Word problems are the heart of the GMAT Focus Quantitative section, and they are also the part where most candidates quietly bleed points. You can be fluent in algebra, comfortable with fractions, and quick on a calculator, then still walk out of the test centre wondering why a passage about a train and two cities cost you a 685. The reason is rarely arithmetic. The reason is translation. The exam gives you prose, sometimes dense prose, and asks you to turn it into a single equation or inequality. The candidates who score 81+ on Quant have not memorised more formulas than everyone else. They have built a habit of reading prose the way a translator reads a foreign language: looking for actors, actions, quantities, and the relationship that connects them.
On the GMAT Focus, the Quantitative section presents 31 questions in 45 minutes, and the word problems that appear there are not puzzle-fair. They are designed to be solved in roughly 90 seconds, which means a long chain of algebra is itself a warning sign. The smartest move is usually to set up the relationship first, then decide whether to solve by algebra, by substitution of answer choices, or by a quick ratio shortcut. This article walks through that habit from the inside out: how to read the prompt, how to name your variable, how to choose between equation-building and backsolving, and how to recognise the four or five prompt families the GMAT Focus keeps recycling across editions.
Why GMAT Focus word problems punish fluent calculators
The single most common mistake I see in diagnostic sessions is also the most counterintuitive: strong maths students lose more points on word problems than weak ones, because the strong students trust the numbers and the weak students hesitate long enough to re-read. The prompt on a GMAT Focus word problem is engineered to look like a maths problem with extra words, but the words are the maths. If you skip the prose and reach for an equation, you will often construct the right equation for the wrong story. The exam writers know this. They place a distractor that matches the equation a hurried reader would build.
Consider a typical rates question. A candidate reads "Pipe A fills a tank in 6 hours, Pipe B in 4 hours, and Pipe C drains the tank in 12 hours. If all three are open together, how long does it take to fill the tank?" The fast reader writes 1/6 + 1/4 + 1/12 and adds the fractions. The careful reader notices that Pipe C drains, so its rate is subtracted, not added. The answer choice that uses 1/6 + 1/4 − 1/12 is the only one that respects the prose. The trap answer, the one a fluent calculator picks, uses 1/6 + 1/4 + 1/12. That single sign is the difference between a 645 and a 685 on the GMAT Focus, and the candidate who lost the point will almost never remember it as a prose problem. They will remember it as an arithmetic problem they 'just made a mistake on'.
Three habits break this pattern. First, read the last sentence of the prompt before reading the rest. The question stem tells you what the equation has to produce, and that shape often reveals which quantities are inputs and which are outputs. Second, underline the verbs and the direction of each verb. 'Drains', 'gives away', 'loses', 'depreciates' all flip the sign of the rate they attach to. Third, when you finish setting up the equation, glance at the units. If the right-hand side is supposed to be hours and the left-hand side is currently a unitless number, you have either multiplied instead of divided or skipped an inverse-rate conversion.
The seven word-problem archetypes the GMAT Focus recycles
Almost every word problem you will meet on the GMAT Focus Quant section fits one of seven families. Naming the family before you start computing is itself a time-saver, because each family has a default setup, a default trap, and a default solve-method. If you cannot name the family within 15 seconds of reading the prompt, you are probably missing a piece of information.
1. Combined-work problems. Two or more agents (pipes, machines, workers) act on a shared object. The setup is always in terms of fractional rates that add, with drains subtracted. The trap is sign confusion on subtraction. The default solve is a common-denominator addition, then taking the reciprocal of the total rate.
2. Distance-rate-time problems. Two objects move, sometimes toward each other, sometimes in the same direction. The setup is d = r·t for each object. The trap is failing to align starting times or directions. The default solve is to express one distance in two ways, which often eliminates time entirely.
3. Mixture and concentration problems. Liquids of different strengths are combined, sometimes with replacement. The setup is total solute = sum of (concentration × volume). The trap is forgetting that the final concentration applies to the final volume, not the original. The default solve is a weighted-average equation.
4. Profit, revenue, and cost problems. A business scenario with fixed and variable costs, sometimes with discount stacks. The setup is revenue − cost = profit, then expressed as a margin. The trap is double-counting a discount or a tax. The default solve is to assign a clean base value, usually 100, and read the percentage changes as multipliers.
5. Ratio and proportion problems. Two quantities are linked by a ratio, then a third quantity changes the absolute values. The setup is a system of two linear equations in two unknowns where the ratio gives you the multiplier. The trap is the candidate who solves for x and y separately when only the ratio was asked. The default solve is to let the ratio be k·a and k·b, then substitute.
6. Overlapping-set problems. Two groups share a region, sometimes with a third group subtracted. The setup is the standard two-circle or three-circle Venn formula: total = A + B − both + neither, or the three-circle variant. The trap is forgetting to subtract the triple overlap or to include 'neither' when it is part of the total. The default solve is a labelled grid with the four or eight regions filled in.
