# The measure of Disorder?

1. May 2, 2005

### nmondal

Ok, so we have a parameter X say.
Now, we have a mean x-bar, that is the mean, X being random.
Now, by basic statistics, the variation , that is

V(X) = E (X)^2 - (x-bar)^2

is the disorder. Is it so?
The interpretation of the disorder by me is correct or not?
I have nice property of disorder then!

:!!)

2. May 3, 2005

### El Hombre Invisible

I don't think so. Say X is the position of a molecule of a gas in a box. You have determined the variation from the mean position of every molecule in your function. How big is the box? One number that would describe the variation in a box, say, a metre cubed may well seem disorderly, but if the box was a kilometre cubed the same value would seem highly (almost impossibly) ordered.

3. May 3, 2005

### nmondal

Position does not makes sense in case of a Gas.
It makes sense sometimes in the context of a drunk person.
Think it from that perspective.

Cheers.
BTW, if the box is 1 KM, sir, then again we have certainly more disorder to achieve...remember the diffusion?
That is the phenomenon.

4. May 3, 2005

### El Hombre Invisible

Eh? I'm not sure if you're arguing with me or agreeing. Yes, you may calculate another value and show that the disorder has increased, but nonetheless your original value did not say anything about the disorder of the system and you would no reason to believe the next one would. Even if you take the size of the box into account and make sure you take your measurement when the variation had converged, the variation would still change with other parameters (such as sample size) while the disorder would remain constant.