Underground fluids: non-zero MinAmount heavily skews effective maximum

[MauveCloud]

In a recent playthrough (playing locally, not on a server), I decided to be more generous with the overworld oil, setting it to minimum 100 and maximum 6250 for each type (as opposed to 0-625 default), and the first oil drilling rig I set up got me 10885 L of light oil per cycle. That's running at MV, so presumably the base rate was actually 21770 L/cycle, way above the maximum I set in the config.

I tried separately testing the algorithm from

GT5-Unofficial/src/main/java/gregtech/api/objects/GT_UO_Fluid.java

Lines 55 to 60 in 77a3c95

	public int getRandomAmount(Random aRandom){//generates some random ass number that correlates to extraction speeds
	int div = (int)Math.floor(Math.pow((MaxAmount-MinAmount)*100.d*DIVIDER, 0.2d));
	int min = (int)Math.floor(Math.pow(MinAmount*100.d*DIVIDER, 0.2d));
	double amount = min+aRandom.nextInt(div)+aRandom.nextDouble();
	return (int) (Math.pow(amount, 5) / 100);//reverses the computation above
	}

and found that configuring for 100-6250 actually produces output of around 91-35241. The default 0-625 config produces output of 0-624, which is close enough. However, just changing the minimum to 1 shifts the output to 0-1831.

[MauveCloud]

I don't really understand why the current method is using the min and max to calculate a parameter for the nextInt function, but the failure when MinAmount is non-zero is probably related to the Binomial Theorem. I tried a few alternatives, and I think something like this will have the desired distribution and handle the min and max correctly:

    public int getRandomAmount(Random aRandom){
        int range = (int) Math.floor((MaxAmount - MinAmount) * DIVIDER);
        int min = (int) Math.floor(MinAmount * DIVIDER);
        int mult = (int)1e6;
        double result = min + Math.pow(aRandom.nextInt(mult), 5) * range / Math.pow(mult, 5);
        return (int)result;
    }