Slashdot Mirror
Why People Are So Bad At Picking Passwords
mrspoonsi writes "Studies suggest red-haired women tend to choose the best passwords and men with bushy beards or unkempt hair, the worst. These studies also reveal that when it comes to passwords, women prefer length and men diversity. On the internet, the most popular colour is blue, at least when it comes to passwords. If you are wondering why, it is largely because so many popular websites and services (Facebook, Twitter and Google to name but three) use the colour in their logo. That has a subtle impact on the choices people make when signing up and picking a word or phrase to form a supposedly super-secret password. The number one conclusion from looking at that data — people are lousy at picking good passwords. 'You have to remember we are all human and we all make mistakes,' says Mr Thorsheim. In this sense, he says, a good password would be a phrase or combination of characters that has little or no connection to the person picking it. All too often, Mr Thorsheim adds, people use words or numbers intimately linked to them. They use birthdays, wedding days, the names of siblings or children or pets. They use their house number, street name or pick on a favourite pop star. This bias is most noticeable when it comes to the numbers people pick when told to choose a four digit pin. Analysis of their choices suggests that people drift towards a small subset of the 10,000 available. In some cases, up to 80% of choices come from just 100 different numbers."
These studies also reveal that when it comes to passwords, women prefer length and men diversity.
We are still talking about passwords, right?
"I believe in Karma. That means I can do bad things to people all day long and I assume they deserve it." : Dogbert
"women prefer length and men diversity"
Fnarr fnarr.
[quote]These studies also reveal that when it comes to passwords, women prefer length and men diversity.[/quote]
Fuck. We'll never win!
Actually it's not a good password if you can't remember it.
So from this article I take it I'm supposed to track down aredhead and have her make my password for me?
She looks like she has a trustworthy face.
Nice cut and paste arsehole. I'll forward a link to Mark Ward of the BBC Technology unit.
http://xkcd.com/936/
So, before choosing an important password make sure you have shaved, had a haircut and dyed your hair red.
(A sex change is asking too much though.)
http://xkcd.com/936/
who where what when now?
people are too lazy/stupid to remember a simple word or phrase critical for logging into essential accounts e.g. your bank, your email, or just your PC. My father had to write down his PC password. It was his dog's name. How can you not remember that?
Also are the most passionate lovers.
Hope is the currency of fools
In other news: The sky is blue, bears shit in the woods, fish swim in water, and this story is a repost from 1995.
... for RMS !
Time for bed, said Zebedee - boing
One of my relatives passwords was their pets name, ie Chloe, Phoebe. I asked "don't you think that is easy to guess?" and they said "No, how would anyone know that was MY password?".
What people don't realise is that hackers arn't usually attacking them specifically but are attacking everyone, anyone with a lame password. I'm pretty sure simple names are high on the list of things to try first.
Basically, bad passwords are a lack of education in how their password is vulnerable, or are just lazy.
A modern day password cracker (brute force) with a reasonably large dictionary can basically break all human generated paswords these days.
First - besides the dictionary, they also try variations - including l33t 5p34k variations, various capitalizations and putting numbers at the beginning or end of the word.
Second, the old trick of picking a phrase and using it? Also done - the dictionaries often pick phrases out of the Bible and other texts and run with those, too. You'd think this would be difficult, but surprisingly not. And there's the variations in the above as well.
A brute forcer that uses a dictionary often enlarges it through variations, which is still far less to check through than a full test-every-combination brute force.
About the only choices left are pure random passwords that the only way to break them is testing every combination.
What is the quality of the password then?
Why are we still using passwords?
Time to deploy client certificates. That can be done pseudonymously. And with Tor even anonymously.
http://eccentric-authentication.org/
As a very well known xkcd points out, a great deal of the problem could be averted if people weer encouraged to use long passphrases with spaces and everything rather than a pass'word'. password as a concept was good enough for the time of it's popularity, to defend against people typing their way into someone else's account. When the model fell apart in a world with much more automation and network connectivity, the 'fix' was 'keep length about the same, but toss some numbers and maybe some punctuation in there'.
The madness comes in when a great deal of the sites I visit put a 12 character *maximum* on a password for their site.
My personal strategy: base64.b64encode(os.urandom(12)) for every site and store the values on a couple of my devices with a phrase that is about 32 characters long (but easy for me to remember and easy to type). hashing a master key with the domain to generate passwords like some chrome and firefox plugins (password hasher) can do is similarly nice without having to worry that you won't have access to the copy of the database.. Of course, the annoying thing is my 16 random numbers and letters frequently fail the 'complexity' check and I have to add some punctuation character to it.
XML is like violence. If it doesn't solve the problem, use more.
Please tell me no one is surprised by the general conclusion (haven't we been here a time or ten before?) of these studies. Add to this the corporate or government attitude demonstrated so equivalently here, the lack of effective computer security training, including a complete failing of organizations to have or heaven forbid enforce policies about password practices and you've got a pretty pickle.
Sadly, it took the recent Adobe compromise, to get me to finally start using a password wallet and use different passwords for each Internet service I use. Have to admit I was stunned, by the number of accounts I had when I got through most of the sites I access.
After hearing a few disturbing stories from my wife, about how computer security and passwords are treated at her place of work, I stepped up my training for her and her co-workers that will listen. Based on what I've heard from her the choice of poor passwords is the least of our troubles.
