Where GMAT Focus inequalities actually break: 6 error families most candidates repeat

Inequalities are a small but unusually punishing slice of the GMAT Focus Quant section. In a 31-question, computer-adaptive test where every item is scored and the percentile curve steepens sharply above the 78–83 band, the candidate who treats inequalities as a quick warm-up is the candidate who leaves five to nine points on the table. The maths itself is elementary — direction of a sign, behaviour at a critical point, the difference between a closed and an open endpoint. What makes the topic treacherous is that the GMAT Focus writes inequality stems in a way that rewards a specific reading habit, and most candidates do not have that habit. They read the stem for the shape of the equation and then reach for a number from the answer choices, which is exactly the move that turns a 60-second question into a 3-minute trap. This article walks through the six error families I see most often in candidate work, with full worked stems, the timing budgets a serious scorer should set, and a short tactical block for each pattern. The goal is not to re-teach school algebra; the goal is to give you the diagnostic grammar that separates a careless miss from a structural one, so your prep time on this topic actually translates into scaled-score movement on test day.

The reading habit that decides inequality outcomes before the maths begins

The single most expensive mistake on a GMAT Focus inequality is reading the stem the way you would read an equation stem. Equation stems usually reward a fast scan: you look for the variable, the operator, the constraint, and the question being asked. Inequality stems look identical on the surface, but they carry two extra layers of information that equations do not. First, the direction of the inequality sign encodes a region of the number line, not a single point, and the candidate who treats the stem as 'solve for x = …' is already working the wrong problem. Second, almost every GMAT Focus inequality stem contains at least one subtle quantifier — 'for all', 'for some', 'for no', 'must be', 'could be' — and the candidate who skips that quantifier is solving a different problem from the one the test is asking. The reading habit, in other words, has to be slower on the first pass, not faster. Aim for about 25 seconds of pure reading before you write a single character on your notepad. Read the stem twice. The first pass is for the relationship: is this a linear inequality, a quadratic one, a fractional one, an absolute-value one, or a compound chain? The second pass is for the quantifier and the question form. Only then do you start moving symbols.

Notice what that 25-second budget does. It does not slow you down on a per-minute basis; it actually speeds you up, because you arrive at the right problem in one pass instead of three. The candidate who rushes the read is the candidate who solves the inequality correctly, picks the right endpoint, and then answers the wrong question — answering 'for all x' as if it were 'for some x', or answering 'must be true' as if it were 'could be true'. On a 31-question section where adaptive scoring makes the second module substantially harder, that single misread compounds across the rest of the section. The test is not just checking whether you can solve the inequality; it is checking whether you can solve the right inequality and report back on the right property of the solution set.

Common pitfalls and how to avoid them

• Treating 'must be' and 'could be' as interchangeable. The first asks for an identity of the solution set; the second asks for a possible member. Different question, different answer, different work.
• Dropping the sign when dividing by a negative. It is the oldest inequality error and it still costs points. Train yourself to mark the sign flip with a small arrow on the notepad before moving terms.
• Forgetting the boundary. A strict inequality has an open endpoint; a non-strict one has a closed one. The boundary point is the answer on roughly one in five inequality stems, and skipping it is a free point given back.
• Working on the answer choices before the stem is fully read. The choice-stalking move burns 90 seconds and almost always produces a wrong answer on inequality items.

Linear inequality chains and the silent sign-flip trap

Linear chains look harmless, and that is exactly why they are dangerous. A typical GMAT Focus stem in this family runs: 'If 3 − 2x < 7 and 4x + 1 ≥ 9, which of the following could be the value of x?' The candidate reads the first inequality, divides by 2, flips the sign, gets x > −2, and then reads the second inequality, gets x ≥ 2, and answers the question as if the intersection of those two regions were the whole story. The intersection is x ≥ 2, which is correct. The candidate then checks the answer choices and picks 2 because it is the cleanest number, only to discover that 2 is not listed. The trap is not the maths; the trap is that the test has chosen four of its five options from inside the solution set and one from outside it, and the candidate's eye is drawn to the inside value that looks 'most' right. The fix is procedural: state the solution set in interval notation before you look at the choices, write it on the notepad as [2, ∞), and then ask which choice is consistent with that set. Candidates who follow that step are noticeably more accurate on chain stems than candidates who scan first and verify second.

