As an independent GMAT tutor with an unorthodox streak, I tend to be the GMAT resource of last resort for many test-takers. The majority of my NYC students have already taken at least one GMAT prep course, and most have taken the test at least once (one of my students took the test seven times before she called me, and cracked 700 for the first time ever this weekend… yay!). I love the challenge of trying to help students beat the GMAT after they’ve already exhausted every other resource, and I’m always honored when somebody puts their faith in me after months–or years–of falling short of their score goal.

Because the prospective students who contact me tend to have so much GMAT experience already, I usually ask the same question: what do you think is holding your GMAT score back? As you might guess, I get a huge variety of answers, but there’s one common thread: nearly everybody blames combinatorics (combinations and permutations), probability, rates, and/or overlapping sets for at least part of their struggles on the quant section. Let’s call those four topics the Evil Four.

I’m not going to tell you that rate questions aren’t important. I’m not going to tell you that combinatorics, probability, and overlapping sets aren’t important. But I will tell you that the Evil Four are never the difference between, say, a 650 and a 700 on the GMAT.

Here’s the thing: when you really look hard at the GMAT, none of the Evil Four seem to appear all that often. It’s possible that you’ll see two rate problems, but it’s more likely that you’ll see only one. It’s possible that you’ll see a combined total of three probability and combinatorics questions, but it’s also possible that you’ll see exactly none–even if you’re doing well on the exam. And overlapping sets questions aren’t so common either: very few test-takers claim to see more than one or two of them on the GMAT.

Think of it this way: you might only see a grand total of four or five questions from the Evil Four on the actual GMAT. That’s only a tiny proportion of the 78 questions on the test, yet I routinely speak with people who seem to spend as much as 50% of their study time on these topics. But if the Evil Four appear in less than 10% of GMAT test questions, why is everybody so stressed out about them?

Honestly, I think that test-takers tend to focus on these questions because they’re so damned memorable compared with the rest of the questions on the GMAT. You won’t necessarily remember the algebra questions (too boring), the arithmetic questions (too pedestrian), or the number properties questions (too abstract); the crazy rate questions with two trains headed toward each other seem to stick in our heads much more easily.

But the unglamorous reality is that algebra (including word problems and functions), arithmetic (including word problems, percents, exponents, roots, and estimation), and number properties (including all sorts of fun with factors, multiples, and remainders) questions are the core of your GMAT quant score. If your accuracy on these three topics is less than stellar, your GMAT score will also be less than stellar.

Again, I’m not saying that rates, combinatorics, probability, and overlapping sets aren’t worth studying. I just think that these questions cause a disproportionate amount of anxiety, and I think it’s crucial to keep them in perspective. If you’re trying to raise your GMAT score from, say, a 47 to a 51, you probably need to study the living hell out of everything. But if your goals and starting score are more modest–or if you need to substantially improve your composite score–don’t overemphasize this stuff.

At the very least, make sure that you don’t focus on the Evil Four at the expense of more important topics. If you need to make a big improvement in your GMAT quant score, spend the bulk of your time developing your accuracy on algebra, arithmetic, and number properties. Since those three topics appear in more than two-thirds of GMAT quant questions, it would be wise to keep them at the front of your mind as you create your GMAT study plan.