hey buddy,
i normally go for -2,-1, 0,1,2
similarly you can choose for y

This completely depends on the problem type - in normal problems I use integers between 1 and 9 - but fraction problems you would be better using a common denominator. In must be problems or on a data sufficiency you have to use numbers that help you on the specific concept - therefore if you are dealing with an inequality problem you have to work with positive/negative numbers but if you are dealing with an algebra based problem other integers might prove better.

I prefer to stay away from picking number on DS unless i have to - look for that core concept instead and you should do much better.

Becky

Becky,
you are absolutely right. For data-sufficieny integer questions the more we use algebra and simplify or understand the core concept, the easier the question gets. And in many cases picking numbers only after we have combined the inequalties seems to be the correct approach.

Picking numbers in problem solving questions requires a different approach than it does for data sufficiency problems.

Beware of one and zero because they give atypical answers, especially if there are powers or roots involved. If you have to take the root of a variable pick a perfect square such as 4 or 9. If the answers are fractions and you notice that the denominators tend to be the same, or multiples of the same number, choose the largest multiple. If the variables are multiplied by constants, a lowest common multiple can often simplify the problem.