of GDP (say €1T) = €50B. Same number, 1,000× different reality.
The rule: A percentage is meaningless without knowing its base. Always ask "X% of what total?"
Trick 5 · small sample sizesin headlines daily
Headline says
"100% increase in rare disease cases in our town!"
Reality
Last year: 1 case. This year: 2 cases. That is a 100% increase — from 1 to 2.
With tiny counts, random variation produces huge percentage swings that mean nothing.
The rule: Large percentage changes on small absolute numbers are almost always noise. Ask for the raw counts first.
Part C · interactive percentage calculator — do it in your head
The mental maths toolkit — pick a scenario
%
Part D · ratios in the real world — what they actually mean
P/E ratio (stocks)
Price ÷ Annual earnings
A P/E of 20 means you pay €20 for every €1 of annual profit. S&P 500 average ~15–25. A P/E of 100+ means investors are betting heavily on future growth (like Tesla, early Amazon).
Debt-to-income ratio
Monthly debt ÷ Monthly income
Banks want this below 43% for mortgage approval. If you earn €3,000/month and owe €1,000/month in debts, your ratio is 33% — acceptable. Above 50% = financial stress risk.
Body Mass Index (BMI)
Weight (kg) ÷ Height (m)²
18.5–25 = normal. 25–30 = overweight. 30+ = obese. A 75 kg person who is 1.75m tall: 75 ÷ 3.0625 = BMI 24.5 (normal). Crude measure — ignores muscle mass.
Gini coefficient
0 = perfectly equal, 1 = one person owns everything
Nordic countries: ~0.25–0.28. USA: ~0.41. Brazil: ~0.53. The single most widely used measure of income inequality in a country.
Part E · the compound growth trap — why % growth adds up fast
The rule of 72: how long to double at any growth rate?
7%/yr
Doubles in
×10 in
€1,000 → in 30 yrs
Rule of 72: Divide 72 by the growth rate to get approximate doubling time. 72 ÷ 7% ≈ 10.3 years. 72 ÷ 3% = 24 years. 72 ÷ 12% = 6 years. Works for any compound growth — inflation, investment returns, population, debt.
Part F · reading percentages in the news — a checklist
Every time you see a % in a headline, ask these 4 questions:
1
X% of what?
What is the base? Is it the total, a subset, last year's value?
2
Percentage or percentage points?
A change "from 5% to 8%" is either +3 pp or +60% — both technically correct, wildly different in feel.
3
Compared to when?
What is the time window? Who chose it, and why?
4
What are the raw numbers?
A 200% increase from 1 to 3 is very different from a 200% increase from 1,000 to 3,000.
Part G · test yourself
1. Interest rates rise "from 2% to 3%." A newspaper says this is a "50% increase in interest rates." A politician says it's "just a 1 percentage point rise." Who is right?
Both are technically correct — which is exactly the problem. In percentage terms: 3 is 50% more than 2, so "50% increase" is mathematically valid. In percentage point terms: the rate moved from 2 to 3, which is +1 percentage point. For borrowers with mortgages, the politician's framing feels smaller but the financial impact is substantial: a 1 pp rise on a €200,000 mortgage means ~€2,000 more per year in interest. The absolute cash impact is what matters for personal finance decisions — always convert to actual amounts.
2. A share price falls 50% in a crash. How much must it rise to get back to the original price?
100%. Start at €100. A 50% fall → €50. To go from €50 back to €100, you need a €50 gain — which is 100% of €50. This is the moving baseline trap: percentage losses and gains are not symmetric. It takes a 100% gain to recover from a 50% loss. A 75% loss needs a 300% gain to recover. A 90% loss needs a 900% gain. This is why investors obsess about avoiding large drawdowns — it is mathematically much harder to recover than to fall.
3. A company reports "average employee salary: €85,000." A union says "typical worker earns €45,000." Can both be true?
Yes, absolutely — and this is extremely common. If a company has 98 workers earning €45,000 and 2 executives earning €2.35 million each, the mean is: (98×45,000 + 2×2,350,000) ÷ 100 = (4,410,000 + 4,700,000) ÷ 100 = €91,100. The median is €45,000. Both numbers are honest. The company uses the mean because it's higher; the union uses the median because it represents the typical worker. For any distribution with high-end outliers, always prefer the median.
4. GDP grows at 2.5% per year. Using the rule of 72, approximately how long until the economy is twice as large?
About 29 years. 72 ÷ 2.5 = 28.8 years. The exact answer (using compound interest formula) is 28.1 years — the rule of 72 is remarkably accurate. At 3% growth: doubles in 24 years. At 1% growth: doubles in 72 years. China's ~6% average growth over recent decades: doubles in 12 years — which explains why its economy grew from roughly $500B in 1990 to over $17 trillion today (roughly 34× in 35 years, consistent with doubling ~2.8 times).
5. A product is "30% off" and then "an additional 20% off at the register." Is the total discount 50%?
No — it's 44% off, not 50%. The moving baseline again. Start at €100. First 30% off → €70. Then 20% off €70 = €14 off → €56. Total saving: €44 on a €100 item = 44% off. Retailers use this trick deliberately: "30% + 20% off" sounds like 50% but delivers 44%. The general rule: to combine percentage discounts, multiply the "remaining" fractions. (1 − 0.30) × (1 − 0.20) = 0.70 × 0.80 = 0.56, meaning you pay 56% = a 44% total discount.