Powers of 10 — The Language of Fermi Estimation
How powers of 10 serve as the skeleton of back-of-envelope reasoning — reading magnitudes, combining estimates, and knowing when precision is beside the point.
The Core Idea
A power of 10 is an order of magnitude. When you say something is “in the millions,” you’re saying it lives near 10⁶. The question is never the exact number — it’s which rung of the ladder it sits on.
The ladder:
| Power | Value | Label |
|---|---|---|
| 10⁰ | 1 | — |
| 10¹ | 10 | tens |
| 10² | 100 | hundreds |
| 10³ | 1,000 | thousands |
| 10⁶ | 1,000,000 | millions |
| 10⁹ | 1,000,000,000 | billions |
| 10¹² | 1,000,000,000,000 | trillions |
Each step up is a 10× increase. Two steps is 100×. Three steps is 1000×. The difference between millions and billions is not “a lot more” — it’s a factor of 1,000.
Reading Landmark Numbers in Powers of 10
The landmark number table becomes cleaner when you strip it to magnitudes:
| Quantity | Rough value | Order |
|---|---|---|
| World population | 8 × 10⁹ | 10¹⁰ |
| India population | 1.4 × 10⁹ | 10⁹ |
| Large city | 10⁷ | 10⁷ |
| Seconds in a year | 3 × 10⁷ | 10⁷ |
| Earth radius (m) | 6.4 × 10⁶ m | 10⁶ |
| Distance to Moon (m) | 4 × 10⁸ m | 10⁸ |
| Human height (m) | 1.7 | 10⁰ |
| Human mass (kg) | 70 | 10² |
| World GDP (USD) | 10¹⁴ | 10¹⁴ |
Once the magnitudes are internalised, arithmetic between them is just adding exponents.
Multiplying and Dividing in Powers of 10
The rule: multiplication → add exponents. Division → subtract exponents.
10³ × 10⁵ = 10⁸
10⁹ / 10⁶ = 10³
In practice: strip each number to its magnitude, operate on the exponents, reassemble.
Example: How many litres of blood does India pump per second?
- India population: ~10⁹
- Heart rate: ~10¹ beats/min = ~10⁻¹ beats/sec (roughly 1 beat/sec, so 10⁰)
- Blood per beat: ~70 mL = 7 × 10⁻² L
10⁹ × 10⁰ × 7×10⁻² = 7 × 10⁷ litres/sec
That’s 70 million litres per second. The exact figure matters less than knowing it’s “tens of millions” — not hundreds, not thousands.
The Geometric Mean for Range Estimates
When you’re unsure if something is 100 or 10,000, don’t average them arithmetically (5,050 — absurd). Use the geometric mean: the midpoint on the log scale.
√(100 × 10,000) = √(10⁶) = 10³ = 1,000
The geometric mean of 10² and 10⁴ is 10³. When your uncertainty spans orders of magnitude, this is the right “middle.”
Combining Estimates: The Chain Rule
Fermi problems are chains of magnitudes. Each step introduces uncertainty; the errors compound — but so do the cancellations. A rough rule: if each factor is uncertain by half an order of magnitude (±3×), a 5-factor chain is uncertain by about 1.5 orders of magnitude. That’s still enormously useful — it rules out answers that are wrong by 100×.
Example: How many piano tuners are in Mumbai?
- Population: 2 × 10⁷
- Fraction with pianos: ~1 in 1,000 = 10⁻³
- Pianos: 2 × 10⁴
- Tunings per piano per year: ~1 → 2 × 10⁴ tunings/year
- Tunings per tuner per day: ~4, working days: ~250/year → 10³ tunings/year per tuner
- Tuners: 2 × 10⁴ / 10³ = ~20
Each step is a rough magnitude. The answer lands at 10¹ — tens of tuners. That’s the useful result. Whether it’s 12 or 35 is refinement; whether it’s 2 or 200 would reveal a broken assumption worth fixing.
Sanity Checks with Powers of 10
The fastest use of this framework: given an answer, does the magnitude make sense?
- Cost of a new smartphone: ~10⁴ rupees. An answer of 10² (₹100) or 10⁶ (₹10 lakh) fails immediately.
- Distance from Mumbai to Delhi: ~1,400 km = ~10³ km. If your map calculation gives 10⁵ km, you’ve made an error.
- Human heartbeats in a lifetime: ~10⁰ beats/sec × 3×10⁷ sec/year × 7×10¹ years ≈ 2 × 10⁹. Two billion beats — about right.
The magnitude check is faster than the full calculation and catches most errors.
Why It Works
The real world spans 40+ orders of magnitude — from the Planck length (10⁻³⁵ m) to the observable universe (10²⁶ m). Our intuition is calibrated for the narrow band we live in — meters, kilograms, hours. Powers of 10 extend that intuition outward, giving a common scale to compare things that don’t naturally sit next to each other.
The habit to build: before calculating, write down the magnitude you expect. After calculating, compare. A mismatch means either an error in the arithmetic or a wrong assumption — both worth knowing.