Unless the underlying problem of poor culture surrounding computer security is changed and an understanding of the associated risks is cultivated, it won't matter one whip whether users can choose "Good Passwords TM".
" up to 80% of choices come from just 100 different numbers."
It gets worse, as 100% of those are chosen from just 10 numerals.
If we start with the asumption that that passwords must be memorized somewhat, we are better remembering things with an attached meaning than something random, and those meanings make usually bad passwords. But, we don't need to remember all passwords, there are password managers for making and storing a bunch of meaningless, secure passwords, and for the keys you must remember (the password manager one at the very least) there are some mnemonic tricks that can help to have safe enough passwords.
Must be an idle day at the BBC. A couple paragraphs of statistical wank about physical attributes seeming to correlate with password quality. Then a rehash of old news about bad passwords being easy to crack. My hair is unkempt and I have a 62 character password encompassing a good chunk of ASCII printable characters. Bring on the "compensating for something" jokes. ;)
I am going to shave, so my passwords get better.
Okay, how many of you use the digits of pi when you have to pick your own PIN?
On passwords, what was once thought to be good password security is no longer true. The length of a password matters more than diversity and given the right instructions, can be much easier to remember than complex passwords.
My current suggestion for passwords is this: Pick three (or more) random words. mongoose, screwdriver, automobile. Now you have a password you can remember, but is very hard for a computer to "crack" and you only have to remember three things, as opposed to memorizing eight (or more) things that don't make any sense.
And, to make it unique for each System you log in to, add in the name: Amazon Mongoose Screwdriver Automobile, or Ebay or whatever.
Agent K: A *person* is smart. People are dumb, stupid, panicky animals, and you know it.
I devised my best password for my luggage. I'm too tired after doing that to worry about online passwords
rewriting history since 2109
I love them.. I trawl through them laughing at the passwords on them, at least so far as mine have never shown or close variants of them.
You have 5 Moderator Points!
Which Helpless Linux zealot/MS basher do you want to mod down today?
It's long been known that using a password is insecure and dangerous. Public key authentication is the bare minimum I'd accept these days.
My problem is being able to correctly type long character strings containing caps and special characters without visual feedback.
I could make my passwords much longer if I could see them as I type them.
there are a lot of sites, that require setting up and account, i could care less about. i use a junk email account and a simple junk password. those accounts, if they are hacked, won't give you any useful information to get into another site's account that i do care about. i think many people do the same. those junk sites also get hacked and the stolen lists get published. then the appalling headlines stating "OMG these passwords are so easy!!!" get published... so what...
"people are lousy at picking good passwords"
This begs the question. There is some reasonable expectation that people should learn to properly use the tools of modern society, but in the end, the tools should serve the people, not the other way around. If your car pulled to the left, would you say you were lousy at driving in a straight line? No, you'd say your car was out of alignment and get it fixed.
A password is something we're expected to remember, but we're wrong to pick words or numbers that might be easy to remember, such as familiar names or dates. Even if you say pick a system of choosing passwords to remember rather than an individual password, that's impossible. Every different system and site has different password requirements, so no single easy to remember system will work for all of them.
"You have to remember we are all human and we all make mistakes"
Yes, and Mr Thorsheim's mistake is assuming the issue is with the people who are using the system and not the people designing the system. The truth is,
"password systems are lousy at serving people."
(as an aside, WTF is up with systems that do not allow special characters in passwords? Are they worried about SQL injection? If that's possible from a password field, the system is FUBAR.)
I would hope the list of allowable PINs is shorter than that. The 10 possibilities with the same number repeated all the way through should be disallowed (and usually are), as well as 1234, 4321, and anything else with four consecutive digits. While taking those 24 possibilities out doesn't dramatically reduce the number of possible PINs (only 2.4% reduction) it is still a list of less than 10,000.
Damn_registrars has no butt-hole. Damn_registrars has no use for a butt-hole.
Humans are no good at generating passwords. That is just a fact. The best option is to use a password generator and to change the passwords often. I started using keepassx a couple of years ago and I have never looked back.
I have a really really good password that I use to get into my server at home. All other passwords are for random sites (like slashdot) and I use a very simple password for them. Does this make me 'bad at picking passwords', or do I simply not care if someone hijacks my slashdot account, ruining my excellent karma?
A good password is one that you don't mentally consider a word or string of words, as much as it is a dance that you do with your hands and fingers, really really fast.
Politics; n. : A religion whereby man is god.
Hire Allyson Hannigan to choose your passwords. PROBLEM SOLVED in sexiest way possible.
Gamingmuseum.com: Give your 3D accelerator a rest.
... That the average inhabitant of the galaxy has 2.4 legs and owns a hyena, right?
I use regexes related to the site name/function. (*)
Now the hackers have 2 two problems when they want to break into my account!
* I actually I do incorporate regex like strings.
I am Slashdot. Are you Slashdot as well?
Any grade school education in a developed country ought to include at least a couple of hours on computer security, including how to pick a good passphrase. Everyone doesn't need to learn information theory and complexity theory, but teaching kids that passwords have different amounts of entropy depending on how they are chosen and which level of entropy you need in order to be safe against various types of attacks should definitely be possible. Roughly 10 bits for every word chosen at random from a list of 1000 different words. Roughly 80 bits needed for protection against most hackers. 8 words needed. Here's a card of printed words and a dice. If they can learn history and biology they can learn this.
