Fermi Problems: How to Estimate Almost Anything

How many air molecules are in your room? How heavy is the whole atmosphere? Questions like these look impossible until you try them, and then they turn out to need only a few rough numbers and some multiplication. Physicists call them Fermi problems, and the skill of answering them “within an order of magnitude” is the idea behind our Enrico Fermi sticker.

What is a Fermi problem?

A Fermi problem, also called an order-of-magnitude problem, is an estimation question answered with justified guesses about the quantities involved, often with rough upper and lower bounds Fermi problem, Wikipedia^[1]. It is named after Enrico Fermi, who was known for making good approximate calculations with little or no actual data Fermi problem, Wikipedia^[1]. The goal is not three significant figures. It is the right power of ten, plus a clear view of which assumptions matter most. If the estimate and the real answer disagree badly, that is useful too: it tells you which assumption to question Fermi problem, Wikipedia^[1].

How did Fermi estimate the Trinity bomb with paper?

The most famous example comes from Fermi himself. At the first nuclear test, on 16 July 1945, he was at the Base Camp about ten miles from the explosion Fermi, Trinity account (1945)^[2]. About 40 seconds after the flash the blast wave arrived, and he dropped small pieces of paper from about six feet before, during and after it passed Fermi, Trinity account (1945)^[2]. In his own words:

“The shift was about 2½ meters, which, at the time, I estimated to correspond to the blast that would be produced by ten thousand tons of T.N.T.” Fermi, Trinity account (1945)^[2]

A 2021 re-analysis of trinitite, the glass formed from the melted desert sand, put the yield at 24.8 ± 2 kilotons of TNT Selby et al. 2021^[3] Trinity (nuclear test), Wikipedia^[4]. Fermi's on-the-spot number was low by a factor of about 2.5. That is exactly what an order-of-magnitude estimate promises: the right power of ten from a sheet of falling paper and a few minutes of thought.

Why do rough guesses work so well?

A Fermi estimate is usually a product of several factors. Multiplying numbers is the same as adding their logarithms:

log₁₀N = Σ_i log₁₀f_i

where N is the final estimate and f_i are the individual factors. If each guess is off by a random factor, too high in some steps and too low in others, the errors behave like a random walk on the logarithmic scale Fermi problem, Wikipedia^[1]. The spread of a random walk grows only as the square root of the number of steps:

σ_total = σ·√n

Here σ is the uncertainty of one factor, measured as a spread in log₁₀ (dimensionless), and n is the number of factors. Wikipedia's worked case is a nine-step estimate in which each step is off by a factor of 2. The total spread is then a factor of 2^√9 = 2³ = 8, so you expect to land within one-eighth to eight times the true value, far better than the worst case of 2⁹ = 512 Fermi problem, Wikipedia^[1]. Our simulation below draws many such nine-step chains and confirms that about two-thirds land within a factor of 8.

The catch is the word random. If every guess leans the same way, the errors add instead of cancelling, and the argument fails Fermi problem, Wikipedia^[1]. A good Fermi estimator tries to keep each step unbiased.

Worked example 1: how much does the atmosphere weigh?

Air pressure at the ground is the weight of the air above each square metre. So the total mass of air is roughly the sea-level pressure times Earth's surface area, divided by gravity:

M_atm ≈ 4πR_E²·p₀ / g

With R_E = 6.371 × 10⁶ m for Earth's mean radius, p₀ = 101,325 Pa and g = 9.81 m/s², our arithmetic gives M_atm ≈ 5.27 × 10¹⁸ kg. The measured total mean mass is 5.1480 × 10¹⁸ kg, about 2.5% less than the value inferred from sea-level pressure and Earth's area Atmosphere of Earth, Wikipedia^[5]; the difference is air displaced by Earth's mountainous terrain Atmosphere of Earth, Wikipedia^[5]. One line of reasoning lands within a few per cent.

Worked example 2: how many molecules are in a room?

Take a room 5 m × 4 m × 2.5 m, so a volume of 50 m³. For an ideal gas the number density is n = p/(k_BT), with p in pascals, T in kelvin and k_B = 1.380649 × 10⁻²³ J/K. At 101,325 Pa and 293 K this gives about 2.5 × 10²⁵ molecules per cubic metre, so the room holds roughly 10²⁷ molecules. The answer barely changes if the room is a little bigger or warmer; the only important inputs are pressure, temperature and the size of the room.

Worked example 3: the Fermi paradox

Fermi's name is also attached to a larger estimate. The Fermi paradox asks why, if technological civilizations are likely to exist in our galaxy, we have never encountered one, and the Drake equation is a well-known attempt to estimate their number as a product of factors Fermi problem, Wikipedia^[1]. Here the method shows its limits: several of the factors are so uncertain that the spread on the log scale is huge, and the result is an open question rather than an estimate. That puzzle is drawn on our Fermia: The Fermi Paradox tee.

Who was Enrico Fermi?

Fermi (1901–1954) won the 1938 Nobel Prize in Physics and led the team that built Chicago Pile-1, the first reactor to reach a self-sustaining chain reaction, on 2 December 1942 Enrico Fermi, Wikipedia^[6]. He developed the statistics now called Fermi–Dirac statistics, and particles that obey the exclusion principle are called fermions Enrico Fermi, Wikipedia^[6], the subject of why two electrons cannot share a state. His 1933–34 theory of beta decay introduced the name neutrino for Pauli's proposed particle Enrico Fermi, Wikipedia^[6]; read more in what neutrinos are.

Quick FAQ

What does “within an order of magnitude” mean?

Within a factor of about ten of the true value, the standard a Fermi estimate aims for Fermi problem, Wikipedia^[1].

Is a Fermi estimate just a guess?

No. Each factor is a justified guess, and the method is to make the reasoning explicit so the weak steps are visible Fermi problem, Wikipedia^[1].

How do I get better at them?

Practise breaking a question into independent factors, and check your answers against real data when you can. More physics gifts are in the Gifts for Physicists collection.

References

1. Wikipedia, “Fermi problem”. https://en.wikipedia.org/wiki/Fermi_problem
2. E. Fermi (1945), “My Observations During the Explosion at Trinity on July 16, 1945,” reproduced by AtomicArchive.com. https://www.atomicarchive.com/resources/documents/trinity/fermi.html
3. Hugh D. Selby et al. (2021), “A New Yield Assessment for the Trinity Nuclear Test, 75 Years Later,” Nuclear Technology 207, 321–325 (DOI 10.1080/00295450.2021.1932176; metadata verified via Crossref). https://doi.org/10.1080/00295450.2021.1932176
4. Wikipedia, “Trinity (nuclear test)”. https://en.wikipedia.org/wiki/Trinity_(nuclear_test)
5. Wikipedia, “Atmosphere of Earth”. https://en.wikipedia.org/wiki/Atmosphere_of_Earth
6. Wikipedia, “Enrico Fermi”. https://en.wikipedia.org/wiki/Enrico_Fermi

Written by Pixelated Physics. Every factual claim is linked to the source we checked; points that are still debated are labelled open.