GMAT-Test taking strategy

Love this topic! A few things I'd add:

1) If you can get a question right, get it right

Take a few seconds to double check your work if you know that you're susceptible to silly mistakes. If you know you'll finish but fear that the calculations will take even 30 seconds more than they should, finish them and get the problem right. The reason? On a computer-adaptive test, wrong answers to questions you should get right hurt you more than right answers to questions that really should be above your head help you. Or more simply "easy wrong outweighs hard right". Don't rush through problems just to "save time for harder problems". If you get those moderate problems wrong, you may never see the hard problems; and even if you get them right, there's no guarantee that extra time on a hard problem will mean you get it right.

2) The first 10 questions are very important, but not the way you might think

The first 10 questions do have a little extra importance over the last 10 because your results on those first 10 stay with the DELIVERY system for the rest of that section. Long story short, the adaptive function of the CAT system really serves two functions:

-Scoring (for which all questions count the same)
-Delivery (and here the system uses your results on all previous questions to determine which question to serve you next)

For scoring, question 36 and question 2 count exactly the same. But for delivery, question 2's results are in the system a lot longer - every subsequent question the system delivers you will be impacted to some degree by how you did on question #2.

Now, people overstate the impact of that delivery function and try to game the system by doing everything they can to get the first 10 right. And from that point the scoring system will figure you out...if you get the first 10 right and struggle from there, the delivery system will take a while to catch up but it will catch up, and when you end up with a lot of guesses and wrong answers the scoring system will have a lot of reasons to tear you down.

The actual key relates back to my first point - don't miss easy questions, especially early in the test. If you get 7 of the first 10 right, you're well on pace for a great score. You don't need to get all 10 - your right answers will mean that the delivery system is already throwing you very hard questions. The key is to not miss, for example, 7 of the first 10. If you do, the delivery system is showing you easier questions and you run the risk of running out of opportunities to "earn" and succeed on harder questions. Especially because you're human - you will make a dumb mistake on at least one problem you should get right. If you make those mistakes early, your margin for error is very thin later in the section. So the takeaway - be patient early (double check your work, don't stress the clock too much) and set yourself up for success. But DON'T kill yourself trying to get them all right. You won't.

3) Don't be afraid to guess / Learn to let go

People often rush through "easy" questions to save time for harder questions, when the above advice really tells you to do the opposite. But obviously the clock will be factor. How do you save time to be patient on questions you should get right?

-Know when to let go of a question. If after a minute you just don't see a path to the finish line, it's ok to blindly guess. There's a great story that one of the "godfathers" of the GMAT told me once (he's seen/known as the main guy who oversees the test for GMAC). He was taking the test for R&D purposes and as a math PhD he really wanted a perfect score. One tough geometry question was giving him trouble and he spent maybe 6-7 minutes on it trying to get it right, laughing to himself that he knows it's horrifically stupid to spend that much time on any one question. As the main man at GMAC he was able to check the "back end" results the next day...when it turned out that that question was one of the unscored experimentals - it didn't even count! Even the guy in charge of the test has to guess sometimes; it's ok for you to do it, too.

-If you're going to struggle on time, it's better to blindly guess once or twice (to earn your 2-4 minutes back) than to rush to save 10-15 seconds on each question. Silly mistakes will really hurt your score, but since you *know* you'll get at least 7-8 questions wrong (*everyone* does) you might as well guess once or twice when a question just looks like it's not going anywhere. Take your 20% (or hopefully better if you can give it a quick educated shot) chance so that your must-haves are 100% chances and not 90.

4) Take practice tests so that you know what this all feels like

Like Hamid said above, there's no substitute for practice. If you've done several practice tests you can get a good internal clock going to let you know when you're moving too slowly or too quickly, and you can play around with certain milestones (if you've used 22 minutes after the first 10, do you still feel comfortable throughout?) to write down or keep in mind. There's no substitute for practice.

Hi,

Following from my last post and upon some significant deliberations on tackling the "timing" aspect of the test, I have come up with the following strategy. I need your advice on the same:

Quant:Currently on 47/48, need to hit 49/50

I would be extra cautious in the first 20 questions and take approx 40 min to complete. Therefore, the time left would be 35 min for the next 17 questions ie. approx 2 min per question. The MAJOR strategy is to stick to the average time(depending upon the question type). Also knowing my areas of weakness(probability for that matter), I would not spend any extra time(if I am stuck). However, will spend an extra 15/20 sec on a question of strength(if needed).

Verbal:Currently @ 33/34, need to hit 39/40

As in quant, I would be extra cautious on first 20, spending approx 40 min. For the next 21 questions, I should be having 17 min, which is a negative time scenario already(5/6 questions behind). I would be adopting the same strategy as in quant. However, to cover up for the 5/6 questions lag, I can do the following:

A..Throwaway some questions from my weak areas by not even looking at the questions

OR

B..Give approx 30 sec to try to solve a question from my weak area and then take an educated guess

Also I have observed(from my MGMAT CAT scores) that the steepest drop in %ile estimate happens if I am missing a RC question. Is this true? In that case, should I be extra careful in dealing in questions from the weak area of RC(inference, detail)?

Please provide your valuable inputs on the above. I have about 10 days to go and my last GMAT Prep score was 700(guess I had seen some questions earlier). I am hitting 660 ish on MGMAT CATs. I need to take it to 720. Is it possible by getting the timing aspect correct?


Regards

Argha