fta2012/replicatedrandomJava Random state replication tool that predicts future random values by recovering internal state from observed outputs, demonstrating weaknesses in LCG-based PRNGs.
See README in node folder
Given a double generated from nextDouble of a java.util.Random, recreate a Random with the same state. This ReplicatedRandom can then be used to predict the future values of the original Random.
Example:
Random r = new Random();
ReplicatedRandom rr = new ReplicatedRandom();
rr.replicateState(r.nextDouble());
System.out.println(r.nextDouble() == rr.nextDouble()); // True
System.out.println(r.nextDouble() == rr.nextDouble()); // True
System.out.println(r.nextDouble() == rr.nextDouble()); // True
Also works with Math.random() since it uses Random internally:
ReplicatedRandom rr = new ReplicatedRandom();
rr.replicateState(Math.random());
System.out.println(Math.random() == rr.nextDouble()); // True
System.out.println(Math.random() == rr.nextDouble()); // True
System.out.println(Math.random() == rr.nextDouble()); // True
Also works for nextInt() but requires two consecutive values:
Random r = new Random();
ReplicatedRandom rr = new ReplicatedRandom();
rr.replicateState(r.nextInt(), r.nextInt());
System.out.println(r.nextInt() == rr.nextInt()); // True
System.out.println(r.nextInt() == rr.nextInt()); // True
System.out.println(r.nextInt() == rr.nextInt()); // True
See this blog post for details.