The sign-flip trap lives a layer below the chain. It is the candidate who turns 3 − 2x < 7 into 2x < 4, then divides by 2 and forgets to flip, getting x < 2 instead of x > −2. The two answers look superficially similar, but only one of them intersects correctly with the second inequality. On test day, the candidate will often solve both inequalities correctly, intersect them, and end up with a region that does not match any of the choices — at which point they panic, change a sign somewhere else, and walk into a second error. The mitigation is purely mechanical: every time you divide or multiply an inequality by a negative number, write a tiny downward arrow next to the sign on the notepad. After two or three practice sessions, the arrow becomes automatic. It looks like a small thing, but in a section where you have roughly two minutes per question on average, the difference between an error-free chain and a sign-flip chain is the difference between a 79 and an 83 on Quant.

Worked example: chain with a 'could be' quantifier

Stem: 'If −4 < 2 − 3x < 10, which of the following could be the value of x?' The chain rewrites as two inequalities: 2 − 3x > −4 and 2 − 3x < 10. The first becomes 3x < 6, so x < 2. The second becomes −3x < 8, so x > −8/3. The solution set is the open interval (−8/3, 2). The question form is 'could be', so any number inside that interval is a valid answer. A candidate who treats 'could be' as 'must be' will hunt for the midpoint, which is the wrong question. A candidate who works the chain and then reads the question will pick the only option inside the open interval and move on. The whole item, start to finish, should sit at about 75 seconds for a candidate who has trained the chain-reading habit, and at about 130 seconds for a candidate who has not.

Quadratic inequalities: reading the parabola before you solve

Quadratic inequalities are the inequality family where reading the stem pays the highest dividend. The standard error pattern is the candidate who jumps straight to factoring. They factor the quadratic, find the two roots, write 'x is between the roots' without checking the sign of the leading coefficient, and pick an answer from the wrong region. The fix is to spend 15 extra seconds sketching the parabola on the notepad. A rough sketch — a U shape or an inverted U shape, with the two roots marked on the x-axis — turns the problem from an algebraic exercise into a visual one. Once the parabola is on the page, the solution set is obvious: the region where the curve is above the x-axis for a 'greater than' stem, or below it for a 'less than' stem. The candidate who has sketched the curve does not have to remember whether the inequality sign is 'outside' or 'inside' the roots, because the curve is right there in front of them.

There is a second quadratic trap that shows up on the GMAT Focus with enough regularity to be worth naming. It is the case where one of the roots is a double root — the quadratic is a perfect square, and the inequality reads (x − 3)² < 0. There is no x value that satisfies a strict less-than against a squared term, so the answer is 'no solution'. Candidates who reflexively write 'x = 3' are solving the equation, not the inequality. The same logic applies to (x − 3)² ≤ 0, where the only solution is x = 3, not an interval. The double-root case is rare, but it is a free point for the candidate who pauses for the sketch and notices that the parabola touches the axis without crossing it.

Worked example: quadratic with a 'must be true' quantifier

Stem: 'If (x − 1)(x + 4) < 0, which of the following must be true?' The parabola opens upward, the roots are −4 and 1, and the region where the curve is below the x-axis is the open interval (−4, 1). The candidate who answers 'x < 1' is right only if the question is asking for a necessary condition; the candidate who answers '−3 < x' is right only if the option is inside the interval. The right move is to identify the interval first, then read the question form, and only then look at the choices. Roughly one in three quadratic stems in the GMAT Focus Quant section uses a 'must be' or 'must not be' quantifier, and that quantifier is doing real work in the answer. The candidate who treats the stem as 'find the interval' without checking the quantifier is solving a different problem from the one being asked.

