Here, your solution is to parse the strings as floating points rather than integers. For example:
var a = '20.796',
b = '20.190';
if (parseFloat(a) > parseFloat(b)) {
// TODO: .. code ..
}
Your code right now is parsing the strings as integers. Integers are whole number values and CANNOT contain a decimal value, meanwhile floats or floating point numbers CAN contain a decimal value. When you call 'parseInt()' on the floating point, it truncates (or removes) the decimal value and just keeps the whole value. Which is obviously not what you're looking for.
PS: I'm guessing you're new to JavaScript, and so I just want to wish you good luck with learning it. Personnaly, I find JavaScript to be a very beautiful language if you learn it well.
parseFloatinsteadif (parseFloat(a) > parseFloat(b))... that would work fine for your particular case