Understanding Offline Password-Cracking Methods: A Large-Scale Empirical Study

Figure 4

Further evaluation of cracking under extensive-knowledge. Each value in this figure represents the fraction of each length or structure type of passwords in the cracked dataset accounts for the corresponding subset of passwords in the targeting dataset (e.g., in [1], 0.90 indicates that 90 percent passwords with length of 6 in the Rockyou targeting dataset can be cracked by the Rockyou-trained Best64 method). (a) Rockyou-length based evaluation. (b) Xato-length based evaluation. (c) Yahoo-length based evaluation. (d) Duduniu-length based evaluation. (e) CSDN-length based evaluation. (f) Sinaweibo-length based evaluation. (g) Rockyou-structure based evaluation. (h) Xato-structure based evaluation. (i) Yahoo-structure based evaluation. (j) Duduniu-structure based evaluation. (k) CSDN-structure based evaluation. (l) Sinaweibo-structure based evaluation.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

(i)

(j)

(k)

(l)