Fractional inequalities: the domain check that 80% of candidates skip

Fractional inequalities are the inequality family with the highest error rate per question on the GMAT Focus, and the error is almost always a missing domain check. A typical stem reads: 'For what values of x is (x − 2)/(x + 3) > 0?' The candidate crosses the two factors, gets the interval (2, ∞), and forgets that x = −3 makes the denominator zero. The correct solution set is (−∞, −3) union (2, ∞). The two pieces are not connected, and the candidate who treats them as a single interval will pick a wrong answer. The fix is procedural: every time a fractional inequality appears on the notepad, write a small 'D ≠' line that lists the values of x that make any denominator zero, and exclude them from the solution set before you look at the answer choices. The check takes 10 seconds and removes an entire class of error.

There is a second fractional trap that lives in the sign chart. The candidate is told that (x − a)/(x − b) < 0, with a < b, and is asked for the solution set. The reflexive answer is 'x is between a and b', which is correct in the open interval, but the candidate then has to remember that x = b is excluded, and the candidate who forgets that boundary ends up with a 'must be' answer that includes an endpoint. On a 31-question section, the difference between an open and a closed endpoint is rarely the difference between a 705 and a 555, but on a single inequality item it is the difference between a marked and an unmarked point, and the test rewards the candidate who marks it.

Common pitfalls and how to avoid them

• Forgetting to exclude values that make the denominator zero. Write the domain exclusion on the notepad before you solve.
• Treating the sign chart as a single interval when it actually has two pieces. Sketch the number line and shade the two regions separately.
• Multiplying both sides by the denominator without checking its sign. The sign of the denominator decides whether the inequality sign flips; skipping that check is the same error as the linear sign-flip, in a harder costume.
• Ignoring negative numerators in the sign chart. A negative numerator flips the sign of the whole expression and changes which region is positive. Count the negatives explicitly on the number line.

Absolute-value inequalities: the two-case split that is not optional

Absolute-value inequalities are the family where the candidate's hand-eye coordination with the stem is the deciding factor. The stem '|2x − 5| < 9' is a single line, but the solution is the intersection of two linear inequalities: 2x − 5 < 9 and 2x − 5 > −9. The candidate who treats the absolute value as a cosmetic wrapper and solves only one side will get an answer that is half-right at best. The two-case split is not optional. Every absolute-value inequality on the GMAT Focus asks you to write the same expression with the sign and without it, and the candidate who skips one of the two cases is leaving at least one piece of the solution set on the table.

There is a useful shorthand that makes the two-case split almost automatic. An absolute-value expression is a distance on the number line: |2x − 5| is the distance between 2x and 5. The inequality |2x − 5| < 9 then says that the distance is less than 9, which means 2x is within 9 units of 5, in either direction. The visual is a number line with 5 at the centre and 9 units to either side. The candidate who draws the line — 5 − 9 on the left, 5 + 9 on the right — solves the inequality in their head without writing the two cases at all. The shorthand is faster and more accurate, and it scales to the harder absolute-value stems where the expression inside the bars is itself a function of x, not a simple linear term. For most candidates reading this, the visual is the move I would personally pick over the algebraic two-case split, because the visual encodes the boundary points in a way the algebra does not.

Worked example: absolute value with a 'could be' quantifier

Stem: 'If |3x + 6| ≤ 12, which of the following could be the value of x?' The visual: the distance from 3x to −6 is at most 12, so 3x sits in the closed interval [−18, 6]. Divide by 3 to get x in [−6, 2]. The 'could be' quantifier says that any x inside that closed interval is valid, including the endpoints. A candidate who treats the inequality as strict and excludes the endpoints will get the same answer on this stem, but on a stem where the answer choice is exactly the boundary value, the open-versus-closed distinction decides the item. The right move is to write the solution set in interval notation before you look at the choices, and to circle the boundary type — open or closed — so the question form is matched to the right set.