Are so good, that they cannot be remembered, and need to be stored and then they are not passwords but tokens (something you have, not something you know). And then they become like keys that they give you for your house or car or your workplace. I understand why passwords came to be used on computers when hardware and software was much more limited, but in 2013 and beyond it would seem like a more reasonable approach would be to use a simple hardware solution via USB and replicate the data. But then nobody would trust the replication people, so we end up with passwords that cannot be remembered which become keys which are not replicated.
I got a new job some time ago, and they don't believe in single sign on or even using one password for common users (ie, root), and its a PITA to have to copy/paste a password list all day long. I don't know my passwords anymore and don't even try to remember them.
The only thing that is important is that your password cannot be guessed. Using highly complex cipher algorithms to create overly complex strings of letters/numbers/symbols is a waste of time. Just come up with something that isn't a dictionary and is reasonable in length. Depending on how much I care about the security of an online account I use 1 of 6 different passwords ranging from 5 characters (made-up/non-dictionary word) to a 12 character string of numbers/letters/symbols that is easy to remember (it's kind of a phrase but non-dictionary). Never been hacked in over 20 years of using the same passwords. The only password I ever bother to modify over time is the one for my checking account and email account since those are the only 2 that could have a direct impact on my life. If someone stole my credit card info I couldn't care less, but if someone drained my checking account, that would be a major nuisance.
Why can't my home computer manage passwords. Seems like it's smart enough to generate a password, pass it to the secure site, then at log off generate another password pass it to the site and then log off. Let the computers handle the task. Then have one master password or some other technique to log onto the computer that can only be used from the keyboard.
How do you get a bunch of people to give you their passwords? Sounds like someone has set up a scam site that doesn't hash passwords.... I wonder if we should trust people like that?
Every time I see articles like this, I feel compelled to bring up the solution I'm using, which is (so far) the single best solution I have been able to find.
It's called 1Password. Runs on Mac, Windows, Linux (read only I think), iOS, Android, and has plugins for all major browsers.
It records your login details for you, has a password generator that you can customize in various ways, and stores an AES encrypted archive on dropbox so that all your devices can sync together.
Now I can safely create new logins everywhere with abandon, because I'm not afraid that if one service is compromised (*cough*Adobe*cough*) I'm not afraid something else is at risk.
It can generate passwords up to 50 characters in length with your choice of number of digits and symbols. It can even make easily pronounceable passwords if you need, and avoid ambiguous characters (eg O (oh) and 0 (zero) ).
It's a little pricey, but IMO it's worth every penny because there is no other product out there that is this easy to use, AND supports so many platforms all at once.
I also blame sysadmins who frequently don't understand that security is contextual; you do not need the same level of password complexity for a gardening forum or slashdot that you need for your bank account. But you still see ridiculous requirements for low-security sites.
If pass phrases are inherently far more secure, why do we still prompt people to create and use a *password* and then make a big stink that they did *exactly that*? Just because they do that poorly we shouldn't hold that against them since the process itself doesn't do anything to help them do so better--it's actually at odds, whereas simply indicating the different process of selecting a pass *phrase* does.
Why not simply change the labels and validation (since when should a site ever *prohibit* any specific character from a pass phrase?!!) to say "pass phrase" to urge people in a better direction?
We have bone-headed developers that have "helpfully" sent out emails to every member of a site saying "to improve security we have stripped all non alpha-numerics from your password"... Huh????? a) that means you stored my pass phrase *in plain text* in your database, then b) you *shortened it*! and c) you reduced the available combinations and d) turned my pass phrase into a password.
We have *banks* adding "site lock" security--reducing the security of their websites and *lying* to their users telling them that a) it increases their security and b) *trust the site lock image to indicate that it's really the correct site* rather than educating them to check the *SSL cert*!
Perhaps we need an article similar to "what every developer needs to know about character encoding" but for "handling user credentials". It's obvious that it's not just users that don't get it--but many developers and businesses also.
The proper way is to use a good password manager with the following features:
1) cloud-based sync, so you can access it from any computer or mobile device
2) multifactor authentication, such as a USB stick or a grid or biometrics
3) a configurable password generator (i.e. you can choose length, complexity, etc.)
I use LastPass and like it enough to have bought a year's subscription for $12, but there are other good choices out there like 1Password, or you could homebrew up something with e.g. DropBox + KeePass or Google Drive + TrueCrypt + something that can read TC volumes on iOS/Android.
Generate a different random password for each site needing an account, as complex and as long as the site will allow for, and with LastPass at least you can attach a note to each site's entry so you could enter random line-noise answers for security questions like "What is your mother's maiden name?", thus making crackers work much harder. I've also got LP set up for multifactor authentication and with a strong master password.
Hail Eris, full of mischief...
E pluribus sanguinem
That's his password, you insensitive clod!
Oh yeah.... I really love it when I go to a site and try to create a password with punctuation, and it gets kicked because the site doesn't support it.
Really????
I'm talking about some major sites... financial institutions too. Scary and unacceptable.
Huh?
His password is open source and everybody is entitled to read it, modify it, or to sell it as text source if he can find a buyer, as long as the copyright notice remains attached!
I think the website's logos are blue because the marketing department saw that everybody was choosing "blue" for their password...
