Why you need to do well on last few questions on GMAT?

zuleron wrote: 5) experimental questions: I'm still not sure. From what has been said I think you couldn't get 5 experimentals in a row anywhere. The location of the experimental is less important to me than the difficulty of the experimental. If the computer preselects experimentals without knowing how well you will score, then presumably you could be doing really well and see a basic question... or be tanking and randomly see a super hard question. This seems strange. It would make more sense if the computer tailored the experimentals to your level as you take the test.

DanaJ wrote:

zuleron wrote: 5) experimental questions: I'm still not sure. From what has been said I think you couldn't get 5 experimentals in a row anywhere. The location of the experimental is less important to me than the difficulty of the experimental. If the computer preselects experimentals without knowing how well you will score, then presumably you could be doing really well and see a basic question... or be tanking and randomly see a super hard question. This seems strange. It would make more sense if the computer tailored the experimentals to your level as you take the test.

And I'd agree with that last thing for one simple reason: on test day, after the first few 10-15 decent questions, a wave of really tricky questions hit. But in between the hard questions I saw a couple of pretty basic ones, concerning powers. I remember I was a bit confused by those, but I went on without too much thought for it.

zuleron wrote: I have learned a great deal from the various threads on the algorithm and strategies.

1) The bisggest no-no is strings of incorrect answers.

2) The first 10 questions are only more important to the extent that doing poorly in the first 10 leaves you with fewer questions to make up for the poor start. So you might want to spend more time on the 1st 10.

3) On the other hand, spending too much time on the 1st 10 increases your risk of gettting strings of incorrect answers later on in the test as you get pressed for time. So you probably don't want to spend too much time on the 1st 10. And if you do you are going to have to guess on some questions.

4) If you are going to guess on 5 questions (say because you spent more time on the 1st 10), it is better that they be Q12, 18, 23, 29, 35 than Q33-37 because you avoid getting strings of wrong answers.

5) experimental questions: I'm still not sure. From what has been said I think you couldn't get 5 experimentals in a row anywhere. The location of the experimental is less important to me than the difficulty of the experimental. If the computer preselects experimentals without knowing how well you will score, then presumably you could be doing really well and see a basic question... or be tanking and randomly see a super hard question. This seems strange. It would make more sense if the computer tailored the experimentals to your level as you take the test.