What Is a Passphrase? Password Strength by Entropy

PublishedToolPassphrase Generator

Contents
  1. Entropy: turning a count of possibilities into bits
  2. Short and complex: 8 characters, 94 character types
  3. Long but simple word combinations: the real variable is dictionary size
  4. How the tool actually guarantees randomness
  5. The bottom line
  6. Frequently asked questions

When you create a password, the advice you hear tends to split into two camps. One says to make it complex by mixing in special characters and numbers. The other says to string together several words that are easy to remember, even if the result is long. Which one is actually stronger isn't a matter of gut feeling — the math gives you the answer.

Password strength comes down to one question: how many possibilities does an attacker have to try before hitting the right one? Convert that count into bits and you have entropy. If you work out the entropy of a short, complex password and of a long word combination yourself, you'll see why a few randomly joined words hold their own on paper.

Summary ① Password strength is calculated by raising the number of choices per position to the power of the number of positions — that is, entropy in bits. ② An 8-character password built from 94 character types comes to about 52 bits in theory, while 8 English words picked at random from a 140-word list come out higher, at roughly 57 to 64 bits. ③ For this math to hold in practice, the choices have to come from real randomness, not from a person — and only when that condition is met does adding characters or words raise strength honestly.

Entropy: turning a count of possibilities into bits

A single password's entropy follows one rule. If each position offers N different choices and there are L positions in a row, the total number of possibilities is N multiplied by itself L times (N^L). That number gets too big to handle directly, so you take a logarithm and express it in bits. As a formula: entropy (bits) = L × log₂(N).

There are only two variables in this formula. N is the number of candidates per position, and L is the number of positions. Whether that position is a single character or a single word, the math works exactly the same way.

Short and complex: 8 characters, 94 character types

The "complex password" people usually recommend mixes 26 lowercase letters, 26 uppercase letters, 10 digits and special characters (about 32 of them on a standard keyboard) into 8 characters. That gives you 26+26+10+32 = 94 possible choices per position.

With 8 positions, the total number of possibilities is 94^8, which converts to 8 × log₂(94) ≈ 8 × 6.55 ≈ 52 bits. That's the theoretical ceiling for the short and complex side.

Long but simple word combinations: the real variable is dictionary size

Stringing random words together uses the same structure. The only difference is that the number of candidates per position (one word) is set by how many words the dictionary holds. The Passphrase Generator's English word list contains 140 words, so the entropy of a single word position is log₂(140) ≈ 7.13 bits. Interestingly, that's higher than the roughly 6.55 bits per character we calculated above. Measured purely by how much information one position carries, the word weighs more.

Actual strength, of course, is that value multiplied by the number of positions, so how many words you link matters. The tool lets you set the word count from 3 to 8 with a slider, and turning on the option to append a two-digit number (00–99) adds log₂(100) ≈ 6.64 bits on top. Here are the values by word count (English words, with the trailing-number option on).

WordsFormulaEntropy
33×7.13+6.64about 28 bits
4 (default)4×7.13+6.64about 35 bits
55×7.13+6.64about 42 bits
66×7.13+6.64about 49 bits
77×7.13+6.64about 57 bits
8 (max)8×7.13+6.64about 64 bits

At the default of 4 words you get 35 bits, which is actually lower than the 52-bit complex password we calculated earlier. But push the slider to 7 or 8 and you reach 57 to 64 bits, overtaking the 8-character complex password. Adding a few more random words has the same exponential effect as adding more characters.

Doing this calculation by hand every time means knowing the dictionary size, opening a logarithm calculator, and recalculating whenever you change an option. More importantly, the moment a person picks "this word, that word" in their head, the choice is no longer random.

Passphrase Generator

Adjust the word count, language, and separator, and it builds a passphrase with cryptographic randomness while showing the estimated entropy in bits right there.