No, RED! Aaaaaaaaargh!
No sig today...
The most likely way for you to lose your account credentials is poor security on the site where you entered them. It doesn't matter how long or nonsensical your password is if the website database gives them up in plain text through defective coding of the web application. Or stores them somewhat obfuscated but the accompanying information like the hint or username is in plain text. Or the site gives access based on a defective password reset function. Or your email provider has poor security so your email account is hijacked and the attacker uses that access to perform password resets. Or just about anything besides password-guessing.
In the UK, ANY offence (like shouting "war criminal" at a UK politician who voted to authorise some military atrocity) is an ARRESTABLE offence, under fundamental changes to British Law introduced by Tony Blair. And, any arrestable offence allows the police to raid the subjects home, confiscating ALL records and electronic devices.
So, what does this have to do with 'password' propaganda? Well, the single most common way 'law' enforcement goons use to 'crack' encryption is to locate where a password is written down. The more OBSCURE a password is made, the more likely it is to be on paper somewhere in the vicinity of the computer.
Now, as I type, the BBC is purposely spinning this report to tell the sheeple that passwords are 'WEAK' if part of them contains, say, the name of a pet. Notice I said "PART". This fallacy is the carefully planted NSA/GCHQ lie.
Passwords are NEVER cracked section by section. Sheeple do NOT understand this mathematical fact. Sheeple would think "PEPSIBANANA" was a 'weak' password, because it can be assumed that 'PEPSI' and 'BANANA' are weak, and that a cracking program would first find one 'word' and then the other in the cracking process. NOTHING COULD BE FURTHER FROM THE TRUTH.
For password systems with REVERSE TABLE LOOK-UP KEYS, passwords are mechanically cracked by building up unthinkably large databases of passwords and their encrypted key equivalents. Then the State discovers your encrypted key, and checks to see if it is present in the database. Commercial services offer this facility as a way for people to 'recover' (yeah, right) their 'lost' passwords. Safe encryption does NOT maintain an encrypted key that matches the password.
The BEST password systems allow LONG password phrases that allow the statistical combinatorial options to grow so large that easy to remember strings are impossible to crack, PROVIDED the phrase is memorable, NOT common. A common phrase with one word perversely modified is a STRONG password.
Again, despite the best efforts to lie to you about the subject, passwords are NEVER broken part by part. The entire password has to be guessed by the cracking program, unless you are using weak encryption algorithms (eg., anything mandated by the government or standards bodies). Use Truecrypt with a personal 'perverse' phrase that is so memorable, you never need write it down (and YES, it can safely partly include the name of your pet), and your encryption is UNCRACKABLE.
Use an NSA recommended password like "19!sDF3g99MM28DD" and you WILL write it down, and the locations you store the written password WILL be located by anyone that seeks access to your files. The hiding places for written passwords are VERY VERY small in number, and no matter how clever you think your hiding plan is, you will use one of the same small number of locations the security experts already have on their list from DECADES of experience locating written passwords.
hunter2
Here's a crutch for those with too few passwords on too many sites. Just paste it to something like safepassword.sh in /usr/local/bin or similar:
#!/bin/bash //g"
# script: safepassword
# this script depends on sha512sum
if [ "$2" = "" ]
then
echo "usage: safepassword constant_key password_purpose"
echo " where constant_key is a string of printable non-whitespace characters,"
echo " and password_purpose is a memorable string related to the purpose of"
echo " the password, e.g. a website address and year. Since the script removes"
echo " any characters outside 0-9 a-z A-Z it is possible that the password"
echo " could be too short in some cases."
else
echo -n "$1-$2" | sha512sum | xxd -r -p | tr -cd [:print:] | sed -e "s/[^0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ]//g" | sed -e "s/
echo
fi
And to prevent any of the command lines going into your command history, and thus exposing your passphrases, be sure to run (once on each account that will use the shell script):
echo "export HISTIGNORE=\"safepassword*\"" >> ~/.profile
Since sha512sum should work the same way on all operating systems, a script such as this could probably be made for Windows as well as BSD/Linux/OSX.
Those who can make you believe absurdities can make you commit atrocities. - Voltaire
I will occacionally use the first letter of each word of some lyrics. One password that I have in use is 1@tvm0@mmg1h1@v@m.... and it goes on from there. With some character substitution, it is "I am the very model of a modern major general. I have information animal, vegetable and mineral...." I use a bit more of the song, but you get the idea.
I suppose you could use "f$@7y@...." for "Four score and 7 years ago...." but keep going from there.
red-haired women tend to choose the best ...... and men with bushy beards or unkempt hair, the worst. .....women prefer length and men diversity.
I was beginning to wonder where this summary was going after the first few sentences.
Have gnu, will travel.
I used to have a beard and bushy hair and my password was "test123". After I neatened my hair and shaved, I had this overwhelming compulsion to change my password, and now it's UjuW8LxttbsWKqMbDaA4SqSJVST783ty
My bank card PIN is four digits. It's not the year I was born, nor is it any other year (or other four-digit number, for that matter) that you will find in my personal information.
For computer passwords I like the "first letter of a phrase" algorithm, producing passwords like TbontbTitQ and MRwiTDtESSahtuwws. Or pick a phrase, l33t it up a bit, and come up with something like W1nd0ze1sTehSux0r3. Long passwords are good.
