The xkcd-suggested passwords have 44 bits of entropy. Assuming a weak hash like SHA1, a single 4090 could crack such a password in under 10 minutes (source).
My 16 character password, with 70 symbols per character, has log₂70 * 16 ≈ 98 bits of entropy. That corresponds to a cracking time of over 200 billion years with the same parameters.
xkcd’s password system is quite terrible for security. Its only advantage is that it’s relatively secure for how easy it is to remember. If you’re someone who really struggles to remember passwords and would otherwise use something even weaker, go for it, but if you want security then random characters are the way to go.
And your opinion is exactly that and doesnt match security research:
For the following you’re not the target group but others reading this who might want to make their lifes easier. Just from your way of writing I at least don’t expect that minor sources like okta or the NCSC will change your mind.
( article links with high level descriptions and links to their primary sources)
I’m not arguing that random passwords are better for everyone, just that they’re most secure for their length. A 9 word passphrase is just as secure as a 16 character random password, but is far longer.
A 4 word xkcd passphrase is more or less equivalent to a 7 character random password, and is secure with xkcd’s threat model (online brute force attack) but not with other threat models, like a brute force of a weak hash, which is many orders of magnitude faster.
If you’d like to verify the math:
4 word xkcd passphrase: 2048 (possible words) ^ 4 (number of words) = 44 bits of entropy ≈ 17.6 trillion possibilities.
7 word password: 70 (possible characters) ^ 7 (number of characters) ≈ 42.9 bits of entropy ≈ 8.2 trillion possibilities.
(Adding an eighth character raises the number to 576 trillion).
The xkcd-suggested passwords have 44 bits of entropy. Assuming a weak hash like SHA1, a single 4090 could crack such a password in under 10 minutes (source).
My 16 character password, with 70 symbols per character, has log₂70 * 16 ≈ 98 bits of entropy. That corresponds to a cracking time of over 200 billion years with the same parameters.
xkcd’s password system is quite terrible for security. Its only advantage is that it’s relatively secure for how easy it is to remember. If you’re someone who really struggles to remember passwords and would otherwise use something even weaker, go for it, but if you want security then random characters are the way to go.
Take a sentence with 200 characters then.
And your opinion is exactly that and doesnt match security research:
For the following you’re not the target group but others reading this who might want to make their lifes easier. Just from your way of writing I at least don’t expect that minor sources like okta or the NCSC will change your mind.
( article links with high level descriptions and links to their primary sources)
https://www.okta.com/identity-101/password-vs-passphrase/
https://www.4bis.com/passphrase-vs-complicated-passwords-passphrases-are-best/
https://specopssoft.com/blog/passphrase-best-practice-guide/
I’m not arguing that random passwords are better for everyone, just that they’re most secure for their length. A 9 word passphrase is just as secure as a 16 character random password, but is far longer.
A 4 word xkcd passphrase is more or less equivalent to a 7 character random password, and is secure with xkcd’s threat model (online brute force attack) but not with other threat models, like a brute force of a weak hash, which is many orders of magnitude faster.
If you’d like to verify the math:
4 word xkcd passphrase: 2048 (possible words) ^ 4 (number of words) = 44 bits of entropy ≈ 17.6 trillion possibilities.
7 word password: 70 (possible characters) ^ 7 (number of characters) ≈ 42.9 bits of entropy ≈ 8.2 trillion possibilities.
(Adding an eighth character raises the number to 576 trillion).