Quant Memo
Foundational

Comparing Huge Powers Without a Calculator

To decide which of two enormous exponentials is larger without computing either, take logarithms of both — logs turn an impossible multiplication into an easy comparison of two ordinary numbers.

Which is bigger, 23002^{300} or 32003^{200}? Both numbers have around a hundred digits — actually computing either one by hand is out of the question, and even a calculator will choke on the exact digits. You don't need the exact values, though, just which one is bigger, and logarithms let you answer that from two ordinary-sized numbers instead.

The idea: logs turn exponent comparisons into multiplication

A logarithm is strictly increasing, meaning if a>ba > b then loga>logb\log a > \log b, and the reverse holds too — so comparing 23002^{300} and 32003^{200} is exactly equivalent to comparing log(2300)\log(2^{300}) and log(3200)\log(3^{200}). Using the power rule log(xn)=nlogx\log(x^n) = n\log x:

log(2300)=300log2,log(3200)=200log3.\log(2^{300}) = 300 \log 2, \qquad \log(3^{200}) = 200 \log 3 .

In plain English: instead of comparing two astronomically large numbers directly, you compare two ordinary-sized numbers — an exponent times a small, memorizable log value — because the logarithm collapses exponentiation down to multiplication, and the comparison survives the transformation intact since log is increasing.

Worked example 1: 2^300 vs 3^200

Using log1020.301\log_{10} 2 \approx 0.301 and log1030.477\log_{10} 3 \approx 0.477 (worth memorizing): 300log2300×0.301=90.3300 \log 2 \approx 300 \times 0.301 = 90.3, and 200log3200×0.477=95.4200 \log 3 \approx 200 \times 0.477 = 95.4. Since 95.4>90.395.4 > 90.3, log(3200)>log(2300)\log(3^{200}) > \log(2^{300}), so 3200>23003^{200} > 2^{300} — decided from two quick multiplications instead of two hundred-digit numbers. As a sanity check, both logs also tell you roughly how many digits each number has: 1090.310^{90.3} has 91 digits, 1095.410^{95.4} has 96 digits, confirming 32003^{200} isn't just bigger but noticeably so.

Worked example 2: a ratio-based shortcut without a log table

Compare 5445^{44} and 4504^{50} without any log values memorized, using a common trick: rewrite both as powers of the same exponent when possible. Here 4444 and 5050 share a factor of 22: 544=(522)25^{44} = (5^{22})^2 and 450=(425)24^{50} = (4^{25})^2, so comparing 5445^{44} vs 4504^{50} reduces to comparing 5225^{22} vs 4254^{25} — smaller exponents, same conclusion since squaring preserves order for positive numbers. Repeating isn't always this clean, but when exponents share a large common factor, dividing it out first shrinks the problem before you even reach for logs. Failing that, logs always work: log(544)=44log544(0.699)=30.8\log(5^{44}) = 44\log5 \approx 44(0.699)=30.8 and log(450)=50log450(0.602)=30.1\log(4^{50})=50\log4\approx50(0.602)=30.1, so 544>4505^{44} > 4^{50}, a margin close enough that it's worth double-checking the arithmetic carefully.

Function explorer
-2222.8
x = 1.00f(x) = 0.000

The explorer above lets you drag the base and see how logb(x)\log_b(x) grows — the key fact this whole technique leans on is that it's strictly increasing, so it never reorders two numbers when you take logs of both, only rescales them to a size you can compare by hand.

To compare two large exponentials ama^m vs bnb^n, take logs: compare mlogam \log a vs nlogbn\log b, two ordinary multiplications instead of two unmanageable numbers. Memorizing log1020.301\log_{10}2\approx0.301, log1030.477\log_{10}3\approx0.477, log1070.845\log_{10}7\approx0.845 covers most interview cases via products of small primes.

Before reaching for logs, check whether the exponents share a common factor — dividing both down to smaller exponents with the same base structure (as with 5445^{44} vs 4504^{50} above) can make the comparison obvious without any logarithm at all.

Related concepts

Practice in interviews

Further reading

  • Engel, Problem-Solving Strategies, ch. 1
ShareTwitterLinkedIn