Compound and chained inequalities: the three-region number line

Compound inequalities are the family where the stem looks like a single line but the solution lives in two or three disconnected regions. A typical stem reads: '|x − 4| > 2 or x < 1.' The solution set is the union of two regions: x is more than 2 units away from 4, which gives (−∞, 2) union (6, ∞), and x < 1, which gives (−∞, 1). The union of those two regions is (−∞, 2) union (6, ∞), because the second piece of the absolute-value solution already covers everything below 1. The candidate who treats the two conditions as a single chain will get the wrong answer. The right move is to solve each piece separately, draw a number line, and shade the union of the two regions before looking at the choices.

Chained inequalities are the related family where a single variable appears inside a single expression, but the expression has two or more critical points. A stem like '|x − 1| + |x − 4| ≤ 7' asks for the values of x where the sum of two distances is at most 7. The candidate has to identify the two critical points, test the three regions (left of 1, between 1 and 4, right of 4), and combine. This is a slower stem — it sits at about 120 seconds for a candidate who has trained the pattern — but it is the kind of stem that separates a 78 from an 83 on Quant, because the work is mechanical once the regions are identified. The candidate who tries to fold all three regions into a single equation ends up with a sign error on one of the cases and a wrong answer. The candidate who walks the three regions one at a time, tests a single point in each, and shades the result, gets the right answer on the first pass.

Comparative table: inequality families at a glance

Quantifier discipline: the layer most candidates never see

The single biggest scoring move on inequality items is the one that has nothing to do with algebra. It is the habit of reading the question form before you read the answer choices. The GMAT Focus uses a small set of question forms on inequality stems — 'must be true', 'must be false', 'could be true', 'could be false', 'is always', 'is never', 'how many' — and each form asks for a different property of the solution set. A 'must be true' question asks for an identity: a statement that holds for every member of the set. A 'could be true' question asks for a possible member: a statement that holds for at least one member of the set. A 'must be false' question asks for a contradiction: a statement that holds for no member of the set. The candidate who treats all three as 'find a number in the set' is missing the question, not the maths.

The training move is to underline the quantifier in the stem on every practice item, for at least the first ten sessions of prep. After ten sessions, the underline becomes unnecessary, because the quantifier is read automatically. In the meantime, the underline is doing real work. It is forcing the candidate to pause for half a second at exactly the point in the reading where most candidates' eyes slide past the most important word in the stem. In a section where the adaptive algorithm is using your performance on the first module to choose the difficulty of the second, an error caused by a misread quantifier is not a one-question error; it is a module-difficulty error, and it costs you several questions downstream.

Timing, pacing, and when to guess on an inequality item

Inequality items on the GMAT Focus cluster in the 60-to-140-second range, and the section as a whole gives you roughly 31 minutes. That is a generous average of about 60 seconds per question, but the adaptive module structure means the second module skews harder and longer. The right pacing rule is to spend no more than 100 seconds on a linear, quadratic, fractional, or absolute-value inequality, and no more than 150 seconds on a compound or chained item. If you cross those thresholds, the item is no longer scoring efficiently for you, and the right move is to make a strategic guess and bank the time for the next item. Strategic guess means: eliminate the choices that are obviously outside the solution set, pick the one that sits inside the region your sketch identified, and move on. The candidate who sits on a single item for three minutes is not 'being careful'; the candidate is spending 180 seconds of a 60-second budget and losing two other items in the bargain.

The other pacing rule is to triage the inequality family by its visual density. Linear chains are the fastest — they should sit at the low end of the budget. Fractional and compound inequalities are the slowest, and they should sit at the high end. The candidate who treats all inequality items as 60-second items will overrun on the harder families and short-change themselves on the easier ones. The candidate who triages by family has a much steadier pace across the section, and steadier pace is the variable that most directly controls scaled-score movement in the 78-to-83 band. For most candidates reading this, the single highest-leverage week-one move is to log every inequality item from your practice set with a timestamp, then group the items by family and compute the median time per family. The median tells you which families are over-budget and which are under, and the gap between the two is where your next two weeks of practice should sit.

