Forbrukslån Billigst | Pikavippi Vertailu 2019 | Beste kredittkort | Strømleverandør | Beste Hyttealarm | Billigste Forsikringsselskap 2019 | Beste Mobilabonnement | Pris på Betong

Who’s Afraid of the GMAT Algorithm? Not 700-Scorers.

700-Scorers understand learn to think about the GMAT Algorithm in a way that suits them, not one that suits the GMAT itself.


There. I said it. Why would we consciously give away our autonomy to a computer? Chew on this:

“You can’t tell if the machine has gotten smarter or if you’ve just lowered your own standards of intelligence to such a degree that the machine seems smart.” –Jaron Lanier

It seems a bit bizarre to me that we would just assume that the GMAT Algorithm—“the machine” in Lanier’s quote–is infinitely intelligent.

After all, the GMAT was written by a bunch of people in a damp basement in Iowa City, Iowa. A lot of them were academics at one point… and then they weren’t.

That’s why they’re writing the GMAT.

So how are these people so much smarter than the rest of us that they create unanswerable questions?

Hint: they’re not.

Now let me tell you what this is not: a list of purely speculative assertions about how the GMAT Algorithm calculates the score level of every question. This is only worth considering in the rarest of cases and you’re really better off focusing on something that can actually move the needle.

The Algorithm is monolithic. You need to learn to live with it. If you’re trying to game the system, this will only end up in tears. Plus–why don’t you put that effort into studying? That’s an easier way to move the needle.

All that said, it’s actually doing ourselves a disservice ever to think that the test can create a scenario that we can’t solve.

Rather, we need to learn how to think about the GMAT algorithm in a healthier way.

First, bear in mind that most GMAT questions, given enough time and perhaps a clearer explanation of “what the test wants,” are eminently solvable.

Imagine this thought experiment:

Let’s say you’re confident in your knowledge of the topics studied on the GMAT.

If you were given one GMAT question each day (IR plus Quant and Verbal) for 90 days—with unlimited time to answer–how many would you get correct?

It would be pretty much all of them, wouldn’t it? Or at least enough to get a very high score.

So it’s really not about the questions not being solvable, is it? It’s more about the questions being solvable within the time frame.

If you knew the proper tricks and techniques, of course, the questions would be solvable within the time frame.

It also follows that, given a sufficient amount of time, energy, and study, that almost anyone could get a great GMAT score. However, that time could be anywhere from a couple of weeks to several years.

The pertinent question to ask yourself is whether you have enough time to do so before you submit your application!

Actionable advice to mentally conquer the Bastard Algorithm after the break…


Let’s think a bit harder–speculatively, of course, about what the Bastard Algorithm really looks for to calculate your score.

As a note, if you expect that this will help you “game the computer,” just stop reading right now and unsubscribe from my e-mail list. Srsly.

That said, it is important to have some concept of how the scoring works so that you’re not running around assuming that it’s all about “percent correct/incorrect.”

The one useful thing to remember about the algorithm is that the percentage of correct and incorrect questions doesn’t really matter that much—at least not compared to the difficulty of the questions you’re thrown.

Put these two pieces of information together and what you’re looking at is that the GMAT Algorithm makes the exam more like running a gauntlet than it is about “getting 90% of the questions correct.”

In short: you have to perform well under pressure, but it’s assumed that there will be a few blips or errors.

Everyone makes mistakes—in fact, you can miss a handful of questions and still get a Q51, PROVIDED THE GMAT QUESTIONS YOU ARE ANSWERING ARE DIFFICULT ENOUGH.

In fact, assuming that the questions aren’t conquerable will aim your thinking in the wrong way. You need the mental agility to try different techniques to answer the question.

For example, in Quant:

–sometimes you won’t see the clever trick, but brute-force calculation will get the answer well within the time frame!

–sometimes you want to use brute force, but the answer is staring you in the face—without even needing to calculate!

In Verbal:

–sometimes the answer is the obvious one!

–many times the “clever answer” is actually a red herring!

(This is a very important reason to spend time wrestling with tough questions and constantly attempting to improve your solutions!)

Put another way, you’re restricting this ability when you assume that you “can’t do it.”

Now let’s address the common concern that “the GMAT is just smarter than I am.”

Oh, dear.

As long as you tell yourself that the test is smarter than you, then you’re doing yourself a disservice—to paraphrase Jaron Lanier see the quote above, you are diminishing yourself by giving an unfair advantage to this arbitrary, computer-based, eminently scientific and technical exam.

However, the choice to let it be soul-destroying is yours and yours alone.

This exam is ridiculous. This exam is conquerable. It is not a living, breathing thing.

It can adapt, within reason, based on the information that you give it. There are a lot of those—enough that trying to figure them out and “beat the computer” is a thorough waste of your time and energy.

HOWEVER, the GMAT is not “smart” because it is not alive. It does not think; it reacts to fixed inputs that you provide.

The GMAT is not above you, yet to think so diminishes your capacity to conquer the GMAT.

Another important thing to bear in mind is that the GMAT Algorithm is designed to make the test consistent.

That is, a 630 on one day should be worth the same as a 630 three months from now. The only way that this is possible is when the questions are written in a consistent way.

It follows that there is an underlying structure to the questions, and that it is learnable. After all, the writers of the questions need to make them consistent, so they must learn a pattern to write them.

If the writers of the questions can learn the patterns, you certainly can as well.

Let’s break this down into a few useful tenets to study by:

–The GMAT is a total pain in the ass

–The exam is difficult to do under timed conditions.

–The GMAT algorithm is not alive.  The idea of it being “smart” is ridiculous on the surface.

–Giving the GMAT more credit than necessary will restrict your ability to answer questions effectively.

–One can solve any GMAT question given enough time and a clear enough understanding of its mechanisms.

–GMAT questions are scientific and consistent; it is possible to learn their underlying structures.

Remember: don’t give the GMAT too much credit. That’s exactly what it wants you to do.

A final note:

In my less-than-humble opinion–and this will open me to all sorts of criticism from trainspotters on Reddit, I’m sure–the “Zen approach” of simply get more questions correct and fewer questions incorrect is the best way to deal with the GMAT.

Get better at answering the questions and everything else will follow.

Need help doing that? Click here for a free 20-Minute Consultation.