mortgage calc returning negatives

Question

I'm trying to deal with money (probably the wrong way as far as best practices but I thought it would work.. its simple, multiply everything by 100 and then use ints)

    import java.util.Scanner;

public class PFiveFifteen {

public static void main(String[] args) 
{   
    Scanner in =new Scanner(System.in);

    int i = 0;
    double principalIn;
    int duration=0;
    double interestRate=0;
    int thisYear = 0;
    int iInterestRate = 0;
    int iPrincipalIn = 0;

    System.out.println("Please input initial amount: ");
    principalIn = in.nextDouble(); 
    iPrincipalIn= (int) principalIn *100;
    //System.out.println(principalIn*100);

    System.out.println("Please input interest rate as a decimal: ");
    interestRate = in.nextDouble()*100; 
    iInterestRate = (int) interestRate;
    //System.out.println(interestRate*100);

    System.out.println("Please input duration of investment in years: ");
    duration = in.nextInt(); 
    //System.out.println(duration);

    while(duration > i ) {

        if ( i == 0) {
            //System.out.println("Preloop value" + (iInterestRate));
            thisYear = (iPrincipalIn * iInterestRate)+ iPrincipalIn;
            //System.out.println("Postloop value" + (iPrincipalIn));
            i++;
            System.out.println("The initial value is " + thisYear);
            System.out.println("Year # one one: " +i);
            }
        else

    thisYear = (thisYear*iInterestRate) + thisYear;
    i++;
    System.out.println(thisYear/10000);
    System.out.println("year # " +i);
    }

But I'm somehow getting negative numbers back in console

    Please input initial amount: 

10000

Please input interest rate as a decimal: 100.00

Please input duration of investment in years: 10

The initial value is 1411065408

Year # one one: 1 141106

year # 2 -119738

year # 3 -72104

year # 4 4904

year # 5 84261

year # 6 21917

year # 7 154073

year # 8 -145664

year # 9 63722

year # 10

What in the world is going on here? I could get if there where whacky decimals or soemthing ( even though I cast it all to ints) but what the dickens is taking all positive inputs to below zero??

Answer 1

Your idea of representing money amounts as integers is an OK one, but you have to be very careful of the scaling, especially when you multiply (or divide). Say you start with a principal of $10, which you represent as 1000. Then you enter an interest rate of 0.06 (I assume that the user is entering an actual rate and not a percentage, so 0.06 means 6%). You then represent that as 6.

So if your principal is P, the int you're using to represent it is P*100. And if your interest rate is I, the int you're using to represent it is I*100.

So now you want to compute P*I. Since you're representing everything as 100 times the actual value, you presumably want to get an int whose value is P*I*100. But when you multiply the two ints you have, P*100 and I*100, the result is P*I*10000 -- a factor of 100 too high. That's because you multiplied the scale factors together along with the values. I can see that you did output thisYear / 10000, but you didn't change thisYear. So as you keep going through the loop, you will get values that are a factor of 10000 too high, and 1000000 too high, and so on, and eventually you will overflow. (I wish things really did work that way, but unfortunately I can't talk my bank into using your algorithm.)

To adjust, you'll need to divide this result by 100. (I'd add 50 to it first, so that when you divide by 100 the result will be rounded.)

Answer 2

looks like using ints was just ill conceived.

I did get it to play right with smaller numbers (the biggest issue was having to again bump up the principal by 100 to keep the correct place holder for when the interest which was (p*100)*(i*100) otherwise it was something *100 being added to the interest which was *10000. Sadly any realistic amount of money (like say 250000 for a house) overflows it before the first loop. Though 200000 makes it all the way to loop 30.

Plus interest rates aren't always hundredths, sometimes they are something like .03875

My code lacks the ability to determine the length and even if it had it bumping that up to an int, and keeping everything else like it (multiplying everything by 100000) would definitely not work.

Props to community for not calling me an idiot though! on Youtube et al, I would have gotten answers that would have put me into a corner crying!