Analysis of 400,000+ Stolen Yahoo! Passwords
On 12th July 2012, more than 400,000 emails and passwords for Yahoo! Voices were stolen via an SQL injection and published online. The passwords were reportedly stored in plaintext, making this security breach even more serious. If you are a member of Yahoo! Voices, change your password immediately, and if you use the same password on other sites, make sure to change them as well.
I performed the following password analysis with the help of pipal, a very popular and powerful password analyzing tool. The full pipal report is located here, with a longer report (showing the top 100 of each category) here.
10 Most Popular Passwords
123456 = 1667 (0.38%) password = 780 (0.18%) welcome = 437 (0.1%) ninja = 333 (0.08%) abc123 = 250 (0.06%) 123456789 = 222 (0.05%) 12345678 = 208 (0.05%) sunshine = 205 (0.05%) princess = 202 (0.05%) qwerty = 172 (0.04%)
Despite numerous warnings by security professionals, the most popular password is still “123456″, followed by “password” in second place. These are highly insecure passwords, not just because of their length or complexity (which is very low), but because they are at the top of most password lists that attackers use to try to compromise an account. Remember, brute-forcing a password is always a last-ditch attempt at gaining access to an account; a clever attacker will always try common passwords first, and if your password appears in a password list online, you should never use it!
The fact that these passwords were even allowed reveals substandard practices in Yahoo’s password policy. To boost security, a user should be required to have a password that contains both upper and lowercase letters, as well as numbers and symbols. For additional security, the chosen password should be rejected if it matches one found in common password lists.
8 = 119214 (26.92%) 6 = 79650 (17.99%) 9 = 66058 (14.92%) 7 = 65654 (14.83%) 10 = 54815 (12.38%) 12 = 21785 (4.92%) 11 = 21261 (4.8%) 5 = 5325 (1.2%) 4 = 2748 (0.62%) 13 = 2585 (0.58%) 14 = 1433 (0.32%) 15 = 773 (0.17%) 16 = 442 (0.1%) 3 = 303 (0.07%) 17 = 252 (0.06%) 20 = 169 (0.04%) 18 = 116 (0.03%) 1 = 116 (0.03%) 19 = 78 (0.02%) 2 = 67 (0.02%) 21 = 6 (0.0%) 22 = 4 (0.0%) 29 = 3 (0.0%) 30 = 2 (0.0%) 24 = 2 (0.0%) 28 = 2 (0.0%)
As you can see, most people are still using short passwords. Indeed, a whopping 61.66% of people are using a password that is 8 characters or shorter. If you include passwords with a length of 9 or 10, then the number jumps to 88.96%. When a dictionary attack fails, the main thing stopping a brute-force from succeeding in a specific amount of time is the length of the password. For each additional character a password has, the amount of time needed to brute-force it increases by a factor of 95 (assuming the brute-force is trying all types of character). Even if the password only contains lowercase letters, an additional letter will increase the time required by a factor of 26.
8 characters and longer is usually cited as the recommendation for password length, but with cracking speeds up due to improvements in processing power, that number should probably be closer to 12, if not more. Remember, a long complex password need not be hard to remember.
Only lowercase alpha = 146512 (33.09%)
This small statistic shows a staggering lack of password complexity. Almost a third of passwords only contained lowercase letters, making the task of brute-forcing them much easier.
loweralphanum: 224085 (50.6%) loweralpha: 146512 (33.09%) numeric: 26080 (5.89%) mixedalphanum: 23233 (5.25%) loweralphaspecialnum: 6053 (1.37%) mixedalpha: 5122 (1.16%) upperalphanum: 3416 (0.77%) mixedalphaspecialnum: 3327 (0.75%) loweralphaspecial: 2103 (0.47%) upperalpha: 1776 (0.4%) mixedalphaspecial: 489 (0.11%) upperalphaspecialnum: 233 (0.05%) specialnum: 189 (0.04%) upperalphaspecial: 51 (0.01%) special: 20 (0.0%)
As these additional statistics show, more than half the passwords only contained lowercase letters and numbers (the numbers only increase the brute-forcing attack by a factor of 10). Barely one percent of the passwords could be considered “complex”, containing upper and lowercase letters, numbers, and symbols.
Yahoo! is of course to blame for the passwords being accessible to hackers, as well as storing them in such an insecure way. Their password policy which apparently lets users choose single characters for a password is absurd, and a full investigation should be carried out to find out how on earth the users were left this vulnerable. There were some decent passwords in the list, and those were made completely useless through Yahoo’s ineptitude.
That said, it should be noted that regardless of Yahoo’s ineffective defences and security policies, a great deal of these user chosen passwords were highly insecure. It is up to the user to choose a decent password, rather than relying on a system which you should not really trust (as users, we do not know what security weaknesses a system has, or how it stores important data). It is best, therefore, to create a unique complex password (or passphrase) for each account you have online, and to use a good password manager to help you keep track of them.