The worst public web site I've encountered for silly password requirements is U.S. Customs eAPIS, which you use to send your information if you're going to fly privately to the U.S.A. Not only does it enforce silly password requirements, it doesn't tell you about them until after you have typed in your new password and it tells you why your password sucks. Yes, I end up writing them down.
...laura
Unkempt hair and bushy beard? Yup thats me. You know, I DID pick out terrible passwords when I was younger and early on in my career. However, being a sysadmin I had to learn to be better. First I thought I improved on my own....then I got called into the security guys office and he pointed at a jumble of letters on the board and said "Recognize anything there?".... my password was clearly embeded in the jumble. Damnit!
Soon after I learned to use mnemonics and never looked back.... not till I found out about passphrases ala xkcd's "Correct Horse Battery Staple", and password vaults. Now I don't choose my passwords, I generate them...and I only have to remember one really good one.
"I opened my eyes, and everything went dark again"
people are as dumb as a box of rocks, and most Obama voters 'round these parts have an IQ of room temperature
"women prefer length and men diversity"
Yes... yes they do.
"If any question why we died, Tell them because our fathers lied."
You jest, but that's not too far off.
[When] passwords first appeared at the MIT AI Lab I [decided] to follow my belief that there should be no passwords. Because I don't believe that it's really desirable to have security on a computer, I shouldn't be willing to help uphold the security regime.
See http://oreilly.com/openbook/freedom/ch07.html
Perhaps everyone quoting that xkcd should be aware that such passwords are no longer safe.
If you think to yourself after reading the first page, "But all of those long passwords were phrases, not nonsense strings!" then you should keep reading to page 2's sidebar for the list of passwords that were cracked using the methods in the article. Crackers have dictionaries of billions of words now and can try combinations and variations at GPU-fueled speeds. Length only protects you if and only if you can exhaust dictionary attacks.
The only safe password is long and either randomly generated or indistinguishable from it. Using some other device to store and auto-fill your passwords like a password manager or a device like a YubiKey is the only long-term solution. Humans are the weakest link.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
I've been saying it for years: length! Thisshittasteslikechicken! Will take many, many years for any algorithm to crack. http://www.securityadminisanidiot.com/ will also assure security. Why don't management and administrators understand this?
If they can crack a website's passwords at GPU speeds it means the site is already been compromised.
That's why I don't bother making really strong passwords for most websites. It's a waste of my time - the site is more likely to get hacked then my password bruteforced over network connections. Every few months there's a web service getting pwned.
It's silly to waste time making your password much stronger than a typical website's admin password.
FWIW I've encountered at least one online bank that actually limits passwords to 8 characters for some unknown stupid reason.
That which is tricksy by nature is tricksy by virtue.
-- Tigger warning: This post may contain tiggers! --
I also blame sysadmins who frequently don't understand that security is contextual; you do not need the same level of password complexity for a gardening forum or slashdot that you need for your bank account. But you still see ridiculous requirements for low-security sites.
So that is who stole my gardening tips!
-- Tigger warning: This post may contain tiggers! --
I hate studies like this. Do people pick common passwords, of course they do. Does everyone pick an easy to guess password, of course not. Can it be blindly determined, for any given user, if their password is "simple" or "complex"? No.
The article puts the blame on the end user, when the truth is the problem is with the websites storing the passwords in plain text or as un-salted hashes and not locking out brute force attacks. What the researchers are really arguing is that
1) your account may be compromised if hackers break into the website and steal all the passwords.
2) your password might be easier to guess if it is related to you, hackers are targeting you personally (not likely), and the website doesn't lock the account out.
Don't blame the user, blame the developers and administrators for being lazy and/or inept and failing to protect people from themselves.
Average Intelligence is a Scary Thing
I don't think you understand the concept that the xkcd advocates.
The ars technica article is pointing out that context can grossly reduce the entropy in any given search space. If you're going to test combinations of words from different languages, for instance, you shouldn't bother with "crotalus fthagn" or "Cthulhu atrox" until you've already tried "crotalux atrox" and "Cthulhu fthagn". The point is that you can't beat the password crackers by picking something from an obscure search space -- in other words, it's a classic point against security by obscurity.
The XKCD is making a different point: that passwords comprised of unrelated words deprive the attacker of such information and are resistant to attack not because of the obscurity of the search space in which they're found, but because of its size. Perhaps 44 bits of entropy isn't enough to defeat extensive computational resources, but the point is that six words chosen out of the dictionary at random, all in lowercase, with spaces between them is a better password than "Cthulhu fthagn" because modern datamining techniques mean that it's likely to appear in someone's dictionary after all.
Public key cryptography is both more secure and more flexible. Why don't websites simply allow you to identify with your SSH or PGP keys?
Even github doesn't support this for the website login yet.
Perhaps everyone quoting that xkcd should be aware that such passwords are no longer safe.
If you think to yourself after reading the first page, "But all of those long passwords were phrases, not nonsense strings!" then you should keep reading to page 2's sidebar for the list of passwords that were cracked using the methods in the article. Crackers have dictionaries of billions of words now and can try combinations and variations at GPU-fueled speeds. Length only protects you if and only if you can exhaust dictionary attacks.
The only safe password is long and either randomly generated or indistinguishable from it. Using some other device to store and auto-fill your passwords like a password manager or a device like a YubiKey is the only long-term solution. Humans are the weakest link.
