A simple derivative that I can't get for the life of me

1. Oct 16, 2012

trap101

Trying to get the 2nd derivative of this expression:

ps/(1-qs) where p,q are constants and s is the variable. The solution is 2pq2(1-qs)3

for the life of me I can't get that solution. every time I do it I get 2q/(1-q)2

maybe I should mention that I'm using a probability generating function and trying to get the 2nd moment for the geometric distribution. But that shouldn't have an effect on the simple fact of the 2nd derivative.

2. Oct 16, 2012

jhae2.718

Make sure you use the product or quotient rule and apply the chain rule properly.

3. Oct 16, 2012

trap101

I've been doing that.

4. Oct 16, 2012

Dick

Shouldn't your answer at least depend on s? That said, I don't get the answer you show as a solution either.

5. Oct 16, 2012

ehild

How did you loose the "s" from the denominator? Show your work in detail.

ehild

6. Oct 17, 2012

Levi Tate

I didn't get that answer either that you show as a solution.

I got qp+q^2ps/(1-qs)^3

7. Oct 17, 2012

phiiota

whenever i have a problem like this, i like to move my constants out front and rewrite things in fractions as negative exponents. in your case, i would rewrite the function as

ps/(1-qs)=p*(s)(1-qs)^-1. then, ignore the p for now, since you can just multiply it out, and focus on find the derivative of s*(1-qs)^-1. should be pretty easy from there out. if i want to be more explicit, i write things out in function notation first, apply the chain rule/product rule (as i would here) with f(s),g(s), etc, and then plug in f(s),f'(s), etc, after i've figured out the proper form.

8. Oct 17, 2012

skiller

...which is also incorrect.

9. Oct 17, 2012

Levi Tate

Is it, I did the problem while sleeping, why don't you post a solution oay?

10. Oct 17, 2012

skiller

I could post the solution, but it's not the done thing. People asking the questions should show their work so far. The OP hasn't yet.

11. Oct 17, 2012

trap101

the reason the "s" disappears from the solution is because it is a PGF from statistics so after you differentiate you let s = 1 and obtain your solution.

12. Oct 17, 2012

Dick

"But that shouldn't have an effect on the simple fact of the 2nd derivative." You said you were just trying to find the second derivative. So why is there an s in the given solution?

13. Oct 17, 2012

skiller

Ok, so you haven't asked for the 2nd derivative at all, then. Make your mind up.

14. Oct 17, 2012

skiller

Are you suggesting yours is correct?

15. Oct 17, 2012

Levi Tate

No buddy i said i was falling asleep, so i am going to do the problem again now and i don't know if this boy here is talking about statistics and i don't know if i am right, i never know if i am right, but this here is what i get,

D^2y/(dx)^2= 8pq^3/(1-qs)^5

oh i took the second derivative of the second derivative that was given in the problem.

So it would be the 4th derivative of the solution given in the initial problem, which was said to be wrong for that problem, so this is just the second derivative of that proposed solution then, eh.

16. Oct 17, 2012

skiller

Fine. No problem, try again tomorrow.

Last edited: Oct 17, 2012
17. Oct 17, 2012

Ray Vickson

You say "I can't get that solution". I should hope not: the correct solution is
2pq/(1-qs)^3. Can you get that?

RGV