subhajeet
I totally agree with the example you have mentioned above, but in your case it is definite that we will start with values of x that are greater than 8, it can also be 8.1.
In the question which we are discussing, it has only been specified that A and B are non negetive integers. Nothing has been specified whether A>B or vice versa, so I took those values using which I could solve the problem.
If there is any other please please mention it here. It will help us.
Here is why I think your logic is wrong.
subhajeet
Here using the conditions we have to prove that A^5> B^2
As you mentioned, we are given the following: A > 0 and B > 0
And asked: is A^5> B^2?
At this point, we do not have sufficient information to answer this question (otherwise it wouldn't be a GMAT DS question!). Lets delve into the conditions to see what further information we can glean that might help us answer the question.
subhajeet
Condition1: A^1/3>B^2 -> If I put A=B=2 or A=1 and B=2 then the condition becomes false.
This is where I believe things go wrong in your logic. We are told that A^(1/3) > B^(1/2). This is now an immutable truth, no matter what the true values of A and B are, they must fit the above equation.
At this point you have suggested testing A=B=2; A=1 and B=2
2^(1/3) > 2^(1/2)
1.25 > 1.4
1^(1/3) > 2^(1/2)
1.25 > 1.4
You are correct that these do not hold under 1), but that's the exact information that 1) offers us! Do you see how this is akin to:
Is x > 10
1) x > 8
Let's assume x = 7...
We can't assume that A=B=2 because from 1) this is clearly not true.
The numbers you should be picking are those that
are true under 1). Then we test those against the question. If we can find numbers that
are true under 1), but give contradictory answers to the yes/no question, then we can rule out 1) as offering useful information.
So for instance, some pairs that we might test are the ones that fluke suggested:
A=0.5;B=0.5
A=2;B=1
Fluke hasn't just chosen these randomly, he has chosen them because they hold under 1), yet give us contradictory answers when put back into our yes/no question.
.5^(1/3) > .5^(1/2)
.8 > . <--- is true!
2^(1/3) > 1^(1/2)
1.25 > 1 <--- is true!
OK, so now we have some pairs of numbers that are true under 1), but we still can't answer the question just from this alone. We need to go back to the question, and see if these numbers give a contradiction. Let's have a look:
Is A^5> B^2?
A=0.5; B=0.5
.03 > .25.
No.
A=2; B=1
32 > 1
Yes.
Now we see why 1) is not helpful to us. We have picked two sets of numbers that are
true under 1), but give contradictory answers to the yes/no question. We can now stop testing numbers, and move onto 2).
The numbers you used to rule out question 2) were not helpful for the same reasons as above.
I highly recommend you focus on the numbers fluke suggested and compare them to the ones you used. It's really important to get this approach locked down for DS.