Using software to store and auto-fill your passwords is the worst possible solution (a post-it on the monitor is more secure in practice). The result of that thinking will be trojan key-stores that simply inform their creator what your password is.
The point of the XKCD is that if you select n random words instead of n random characters you can get a password that can be memorized easily, and exploits the larger search space of words (compared to the smaller search space of characters that exist on your keyboard) meaning your password will be more secure and easier to remember.
The guy who complained loudly about his department introducing the requisite to use a password, and stop having account separation based on trust.
A password is something that, almost by definition, should be hard to guess, have no relation to the user, and be difficult to "shoulder-surf".
As such, the very definition of a password means that they are hard for THAT PERSON to generate, and hard to remember.
This really needs any kind of study or discussion?
That's only true if you never reuse passwords, which means you're pretty much forced to use something like Keypass anyway, and might as well make the password secure since it's just as easy to use a 32 character random string as it is a normal human password. If you don't use a password manager, then it's hard to come up with a memorable password for every goddamn site that needs a login these days. It's so damn annoying to google a problem and find a potential solution, but then click on the link and bet told "you must register a free account before you can view this forum."
Every time someone sets up forum software to require an account to simply read it, they should be kicked in the nuts. Requiring an account to post is totally ok, but requiring an account to read is not.
I read the internet for the articles.
Oh, no: someone hacked into all the silly website accounts I have at once. It doesn't really matter to me if I lose my /., reddit, tumblr, facebook etc. accounts at once. My bank has a good password, as does everything else which could reasonably affect me.
Is 1563649 a prime number?
Such passwords were NEVER safe. The reason passphrases CAN be good is that they can be made easy to remember while STILL BEING RANDOMLY GENERATED. Diceware is a good example: You get a LOT of entropy for each word in the phrase, so a short phrase of 5-6 words gives you a good password. Thinking up 5-6 words will give you a terrible password, since there will be very low entropy in your choices.
Not a sentence!
Perhaps everyone quoting that xkcd should be aware that such passwords are no longer safe.
Nonsense. You don't understand the approach XKCD was suggesting; you can't defeat entropy by getting a bigger dictionary. If that were true, then AES-128 would be trivially easy to crack because I can enumerate all of the possible keys. I have a 100% perfect dictionary.
The point that by selecting a set of randomly-chosen words (do not do the selection yourself; use a random number generator) words, you can get a great deal of entropy in a fairly memorable form. It doesn't matter if the attacker knows the exact method you used (as long as it's random), and knows the exact dictionary you selected your words from; he's still going to have to try 2^n possibilities, where n is large enough to make brute force impractical.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
It meets minimum IRS publication 1075 requirements. See page 104
http://www.irs.gov/pub/irs-pdf/p1075.pdf
Would it keep a government agency from brute-forcing on a super computer? no, but remember most password hacks are on websites, such as facebook, and these attacks go for the lowest hanging fruit (dictionary words and stupid combinations 987654321a). The real problem is with sites which allow brute forcing, I had an old skype account, which had the password brute forced (last year), lesson learnt for me about using a dictionary word followed by numbers, now for semi important stuff, like Skype I use a password which has a common element (including symbols) and the site name in the password, this ensures the stored hash is unique to that site. The other day I had a customer (my company sells software applications), send a scan of their passport to our support email, it was a surprise to us as we never request such documentation. The email he was responding to was from a non-existent address on our domain, when it bounced back to him, he found a working address and sent. The email which he responded to, looked just like one of our emails, but with extra paragraphs inserted, saying for security reasons photo ID was required. It was obvious that his email account was compromised (or servers would never send this email with extra information entered, unless they reprogrammed our backend software), and the attacker was reading all his email (inbox), that document would be read when bounced from our servers. This was a individually targeted attack on that individual (traced to Pakistan - as the attacker clicked on the software download link and was logged), it is scary the length this attack went to to get his passport scan.
...girth is more important.
Not sure how that relates to passwords.
The pursuit of absolute tolerance leads to the most rigorous and ludicrous intolerance. - REX MURPHY
Whether those 'silly website accounts' being hacked is a problem depends on the amount of personal non-public information you have stored there. If enough information is compromised, it becomes really easy to use that information for social engineering purposes. They could simply call up your bank and tell them that 'you have moved to a new address and that you lost your bank card and need a new one'. Usually even they accept things like your DOB as valid identification. Retarded, but true more often than not.
That's why I'm born on 01-01-1970 when anybody who (or entity that) has no fucking business knowing my birth date asks.
Would it keep a government agency from brute-forcing on a super computer?
Depends how many words you use. Use enough to get to, say, 80 bits of entropy, and assuming a decent (slow) hashing algorithm, yes it would.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
I knew my red haired fetish would be justified!
FWIW I've encountered at least one online bank that actually limits passwords to 8 characters for some unknown stupid reason.
Likely the backend system is/was a mainframe with RACF, which can have limitation of 8 characters to their passwords. "Newer" mainframes have been extended to 2 sets of 8 characters for a maximum password length of 16 characters, but most implementation don't utilize the full 16. Now WHY banks are using mainframes to store passwords instead of a different authentication system is another discussion altogether...
