"Six words" is the number everybody quotes — and it is correct for a 7,776-word list, where each word is worth 12.9 bits and six of them come to 77.
The lists here are a few hundred words, so each is worth about eight bits. Six of those is under 50 — weaker than an eight-character random password while sounding far stronger.
So the size of the list is printed, and the recommended word count is derived from it. A stated strength is only true about the list that produced it — borrowing the familiar advice while shipping a smaller list gives a passphrase that sounds strong and is not.
And capitalising adds no entropy at all. An attacker assumes the pattern, so nothing is multiplied.
Because “six words” assumes a 7,776-word list, where each word is worth 12.9 bits and six of them come to 77. The lists here are a few hundred words, so each is worth about eight bits — and six of those is under 50, which is weaker than an eight-character random password while sounding far stronger. The default is derived from the size of the list actually being used, so it always reaches the strength the page recommends.
Because reproducing 7,776 words accurately is not something this site can do honestly, and a wordlist with invented entries produces an entropy figure that is a guess. The list here is smaller and its size is printed on the page, because the strength quoted is only true about the list that produced it. The remedy for a smaller list is more words, and the page says how many.
A long one is, and the arithmetic is the whole argument. Each word multiplies the search space by the size of the list; each character of a password multiplies it by the size of the alphabet. Words are worth more per unit of memory — you can remember nine words far more easily than the twelve random characters that would match them, and the nine words are easier to type.
No, and the figure on this page correctly ignores it. Capitalising every word is a pattern an attacker assumes, so it does not multiply anything. It is a readability choice. The same is nearly true of a trailing digit: it adds 3.3 bits, which is less than half of one extra word, and it is offered only because some systems insist on a number.
Because a passphrase has to be remembered, and people remember words in the language they think in. Every entry is ordinary Filipino vocabulary — words you would recognise instantly and can picture. The list is slightly smaller than the English one, so the recommended word count is slightly higher, and the page works that out for you.
Usually, and it is the most readable option — the separator makes no difference to the strength, since an attacker knows what it is. A few systems still reject spaces or silently trim them, which is why hyphens and full stops are offered. If a site refuses a long passphrase at all, that is a sign it is storing passwords in a way you should be wary of.
No. It is generated in your browser from the operating system’s cryptographic random source, and nothing is transmitted, logged or stored. Nothing is kept when you close the page, so save it somewhere before you navigate away — we cannot recover it.
An eight-character random password is 53 bits. So six words from a small list loses to it.
Generated in your browser and never transmitted or stored. The wordlist here is smaller than the well-known EFF list, so the entropy per word is lower and the recommended word count is higher — both figures are printed rather than borrowed. Nothing you type here is sent to our servers — the calculation runs entirely in your browser.