How the tool actually guarantees randomness

The numbers above only mean something if the choices really were made at random. When a person picks words by hand, they drift toward words they like or words they know well, so the real number of possibilities ends up far smaller than the calculated value.

The Passphrase Generator pulls words and numbers each time using the browser's cryptographic random number generator, and it also removes the bias that comes from the modulo operation, so no word in the candidate list shows up more often than another. That's why the "estimated entropy" figure on screen holds exactly, with no gap between theory and the actual result.

Change the word count, the language (English/Korean), the separator or whether to add numbers, and the entropy value is recalculated and displayed right away, so you can watch the bit count while you adjust. For reference, switching the word language to Korean drops the candidate list to 79 words (log₂(79) ≈ 6.30 bits), so the same word count gives lower entropy than English. Uppercase display and separator choice, on the other hand, are just presentation options for readability — they don't add any possibilities.

The bottom line

  • Password strength is calculated by raising the number of choices per position to the power of the number of positions — that is, entropy in bits.
  • An 8-character password drawn from 94 character types comes to about 52 bits in theory.
  • This tool's English word list has 140 words, so one word carries about 7.13 bits — more than one character's roughly 6.55 bits.
  • The default of 4 words gets you only about 35 bits, but 7 or 8 words clears the 52-bit complex password.
  • For these figures to hold, the choices have to be truly random, and this tool meets that condition by drawing words with cryptographic randomness.
Passphrase Generator

Adjust the word count, language, and separator, and it builds a passphrase with cryptographic randomness while showing the estimated entropy in bits right there.

Frequently asked questions

What exactly is a passphrase?

A password made of several unrelated words picked at random and joined with a separator character. Instead of complex character combinations, it increases the number of possibilities through the number of words. This tool lets you adjust the word count, language, separator and whether to add trailing digits.

How do you calculate password entropy?

Decide how many choices each position offers (N), work out how many such positions there are (L), multiply them, then take the logarithm. The formula is "entropy (bits) = number of positions × log₂(choices per position)." For a character password, N is the number of character types you used; for a passphrase, N is the number of words in the dictionary.

Which is safer: a long word combination or a short, complex password?

You can't say either one is always stronger — you have to work out the number of positions and the choices per position. With this tool's 140 English words, 4 words calculate lower than an 8-character password from 94 character types, but 7 or 8 words flip the result. Keep in mind that this is the theoretical value assuming truly random selection.

Can I trust the "estimated entropy" number the tool shows?

The figure is the theoretical number of possibilities based on your current word count, language and trailing-number setting. Because the tool picks words and numbers with cryptographic randomness every time, the passphrase it actually generates has the same randomness as the calculation. That said, reusing the same passphrase across multiple sites creates risks this math says nothing about, so it's safer to use a different passphrase for each site.

This article was adapted from our Korean original. Read the Korean original

Related articles

  1. 8 vs 12 character password: 2.6 trillion times more combinationsSee how password length and character types change brute-force time, with a combination-count table and a 2.6 trillion gap between 8 and 12 characters.
  2. CLS Image Layout Shift: Why It Happens and How to Fix ItWhy images push content around as they load, and how setting width, height, or a same-size placeholder image keeps your layout from shifting.
  3. QR code grid pattern: why it looks that way and why it scansQR codes look like random black-and-white grids for a reason. See what finder patterns do and how error correction keeps a code scannable when covered or colored.
  4. QR Error Correction Levels L, M, Q, H: What Each One MeansL, M, Q, and H recover about 7%, 15%, 25%, and 30% of a QR code. See why higher levels add cells and how a damaged code still reads.
  5. Random Number Picker Algorithm: Math.random vs CSPRNGHow Math.random, CSPRNG, and Fisher-Yates shuffling affect draw fairness, plus how to pick winners without bias.
  6. 301 vs 302 Redirects: Why a 301 Is Hard to Undo301 is permanent, 302 is temporary. How search engines and browsers treat each, why a wrong 301 is hard to undo, and how to check with curl.

Related tools