All of the examples they gave in the article break one of the fundamental rules in that XKCD strip. The words shouldn't be words that are easily associated with each other. Of course picking a quote straight out of fiction is stupid. Four random 4-6 letter words that don't appear together in common language usage would be harder to crack for people using the strategy in the article http://arstechnica.com/security/2013/10/how-the-bible-and-youtube-are-fueling-the-next-frontier-of-password-cracking/
RMS doesn't HAVE a password. He campaigned AGAINST the introduction of them, ffs.
I don't think you understand the concept that the xkcd advocates.
This.
Also, what the article doesn't say is that the programs it uses as examples aren't that fast over a network, so if they're cracking the passwords at full speed, they've already compromised the site. Network speed plus other forms of detecting password crackers (such as locking out after 5 or 10 attempts) really slow down attempts to crack a password. This is why they tend to use dictionary attacks rather than brute force, dictionary attacks are faster and yield decent results.
The XKCD is making a different point: that passwords comprised of unrelated words
This again,
/.ers to know where this is from) but extremely unlikely to be found in a dictionary attack, especially with the punctuation (which is not 100% correct, but they're mistakes I make commonly, Grammar nazi's can bite me). So the only real way an attacker has to defeat this is via brute force, so the longer and more complex the password, the longer it will take over a network.
/. as you do for your knitting forum isn't that bad. However using the same password on your webmail or work account as you have on Facebook is terrible, so important accounts should have unique passwords whilst ones that are potentially vulnerable (such as a forum for your lawn bowls club) should never use a password that is the same as something important... Doubly so if that password is the same as the password you use on the email address you joined the forum with.
Along with being unrelated words they are easy to remember. For example "Shotgun, Raptor; clever girl" are pretty unrelated outside of the context (and I expect most
The other issue is password reuse.
A lot of people get around password resuse by using a password safe (such as key pass) but all this does is introduce a single point of failure. What people need to realise is that reuse can be managed, using the same password for
Calling someone a "hater" only means you can not rationally rebut their argument.
"...women prefer length..." I feel so lied to.
Perhaps everyone quoting that xkcd should be aware that such passwords are no longer safe.
Nope, that article details methods for cracking known phrases, not non-phrase combinations of several random words. Indeed an poster on that article specifically addresses the crackability of the aforementions xkcd pass "phrase" in this context.
Of course "correct horse battery staple" is now a known phrase rather than a non-phrase combination of 4 random words. However, at least before I posted this, the password "honey $anctify Entropy umlaut m1ll10n" was still safer than "to be or not to be that is the question".
Nonsense. You don't understand the approach XKCD was suggesting; you can't defeat entropy by getting a bigger dictionary. If that were true, then AES-128 would be trivially easy to crack because I can enumerate all of the possible keys. I have a 100% perfect dictionary.
The point that by selecting a set of randomly-chosen words (do not do the selection yourself; use a random number generator) words, you can get a great deal of entropy in a fairly memorable form. It doesn't matter if the attacker knows the exact method you used (as long as it's random), and knows the exact dictionary you selected your words from; he's still going to have to try 2^n possibilities, where n is large enough to make brute force impractical.
Uh, ya the point is that nobody uses brute force, so the "n" in your "2^n" is equal to the number of words in his dictionary, not the length of your passphrase.
Look, I'll try to simplify this for you a little bit. You're picking words, but then counting individual characters. Even if you chose the words at random, the characters in each word are NOT at all random. So picking three words at random, which have a total of 20 characters, does NOT give you a 20-character strong password, it gives you roughly a THREE letter strong password.
No matter how many clever comics you read, DO NOT USE REAL WORDS. Ever. Period.
Because nobody really runs brute force, and when they DO try all possible combinations they START by running through as many real words as possible, as well as permutations, so you're basically shifting the worst-case time-to-crack (for the attacker) from all possible combinations to just all possible combinations which contain real words... a MUCH smaller data set.
Also with that in mind, choosing ANY random password NOT containing real words shifts the time-to-crack in YOUR favor, because they are going to run through all the other combinations before ever starting on the random set.
You didn't do the math :-)
If we were to count letters, the "correct horse battery staple" password would have ~117 bits of entropy (26^25 = ~2^117). But it doesn't, it has 44 bits. This is because it's a sequence of four words selected from a dictionary of 2048 entries. 2048^4 = (2^11)^4 = 2^44.
Assuming a good iterated password hashing function like, say, scrypt, 44 bits is pretty decent, and proof against anyone who isn't willing to throw tens of thousands of dollars at cracking that one password.
FWIW, I don't actually use XKCD-style passwords, not because of security deficiencies but because I have to use my passwords far too often to want to type anything that long. I shoot for 50 bits of entropy, but with shorter passwords. My passwords are generally 8 characters long, unless the character set specified by the system is too restricted to achieve 50 bits, in which case I add characters until I achieve the desired level. 50 bits is arguably excessive, but only if you assume that systems implement proper password hashing, with iterated hash functions and salt. I know from experience that you can't assume that, so I add a few more bits to be sure.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
http://www.bugmenot.com
he's still going to have to try 2^n possibilities
Far from that. 2^n is assuming there is a possibility all the words are used. For 2048 word dictionary, with average word size 5, 2^n means a password of length 0 to 10240 (over ten thousand) characters. If we assume humans are typing, it has to be restricted to less than 100 characters, practically less than 25.
