Hi guys, I'm building an AI bot to pass Jagex's turing text (to bot old school runescape without getting banned).

Each action waits till the client knows it (0.00001ms) then adds a HUMAN reaction time. This is to model a human looking at the screen and reacting to the change before performing the next click.

I'm currently implementing a poission distribution with a mean/lamba of 273 ms ( recalculated each poll of the finite state machine, **currently doesn't store previous values....would this cause divergence with limited sampling?).

This should make almost all the RTs fall close to that 273ms value, and a very low falling off chance of much greater RTs that simulate "local" fatigue.

Later I am thinking of synchronizing a mental fatigue weight with the time of day to simulate adenosine cycle fatigue.

Any ideas on other ways to avoid my bots being detected by ML algorithms and chinese slaves who are forced by jagex to hand review flagged accounts suspected of botting?

Here's my java code for the Poisson distributed reaction time too, appreciate if anyone could review it:

I will also post my other anti ban measures ITT my script currently has and may have in the future.


public int getPossion(int lambda) {

int sum = 0;
int n =-1;

while (sum < lambda) {
n += 1;
sum -= Math.log(random());
}
return n;
}

public static int getPoissonRandom(double mean) {
Random r = new Random();
double L = Math.exp(-mean);
int k = 0;
double p = 1.0;
do {
p = p * r.nextDouble();
k++;
} while (p > L);
return k - 1;
}