REF: Thursdays w/ Ron 2/18/10 "GMAT CAT Adaptive Algorithm"

[as2764]

Hello:

I recently watched a video by Ron Purewal from "Thursday with Ron" series, in which the differences between a 'Paper Test' and the 'GMAT CAT' are discussed. In particular, I had the following questions re the timing strategy suggested:

a. Ron explained that because there isn't a 'pool of surplus time' on the GMAT, after a certain amount of time one should consistently give up and move on to the next problem. BUT, since we know that 'coarse adjustment' of the difficulty level is done in the initial few problems (courtesy: the graph), shouldn't we be spending relatively more time on them?

b. If the answer to a. is YES, does it mean that the problems towards the end are less important and are merely 'fine tuning' the score because the 'hill' was already created by the answers in the beginning and that it's ok to err or even miss on the tail end of the section?

c. If the answer to b. is NO, based on another fact that running out of time on answering the questions at the tail end of the section carries more penalty than answering them wrong, should we guess on the harder problems mid-way, for instance, Q20-30 on Quant or Q25-35 on Verbal?

Thank you in advance!
--
Ashish

[StaceyKoprince]

(A) No. The "coarse adjustment" is only one factor. There are others that will penalize you (heavily) at the end for running out of time or for having unexpected questions wrong (questions lower than your current scoring level) or having multiple wrong answers in a row (another consequence of running out of time). The higher your testing level, the heavier the penalty at the end for the scenarios I listed above - in fact, for a 780 level tester, if she got the first 5 questions right (regardless of time spent), she would be permitted only 1 question wrong at the end of the test before a penalty would kick in that would actually lower her overall score. The takeaway there is that she has to get the first 5 right while spending almost no extra time. (The penalty isn't as severe for, say, 500-level testers... but you're going for a higher score than that, aren't you? :)

If you want to read more about how this works, read the scoring section in our e-book The GMAT Uncovered. Just look in your student account.

(B) moot.

(C) As a general rule, you do your best on every question within the allotted time and, if it's too hard for you (as some will be), then you guess and move on. You can reasonably spend about 30 seconds longer than the expected average on SOME problems (because you will spend less than the average on others), but the only good reason to spend *extra* time is when you know exactly what you're doing, but the problem takes longer to read or has an extra few calculation steps or something. When you're thinking, "Well, I'm not really sure, but if I spent a little more time, I'm sure I could..." That's when you stop. Spend any remaining time you have making an educated guess, if you can, and move on before you lose time on that problem.

If you find yourself behind on time, guess randomly on the next "oh, that looks really hard" problem you see (not just the very next problem - if you have to guess, you might as well get to choose which problems!). If you need to do it again in order to catch up, do it again. Don't wait and let the time lag pile up at the end.

[as2764]

The higher your testing level, the heavier the penalty at the end for the scenarios I listed above - in fact, for a 780 level tester, if she got the first 5 questions right (regardless of time spent), she would be permitted only 1 question wrong at the end of the test before a penalty would kick in that would actually lower her overall score. The takeaway there is that she has to get the first 5 right while spending almost no extra time. (The penalty isn't as severe for, say, 500-level testers... but you're going for a higher score than that, aren't you? :)

When you're thinking, "Well, I'm not really sure, but if I spent a little more time, I'm sure I could..." That's when you step. Spend any remaining time you have making an educated guess, if you can, and move on before you lose time on that problem.

If you find yourself behind on time, guess randomly on the next "oh, that looks really hard" problem you see (not just the very next problem - if you have to guess, you might as well get to choose which problems!). If you need to do it again in order to catch up, do it again. Don't wait and let the time lag pile up at the end.

Thanks Stacey. although it seems logical, i didn't realize that penalties are larger with higher score levels. Certainly 700+ :) And with that being said, 750+ is going to be a tough nut to crack!

In conclusion, the way i understand the test is that as you answer correctly, the difficulty level keeps on increasing until your thinking either exceeds its capacity or exhausts the mentally allotted 2.30-3 mins. And in case you do exceed the allotted time, make it up by guessing later but maintain the average time per question.

[StaceyKoprince]

Yes, maintain the average, and be able to STOP yourself if you are exceeding about 2.5m or so, NEVER more than 3m on any single question. The longer you take, the more likely you are to get it wrong anyway (because you don't really know what you're doing) and the more likely you are to get multiple additional questions (possibly in a row) wrong later because you're now behind on time.

You're playing tennis. You don't expect to win every single point, right? Sometimes, you're just going to have to say, "Nice shot!" and get ready for the next point. Same thing here. :)

Also, I just noticed that last time I wrote:
When you're thinking, "Well, I'm not really sure, but if I spent a little more time, I'm sure I could..." That's when you step."

That last word should be "stop." lol :)

[as2764]

It's indeed like Tennis! I was never able to reach the Verbal section past the 24th Q and just managed to complete 34 Qs on Quant, but after learning time-mgmt I have consistently finished all my tests and sections on time.

And I've finally started to score in the 700 range!

Also, I just noticed that last time I wrote:
When you're thinking, "Well, I'm not really sure, but if I spent a little more time, I'm sure I could..." That's when you step."

That last word should be "stop." lol :)

I still got your message as 'step out/away' from the problem :-)

--
Ashish

[StaceyKoprince]

Great! Keep up the good work!