Assume 0-10 words are required for this, reducing 2^n to n^10 (same word can be chosen twice in the same password, of course). Then all permutations of those 10 words are required, so multiply it by factorial 10. Still much lower than 2^n.
Bingo Dictionary - Pragmatist, n. A myopic idealist.
Far from that. 2^n is assuming there is a possibility all the words are used.
No.
With a 2048-word dictionary, you get 11 bits of entropy per randomly-selected word (because 2048 = 2^11). A four-word example like the one Munroe suggested therefore has 44 bits of entropy -- with four words n = 44.
For 2048 word dictionary, with average word size 5, 2^n means a password of length 0 to 10240 (over ten thousand) characters
Ah, I see, you think we're trying to achieve n = 2048? Not at all. The point is to achieve a reasonable level of entropy in a memorable way. If you want a password space of 2^44 with randomly-selected lowercase letters you have to use a 10-letter password, but a sequence of 10 randomly-selected letters is pretty hard to remember. Even if you use an alphanumeric character set, with upper and lower case, and throw in another 10 symbols for a character set of size 72, you'd still need 8 characters.
The beauty of the XKCD approach is that you can much more easily remember four random words -- or four images, especially if you can invent some relationship between them -- than 8 random characters.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Assume 0-10 words are required for this, reducing 2^n to n^10 (same word can be chosen twice in the same password, of course). Then all permutations of those 10 words are required, so multiply it by factorial 10.
Oh, one correction: You already accounted for all the permutations in the initial selection n^10 (assuming n is the number of words in the dictionary). Multiplying by 10! results in over-counting. n^10 is the entropy... and if n=2^11, you've got a 110 bits of entropy which is an incredibly strong passphrase.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Ah, I see, you think we're trying to achieve n = 2048? Not at all
Ok, I am not sure what you meant by n if they have to try 2^n possibilities, and n is not the dictionary size. You still haven't defined "n" for the statement "still going to have to try 2^n possibilities"
Unless you defined n as the log of number of times the cracker has to try. Was that statement meant as a tautology ?
Bingo Dictionary - Pragmatist, n. A myopic idealist.
Right, thanks.
Bingo Dictionary - Pragmatist, n. A myopic idealist.
Yes, 2^n is just the keyspace size. I expressed it that way for analogy with the AES keyspace size, and used 'n' rather than specifying a value because it obviously depends on how many words you use and what size dictionary. I suppose I could have written 2^(word_count * log_2(dict_size)).
The point is that Valdrax was wrong; you can certainly achieve entropy in an XKCD-style key.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
Likely the backend system is/was a mainframe with RACF, which can have limitation of 8 characters to their passwords.
So why doesn't the front-end hash the user's password and use the base64 of the first 48 bits as the mainframe password, preserving up to 48 useful bits of entropy?
a bad password gives the same result as not having a username
In your system, what happens when a user attempts to use an unknown username to register, begin self-service password reset, or visit a user's public profile? A growing number of systems, such as two of the three banks I interact with, will ask for a username on one form and a password on the next to increase security against phishing.
Perhaps everyone quoting that xkcd should be aware that such passwords are no longer safe.
If you think to yourself after reading the first page, "But all of those long passwords were phrases, not nonsense strings!" then you should keep reading to page 2's sidebar for the list of passwords that were cracked using the methods in the article. Crackers have dictionaries of billions of words now and can try combinations and variations at GPU-fueled speeds. Length only protects you if and only if you can exhaust dictionary attacks.
The only safe password is long and either randomly generated or indistinguishable from it. Using some other device to store and auto-fill your passwords like a password manager or a device like a YubiKey is the only long-term solution. Humans are the weakest link.
Using software to store and auto-fill your passwords is the worst possible solution (a post-it on the monitor is more secure in practice). The result of that thinking will be trojan key-stores that simply inform their creator what your password is.
The point of the XKCD is that if you select n random words instead of n random characters you can get a password that can be memorized easily, and exploits the larger search space of words (compared to the smaller search space of characters that exist on your keyboard) meaning your password will be more secure and easier to remember.
Better yet, randomly capitalize and use aural memory to remember where they are. "Correct horse, BATtery staPLE!" If say it aloud a few times (in private, of course), pronouncing it with stress on the capitals, you'll remember it easily, even if it's silly :) Of course you might have to leave out the punctuation, depending on the password field tolerances...which sucks.
"I love animals! Some are cute, others are tasty, what's not to like?" - Betsy Schroeder, Jeopardy contestant
Assuming that wasn't sarcastic, one would define "site's name" in terms of the last part of the hostname before a public suffix. For example, in "it.slashdot.org", the public suffix is "org", and the part before that is "slashdot", giving "sla" as the site name.
Title says it all. You have to remember your password, so you probably won't use a password like "afi9blm#20niv8__q4i".
Pseudo-words - i.e. words that you can read but are in no dictionary - are probably slightly better, but I wouldn't rely on passwords at all in the first place.
BTW if somwone is interested, this tool CAN generate readable pseudo-words like "foliticalling", "uppet" or "furvicially".
FWIW I've encountered at least one online bank that actually limits passwords to 8 characters for some unknown stupid reason.
I used to work for a major US shipping company who has the same policy. They have a corporate single-sign-on system, and one of the systems it feeds into is an old IBM mainframe. The mainframe doesn't allow more than eight characters in a password, so the same limit is imposed on every other system.