When to guess???

[eric.david.grossman]

Hello,

I have been trying to hone my GMAT quant pacing and one of the things I have come to realize is that I am going to have to guess on some questions. Answering a question correctly in 5 minutes seems to hurt more than it helps and I never finish a quant section when i do this. With that in mind, does anyone know if it is better to guess on really tough questions in the middle of the test as opposed to the end? I read elsewhere that GMAC is less likely to put the experimental questions at the end of the quant section because they know a lot of people will be guessing on the last 2-3 questions for pacing reasons hence you are more likely to be guessing on an experimental question if you guess early/middle of the section. Also, given how the MGMAT algorithm seems to work it would appear that if you know you are running out of time by 2/3rds of the way through the test it is better to guess on a few tough questions, save the time, then try to answer the last 10 questions or so correctly and bring your percentile up vs getting several in a row wrong.

Thoughts?

[StaceyKoprince]

This is a great question. You're absolutely right that spending 4 to 5 minutes to get a question right just hurts you, even if you do get it right. (And, frankly, when you spend that long, the odds are not so great - spending that long is mostly just an indication that you don't really know how to do the problem because you don't know the regular solution. :)

Also, as you've realized, you are going to have to guess sometimes. Everybody has to guess some of the time - that's just how the test works.

Your task is to recognize which questions are too hard for you throughout the test - and that's where you should be guessing. It doesn't matter exactly where they are - just whenever you hit something that's too hard for you, recognize it, try to make an educated guess, move on.

If, for example, we're talking about quant questions (which you said is your "timing problem area"), we're supposed to average 2m per question. That means we don't want to go much beyond about 2m30s on any single question. That also means that, by the halfway mark, we should have a pretty good idea of what we're doing in order to answer the question within the expected timeframe.

So, if around 1m or so, you're still lost... you know you need to guess on this question. The good news? You still have about a minute left, so you have time to make an educated guess!

(Note for others reading this: we're addressing quant here because that's this student's problem area. Verbal works a bit differently, though the overall principle still applies: recognize the "too hard for you" Qs before you use up all your time and also guess before you go over time.)

I think you would benefit from a timing exercise: learning about how long one minute is without looking at a watch or stopwatch. If you don't have one already, buy yourself a stopwatch with lap timing capability. When you go to do a set of problems, start the stopwatch but turn it over so you can't see the time. Every time you think one minute has gone by, push the lap button. When you're done, see how good you were - and whether you tend to over or underestimate. Get yourself to the point where you're within 15 seconds either way on a regular basis (that is, you can generally predict between 45 sec and 1min 15 sec). Note: at the same time that you are using the stopwatch to time this "1-minute" thing, also use the OG Stopwatch (in your student center) to track the total time spent on each question.

Now, how do you use that when doing problems? If you're not on track by one minute, make an educated guess** and move on. (The general idea is that if you're not on track by the halfway mark, you're unlikely to figure out what's holding you back AND have time to do the whole problem in the 1 min you have left.)

** This also requires you to know HOW to make an educated guess depending upon the type of problem and the content being tested. So that's something else to add to your study: how to make educated guesses on different kinds of problems.

Here are two articles that can help you develop educated guessing skills (one for quant, one for verbal).

http://www.beatthegmat.com/mba/2010/07/ ... s-on-quant
http://www.beatthegmat.com/mba/2010/08/ ... -on-verbal

The other thing that I think helps is just a certain mindset:

Think of this as a tennis match, not a test. You're going to win some points and the other guy is going to win some points; you're not going to win them all, right? Your goal is to put yourself into position to win the LAST point. Translated, that means you have to put yourself in position to answer the last question - you have to have time to address it. Otherwise, you've lost the last point, and by extension the match. When the other guy hits a winner, don't go running after it so fast that you hit the fence and injure yourself, thereby hurting your chances on the later points. (Translation: don't go way over when the problem is too hard.)

[eric.david.grossman]

Thanks Stacey, that was very helpful! I think one part of the question which remains unanswered is the part about actually when to guess (i.e. if you know you are running out of time is it better to guess in the middle or the end?). It appears from your comment that it is better to have time to do the last couple of questions and save time on middle questions. Does the test weight one more than the other?

Also, I have heard that experimental questions tend to be in the middle as opposed to the end because GMAC needs to see actual responses to these questions hence does not put them at the end where more people are likely to be out of time and guessing. Wondering if you or anyone else has heard of this and thinks it has any merit?

Many thanks!

-Eric

[StaceyKoprince]

Officially, experimental questions are randomly distributed. Unofficially, while you could have some experimentals towards the end, they do have measures in place to ensure that, say, the last 5 questions are not all experimental for one person. That tester might gain an unfair advantage.

Ideally, you guess on the hardest questions as you see them through the section, so that you never have to guess solely because you are behind on time.

If you do get behind and realize that you have to guess to catch up, do one of two things:
1) ideally, guess (almost) immediately and randomly on the next really hard one that you see - as soon as you realize that it's really hard. keep doing that on the reall hard ones till you catch up
2) if you have limited time to catch up, or if you find yourself panicking a bit because you know you're behind, guess immediately and randomly on the very next question. Try the one after, then guess again if you still need to (though, ideally, switch over to method 1 as soon as you can - this method isn't better as a guessing method, but if you're panicking, that's even less useful :)

When you combine all of the factors going into the algorithm, no questions are weighted significantly heavier or ligher than the others (with the exception of the experimentals of course, which are weighted at zero!).

[eric.david.grossman]

Thank you so much Stacey, that is really really helpful!

[StaceyKoprince]

you're very welcome! :)