Why Humans Are So Bad at Estimating Numbers (Explained)

How many ants are there in the world? How long is the Great Wall of China? How many cells does a human body have? For questions like these, most people are off by orders of magnitude. Usually too low. And it is not because they lack knowledge.

Our brain has a problem with very large and very small numbers. Evolution simply did not train us for them. Researchers know the exact mechanisms behind why we get certain numbers systematically wrong. In this article we explain the six most important thinking errors and how to avoid them.

The core problem

Our brain does not perceive numbers linearly, but logarithmically. The gap between 1 and 10 feels the same as the gap between 10 and 100. And that single error sits behind almost every bad estimate.

1. The Anchoring Effect: the first number sticks

Ask someone whether Mount Everest is higher or lower than 3,000 meters. Then ask how tall it is. The answer is almost always way too low. Ask the same person whether it is higher or lower than 15,000 meters. Now the estimate will be too high.

This is called the anchoring effect. Psychologists Daniel Kahneman and Amos Tversky discovered it in the 1970s. Even completely random numbers mentioned shortly before an estimate shift the result. In one famous experiment they spun a roulette wheel, then asked participants how many African countries are in the UN. Those who saw a high number on the wheel guessed higher. Even though everyone knew the wheel had nothing to do with Africa.

Anchors are everywhere: in price tags, in negotiations, in headlines. If you need an honest number, take a minute before you let the first impression in.

2. Logarithmic perception: why a million feels almost like a billion

Ask someone how long a million seconds is. Most guess a few weeks. The correct answer is 11.5 days. Ask the same person how long a billion seconds is. Most will double or triple their answer. The correct answer is 31.7 years.

A billion is a thousand times more than a million, not twice as much. The problem: both words sound similar, both are unimaginably large. Our brain treats them like neighboring floors when they are really on different planets.

A billion is huge

A million seconds ago was 11 days ago. A billion seconds ago the Romans were alive, and the Pyramids of Giza were already 1,000 years old.

3. Benford's Law: why the digit 1 shows up so often

Look at any statistic: populations, stock prices, river lengths. Which digit appears most often at the front? Most people guess 5 or 9, because it feels fair. The correct answer is 1.

In about 30 percent of real-world number sets, the value starts with the digit 1. Only 17 percent start with 2. And just 4.6 percent start with 9. This is not a coincidence, it is a mathematical law. It works for almost every natural set of numbers that spans multiple orders of magnitude. Tax auditors use Benford's Law today to spot manipulated filings. People who invent numbers tend to use the digit 1 far too rarely.

4. The exponential trap: why growth explodes

When something doubles every day, we massively underestimate the end result. Classic example: a patch of water lilies doubles every day and covers the whole pond in 48 days. On which day is the pond half covered? Most people guess day 24. The correct answer is day 47.

Exponential growth runs almost invisibly until just before the end. We think linearly, so we estimate the final value way too low. The same problem shows up in compound interest, in viruses, in computer chips. Once you really grasp that doubling ten times in a row equals a factor of 1,024, you start thinking about the future differently.

5. The availability heuristic: what we see often, we assume is common

Are plane crashes or car accidents more dangerous? Intuitively, crashes feel worse because they make headlines. In reality, over one million people die in cars every year, while only 300 to 500 die in scheduled flights.

Our brain uses a shortcut when estimating: how easily can I recall an example? The more present an event is in media or memory, the more common we think it is. That is why we overestimate shark attacks, terror attacks, and plane crashes. And why we underestimate quiet risks like high blood pressure or diabetes.

6. The scale illusion: very big and very small numbers

How many liters of water fit in the Pacific Ocean? How many atoms are in a single grain of sand? For questions like these, almost everyone guesses too low. The reason is always the same: our brain cannot picture anything bigger than a trillion or smaller than a millionth.

Evolution optimized us for numbers up to about a hundred. Up to that point we can compare, estimate, and calculate well. Everything beyond becomes abstract. Astronomers, chemists, and financial mathematicians work with such extremes every day. Their trick: they work in powers of ten, not absolute numbers. 10 to the 6 instead of one million. That way the brain can keep track.

What you can do to estimate better

Anchor yourself on purpose

Before you estimate, think of a known reference value. How tall is the Empire State Building (443 meters)? How many people live in Germany (84 million)? Starting from a real anchor, you can reason your way to a better guess.

Think in orders of magnitude

Instead of guessing exact numbers, first guess the correct power of ten. Is it thousands, millions, or billions? If you get the order of magnitude right, you already have 90 percent of the answer.

Always ask two questions

For important numbers, ask first: what would be impossibly low? And what would be impossibly high? Your guess sits somewhere in between. This is called a Fermi estimate, and it is a favorite tool of good physicists and managers.

Frequently asked questions about estimating numbers

Why do we almost always estimate large numbers too low?▾

Because our brain perceives numbers logarithmically, not linearly. The jump from million to billion feels much smaller than it really is. A thousand times more ends up feeling like just twice as much.

What is the anchoring effect?▾

A random number seen shortly before an estimate biases the result. Daniel Kahneman and Amos Tversky discovered it in the 1970s. It still works even when you know the anchor has nothing to do with the question.

How does Benford's Law help?▾

It shows that in natural number sets the digit 1 leads about 30 percent of the time, while 9 leads only 4.6 percent. People who make up numbers tend to distribute them evenly, and Benford's Law exposes them. Auditors and statisticians use it routinely.

Why do we underestimate exponential growth?▾

Because we think linearly by default. When something doubles every day, it looks slow 90 percent of the time and only explodes at the end. That is why pandemics, compound interest, and computer chips keep surprising us.

What is the availability heuristic?▾

A mental shortcut: the easier an example comes to mind, the more common we assume an event is. That is why we overestimate events that dominate the news and underestimate quiet risks like high blood pressure.

How do I estimate big numbers better?▾

Start with orders of magnitude instead of exact numbers. Use known anchor values (like Germany's population of 84 million). Then ask: what would be impossibly low, and what would be impossibly high? The truth sits between them.

What is a Fermi estimate?▾

A method physicists use to roughly answer big questions. They break the problem into small parts, estimate each part, and multiply at the end. Named after Enrico Fermi, who used it to estimate the strength of the first atomic bomb with scraps of paper.

Why We Systematically Misjudge Numbers