Score-band implications: why inequality errors are not 'just one question'

Inequalities are not a separate subscore on the GMAT Focus — the section is reported as a single Quant score on a 60-to-90 scale — but the error rate on inequality items is a leading indicator of your position in the 78-to-83 band. A candidate who misses 6 of 31 Quant items will land in the high 70s, and a candidate who misses 4 of 31 will land in the low 80s. The difference between those two outcomes is two questions, and inequality items are exactly the place where most candidates leak the two. The reason is that inequality items are disproportionately represented in the second module of the adaptive section, where the algorithm is using your first-module performance to pick the difficulty of every item. A misread quantifier on a chain stem in the second module is not a one-question miss; it is a signal to the algorithm that you are a lower-difficulty candidate, and the next four items are pulled down to match. Recovering from that signal in the remaining 11 questions of the module is possible, but it requires a run of three or four consecutive correct answers on harder items, and the candidate who has just misread a quantifier is rarely in a state to do that.

The tactical implication is that inequality prep pays a higher score dividend per hour than its share of the syllabus would suggest. A candidate who spends 10 hours on inequalities and reduces their error rate from 40% to 15% on the family is not just adding a few points to their Quant score; they are moving the difficulty of the second module up by a tier, which compounds across the remaining 11 items. That is why this article is built around error families and not around formula review. The candidate who can solve every inequality in this article on paper but who arrives at the test without the reading habit is the candidate who will still miss two of them on test day. The candidate who has trained the reading habit and the notepad discipline will miss one at most, and that one miss is recoverable inside the section. For most candidates, the diagnostic question is not 'do I know the algebra?' — it is 'do I read the quantifier before I look at the answer choices?', and the answer to that question is a much better predictor of test-day Quant performance than any formula list.

Building a 14-day inequalities sprint

A focused inequalities sprint does not need to be long. Two weeks of deliberate practice on the six error families in this article is enough to move a candidate from a careless-miss pattern to a structural-clean pattern, provided the practice is structured around error families and not around problem counts. The first three days should be diagnostic: take 20 inequality items, log the family, log the time, log the error type. Most candidates will discover that 60–70% of their errors live in two families, usually fractional and compound, and the remaining 30–40% are scattered. The next seven days should target the two dominant families, with at least 10 items per family and a hard time budget per item. The last four days should be mixed-family review: 5 items per day, drawn randomly, with the same time budgets, and a written log of the quantifier and the question form for every item. The log is the variable that decides whether the sprint translates into score movement. Without the log, the practice is repetition; with the log, the practice is calibration.

The other lever in the sprint is the notepad. Candidates who arrive at the test with a consistent notepad layout — domain exclusions on the left, sign-flip arrows in the middle, solution-set sketches on the right — are measurably faster on the second module than candidates who improvise. The notepad is a working memory aid, and a working memory aid is exactly what the second module requires, because the second module items are denser and the candidate has less cognitive bandwidth to spare. Build the notepad layout in week one, drill it in week two, and the layout becomes automatic by test day. That is the move I would personally pick over any single content review, because the layout compounds across every item in the section, not just the inequality items.

Conclusion and next steps

The six error families in this article — sign-flip, domain exclusion, two-case split, sign-of-leading-coefficient, quantifier misread, and region-union confusion — are the inequality patterns that decide the difference between a careless miss and a structural clean. The maths is secondary; the reading habit and the notepad discipline are primary. Train those two, and the algebra takes care of itself. The next concrete move is to pull 20 inequality items from your GMAT Focus prep set, log them by family and error type, and identify the two families that are leaking the most points. Those two families are your sprint target. TestPrep İstanbul's diagnostic assessment is a natural starting point for candidates building an inequalities sprint around the six error families above.

Frequently asked questions