← Garden

Password entropy, made visible

🌿 budding·Jul 8, 2026#security#passwords#entropy#lab

“Use a strong password” is good advice and useless advice at the same time. What does strong actually mean? Length helps, but Password1! is eleven characters and still terrible. Entropy helps, but raw bit counts do not tell you that correcthorsebatterystaple beats Tr0ub4dor&3 until something like zxcvbn explains the patterns.

The gap is visualization. Security teams talk in bits and crack times. Users pick pet names with a digit on the end. A good tool bridges that gap: show the math, show the patterns, and make the tradeoffs tangible.

Runs in your browser. Nothing you type is sent anywhere. Scoring and random generation use zxcvbn and Web Crypto, fully offline.

Waiting for input0 bits of entropy
Time to crack

Assumes offline attack at 10,000 guesses/sec (slow hash).

Use the analyzer to type or paste a password and watch the crack-time meter respond. Switch to the diceware tab when you need a passphrase instead of a memorized string of symbols. Source at lab-pw-entropy.

Entropy vs pattern matching

Approach What it measures Blind spots
Charset entropy Bits from length × character pool size Treats aaaaaaaa like random noise
zxcvbn Dictionary words, keyboard walks, repeats Heavier, English-centric
Diceware log₂(7776^words) for unbiased word picks Longer to type, very strong

Charset entropy answers “how big is the search space if every character were random?” zxcvbn answers “how hard is this specific string to guess in the real world?” Both numbers belong on the same screen.

When to reach for it

This is a teaching and sketching tool, not a breach checker. For that, pair it with a k-anonymity lookup like Have I Been Pwned when you are ready to ask whether a password has already leaked.