Crypto Passphrase Entropy: Why Randomness Beats Complexity
For years, password security advice trained people to think that a strong secret should look complicated:
Uppercase letters. Lowercase letters. Numbers. Symbols.
Take something ordinary, change a few letters, add a number and punctuation, and supposedly the result becomes dramatically harder to crack.
For cryptocurrency self-custody, that is an incomplete - and potentially dangerous—way to think about a BIP-39 passphrase.
The more important concept is entropy.
A passphrase is strong when an attacker cannot predict how it was generated. A string that looks complicated to a human does not necessarily contain much entropy if it was constructed using predictable human habits.
That distinction matters enormously when the passphrase may be the final secret protecting a cryptocurrency wallet after the recovery seed itself has been compromised.
Complexity Is Not the Same as Entropy
Consider these two hypothetical passphrases:
DOG25
and:
DOg-25'
The second certainly looks more complicated.
But suppose a person created it by taking the first phrase, changing one letter to lowercase and adding punctuation.
An attacker doesn't have to treat every possible seven-character string as equally likely.
Modern password-cracking strategies can prioritize dictionary words, capitalization changes, numbers, common substitutions and punctuation patterns.
This is one reason modern NIST guidance has moved away from requiring mixtures of uppercase characters, lowercase characters, digits and symbols. Research into breached passwords shows that people tend to satisfy those requirements in predictable ways.
Adding a symbol randomly can add entropy.
Adding an exclamation point because a person thinks passwords are supposed to end with an exclamation point is something entirely different.
Entropy Measures the Search Space
Entropy gives us a better way to think about the problem.
Suppose words are selected independently and uniformly from a 7,776-word Diceware-style list.
Because there are 7,776 possible outcomes for each word:
one random word ≈ 12.9 bits of entropy.
Six independently selected random words therefore provide approximately:
77.5 bits of entropy.
The arithmetic is straightforward:
Entropy = number of words × log₂(7,776)
And the growth is enormous.
| Random Words | Approx. Entropy | Possible Combinations |
|---|---|---|
| 4 | 51.7 bits | 3.66 quadrillion |
| 5 | 64.6 bits | 28.4 quintillion |
| 6 | 77.5 bits | 221 sextillion |
| 7 | 90.5 bits | 1.72 octillion |
| 8 | 103.4 bits | 13.4 nonillion |
A 7,776-word Diceware list contributes approximately 12.9 bits per independently random word, so six words provide about 77.5 bits.
And here is the crucial property:
Every additional random word multiplies the search space by another 7,776.
That is far more meaningful than simply asking whether a passphrase contains a capital letter or a symbol.
What About “Time to Hack”?
Crack-time estimates can help make enormous numbers understandable, but they need to be treated as illustrations rather than guarantees.
For example, if an attacker could somehow test 10 million candidates every second, and on average needed to search half of the available combinations:
| Random Words | Entropy | Illustrative Average Search Time* |
|---|---|---|
| 4 | 51.7 bits | ~5.8 years |
| 5 | 64.6 bits | ~45,000 years |
| 6 | 77.5 bits | ~350 million years |
| 7 | 90.5 bits | ~2.7 trillion years |
| 8 | 103.4 bits | ~21 quadrillion years |
*Illustrative calculation assuming 10 million candidate guesses per second and an average search of 50% of the total keyspace.
The exact time is not the important part.
Change the attack speed and the number of years changes.
The entropy does not.
That is why entropy and search-space size are better measurements than headlines claiming a password will take a particular number of years to crack.
BIP-39 also derives the wallet seed using PBKDF2-HMAC-SHA512 with 2,048 iterations, so generic password-cracking speed estimates should not simply be assumed to represent the exact performance of a real BIP-39 attack.
Random Means Random
There is one enormous condition behind all of those entropy numbers:
The words must actually be selected randomly and independently.
Seven words chosen because they make a clever sentence do not automatically provide 90.5 bits of entropy.
Neither does a favorite quotation.
Neither do children's names.
Neither does a sequence of personally meaningful words.
The entropy calculation assumes that each selection could genuinely have been any one of the 7,776 words with equal probability.
Physical dice are one way to produce that randomness.
A five-dice result has:
6 × 6 × 6 × 6 × 6 = 7,776 possible outcomes.
Each result maps to one word in the list.
Repeat the process for each word and the randomness comes from the dice—not from a human trying to behave randomly.
The Word List Does Not Need to Be Secret
This concept can feel counterintuitive.
An attacker may know the exact word list.
That does not defeat the method.
In fact, the entire list can be public.
Security comes from the attacker not knowing which independently random outcomes occurred.
If six words are randomly selected from 7,776 possibilities, knowing the vocabulary still leaves:
7,776⁶ = 221,073,919,720,733,357,899,776 possible sequences.
The secrecy belongs in the result, not the list.
Four Words Are Not the Same as Six
This is also why simply saying “use several random words” isn't enough.
Four Diceware words already produce a substantial 51.7 bits of entropy.
But moving from four words to six isn't a modest improvement.
Adding those two random words multiplies the search space by:
7,776 × 7,776 = 60,466,176 times.
Four words and six words may look relatively similar on paper.
Cryptographically, their search spaces are separated by more than 60 million times.
That is what entropy reveals.
Protect the Passphrase You Worked So Hard to Create
A high-entropy passphrase is only useful if it survives with you.
Once you have gone to the effort of creating a truly random, high-entropy passphrase, protect it with a durable offline backup. The Black Seed Ink steel passphrase backup solution is designed to preserve that critical secret against threats that paper and ordinary digital storage may not survive, including fire, water exposure, corrosion and long-term environmental damage.
Entropy protects the passphrase from being guessed. A resilient physical backup helps ensure that same passphrase is still available when you need to recover your wallet.
The passphrase backup should be kept separate from the wallet recovery seed so that finding one does not automatically give an attacker both components needed to reconstruct the passphrase-protected wallet.
The Research Lab's Updated View
Passphrase security should not be evaluated primarily by asking:
How many uppercase letters, numbers and symbols does it contain?
A better set of questions is:
How was the passphrase generated?
How much entropy does the generation process provide?
Could an attacker predict the choices a human made?
Can the passphrase be reproduced exactly during recovery?
Is there a durable offline backup?
For high-value cryptocurrency self-custody, randomness is more meaningful than appearance.
A complicated-looking secret may be predictable.
A sequence of ordinary-looking words selected with genuine randomness can represent an enormous search space.
That is the lesson entropy makes visible.
Research Lab Correction
An earlier version of this article presented brute-force estimates based on combinations of uppercase and lowercase characters, numbers and symbols. While that mathematics can describe strings uniformly generated at random from those character sets, it can overstate the security of human-created passphrases because human choices are not uniformly random.
The Black Seed Ink Research Lab has updated this article to focus instead on generation method, measurable entropy and total search space, which provide a more useful framework for evaluating cryptocurrency passphrase security.
Research Lab Disclaimer
This article is provided for educational and informational purposes only and does not constitute financial, investment, legal, tax or individualized cybersecurity advice. Entropy calculations assume independently and uniformly random selections from the stated 7,776-word list. Brute-force time estimates are illustrative and depend on assumed attack speed, implementation, hardware, key derivation and future technology. No passphrase, hardware wallet or security configuration eliminates all risk. Independently verify current security practices, maintain appropriate recovery backups and choose a self-custody strategy appropriate to your individual threat model. Do your own research (DYOR).
