Scoring logic in GMAT

This topic has expert replies
Junior | Next Rank: 30 Posts
Posts: 21
Joined: Tue Feb 12, 2008 12:50 pm

Scoring logic in GMAT

by getk » Sat Oct 11, 2008 10:11 am
Hi friends,

I had a doubt about the GMAT scoring logic/pattern. Many of the sites/tests i have tried give enormous importance to initial questions, after which the rate of percentile remains stagnant.
But i've seen Stacey mentioning about attending a conference and saying that the old GMAT strategy of pacing at the end wont work anymore, as wrong answers at the end will lower the score than what we expect.

So is it true to say questions are having equal importance, and so can devote equal time at the end till end?

regards

GMAT Instructor
Posts: 1223
Joined: Thu May 01, 2008 3:29 pm
Location: Los Angeles, CA
Thanked: 185 times
Followed by:15 members

by VP_Jim » Sat Oct 11, 2008 10:28 am
I think the general consensus - and what I tell my students - is that, yes, the earlier questions are still more important because you need to start getting the harder questions earlier on (and you do this by getting easier and medium level questions right).

Now, that comes with a couple caveats:

1. You still need to leave time to do a decent job on the last few questions. A big string of guesses in a row will probably hurt you, so make sure you have a couple minutes per question towards the end. Now, that doesn't mean you can't guess on one or two questions towards the end if you're running out of time - just don't do them all in a row.

2. I think Stacey mentioned this one before: to a certain extent, you have to be a pretty good GMAT test taker to pace yourself the way that most recommend (e.g., 3 minutes for first 10, 2.5 minutes for next 10, etc.), and if you're good enough to pace yourself properly, you're probably going to score well, anyway.

I tend not to think of pacing as strictly "time per question". Rather, I think of it as how to approach questions. Earlier on, I'm more willing to spend an extra 30 seconds or a minute double checking my work. In the last few questions, if I solve a problem and get an answer I think is right, I'll usually just pick it and keep going. Finally, I'm less likely to take shortcuts on earlier questions, whereas I will usually try to save time by using tricks on later questions.

Pacing is a very individualized thing. You also need to consider how much time you usually have left. I, for example, usually cut it close on math so I'm more concerned with pacing on that section vs. verbal, where I always have 20 minutes left (so I spend as long as I want on every question).

That probably made things even more confusing - that's the GMAT for you!
Jim S. | GMAT Instructor | Veritas Prep

Junior | Next Rank: 30 Posts
Posts: 21
Joined: Tue Feb 12, 2008 12:50 pm

by getk » Sat Oct 11, 2008 10:39 am
thanks Jim for prompt reply. As you said, its confusing :)
it may be secret logic with GMAC.

I also feel ,that string of wrong answers are what GMAT looks into to judge whether u made random guesses or not.

I'm trying it hard for my Verbal part to pace through (rather different from many others), which always ends me up in random guessing at the end :(

Junior | Next Rank: 30 Posts
Posts: 10
Joined: Sun Dec 14, 2008 12:04 pm

by realanoop » Thu Jan 08, 2009 3:10 pm
there is a very simple logic to this...

try taking a random book and decide to flip to a random page.

you will realise that you first few flips will be really large as you try to estimate the total pages and the approximate section where the page you want is.

And your last few flips will be smaller and more accurate taking you to the exact page that you want.

there isn't a faster way to find something in an sorted list. (tech speak)

so this is the exact same way in which your gmat scores HAVE to be calculated.

elementary my dear watson...

GMAT/MBA Expert

User avatar
GMAT Instructor
Posts: 2621
Joined: Mon Jun 02, 2008 3:17 am
Location: Montreal
Thanked: 1090 times
Followed by:355 members
GMAT Score:780

by Ian Stewart » Fri Jan 09, 2009 5:50 am
No, that's not how it works. It's a Bayesian probability estimate, not a search-based algorithm - the test asks, based on how you've answered questions so far, what is the most likely estimate of your ability? Each question is weighted equally in the calculation; the only importance to the sequence of answers is which difficulty level of questions you actually see. Of course, to get a great score, you have to answer some difficult questions, and to see difficult questions, you have to get a lot of questions right at some point. But the difficulty level of the questions you see can swing quite a lot, even later in the test.
For online GMAT math tutoring, or to buy my higher-level Quant books and problem sets, contact me at ianstewartgmat at gmail.com

ianstewartgmat.com