7. Sequence and pattern problems. A list of numbers follows a rule, sometimes a recursive rule, sometimes a closed-form rule. The setup is to compute the first three or four terms to detect the pattern, then generalise. The trap is a candidate who tries to derive a closed form when a recursive count would do. The default solve is to ask: 'What would the 8th term be?' and to verify the rule against the 1st, 2nd, and 3rd.
Notice that none of these setups requires calculus, and none of them requires a clever trick. They require pattern recognition. A candidate who has practised 40 of each family will read a new prompt and instantly see the family label. That label is what frees the 90-second budget.
How to read a word problem the way a translator does
Reading a word problem is not the same skill as reading a Reading Comprehension passage. Reading Comprehension rewards you for tracking an author's argument. Word-problem reading rewards you for extracting an equation. The cognitive task is closer to coding: parse the prose into a structure, then execute the structure.
Three parsing moves cover most of the value. The first move is to identify the actors. In a profits question, the actors might be 'the company' and 'the supplier'. In a rates question, the actors are the agents acting on the object. The second move is to identify the quantities attached to each actor. In a distance problem, each actor has a starting time, a speed, and a position. The third move is to identify the relationship. Most relationships on the GMAT Focus are linear, and they are signalled by a small set of words: 'combined', 'in total', 'gives', 'receives', 'leaves', 'arrives', 'is reduced by', 'is increased by'.
The translation step is where most candidates stop. They identify the actors and the quantities, then they begin to compute. The intermediate step that gets skipped is writing the relationship in English, in a single sentence, before converting it to algebra. For example, the prompt 'A tank is filled in 6 hours by Pipe A alone, and in 4 hours by Pipe B alone. If both pipes are open for 2 hours, then Pipe B is closed, how much longer does Pipe A need to finish the tank?' should be translated as: 'The fraction filled by A in 2 hours, plus the fraction filled by B in 2 hours, plus the fraction filled by A in the remaining t hours, equals one whole tank.' That English sentence is the equation. Once the English is correct, the algebra is a mechanical step.
A useful drill: take any word problem you have already solved, cover the equation, and write the English sentence that the equation encodes. If you cannot write the sentence, you did not actually understand the prompt. You pattern-matched the answer.
Picking the variable: where candidates quietly lose 30 seconds
On a 90-second-per-question test, 30 seconds is a lot. It is the difference between finishing the section and having to guess on the last two problems. One of the cleanest ways to recover those 30 seconds is to spend 10 of them choosing the variable. The wrong variable choice forces you to set up two equations when one would have done. The right variable choice turns a prompt into a single line.
For example, consider: 'The price of a stock rose 20% on Monday, then fell 25% on Tuesday. What percentage of the original price is the Tuesday closing price?' The candidate who picks the original price as the variable and calls it 100 is doing the right thing. The candidate who picks x for the original price and then watches fractions accumulate is doing the wrong thing. In a multiple-choice format, a clean value like 100, 60, 12, or 360 will almost always simplify the arithmetic. If the answer is a percentage, picking a base of 100 turns every multiplier into a number. If the answer is a count, picking a base of 60 or 360 turns fractions into whole-number minutes.
There are three rules I teach for variable choice. Rule one: if the prompt contains a percentage change, let the original quantity be 100. Rule two: if the prompt contains a ratio and a difference, let the ratio be k·a and k·b, where a and b are the smallest whole numbers in the ratio. Rule three: if the prompt contains a fraction of a whole, let the whole be the lowest common multiple of the denominators, so that the fractions become whole-number subtractions. None of these rules is mathematical in the deep sense. All of them are time-saving. The GMAT Focus rewards time-saving.
Algebra setup versus backsolving: a triage decision
There is a persistent myth that the GMAT Quant section is a pure algebra test. It is not. It is a reading-and-arithmetic test with an algebra wrapper. For roughly a third of word-problem prompts, the fastest path to the answer is not to solve the equation. It is to test the answer choices. Backsolving works when the question stem has a specific value, when the answer choices are numbers, and when the unknown is a single quantity that you can plug into a clean equation. It fails when the answer choices are variables in terms of other variables, or when the prompt contains a 'must be true' or 'could be true' qualifier.
The triage is fast. Read the stem. If it asks 'What is the value of x?', and the choices are five numbers, backsolve from the middle choice, C. If the answer checks, you are done in 30 seconds. If it does not, move to B or D depending on whether the result was too high or too low. If the stem asks 'Which of the following could be the value of x?' or 'Which must be true?', the answer choices are statements, and backsolving is a much weaker tool. There, you must solve the equation.
A subtler point: even when you intend to solve the equation, you can use the answer choices as a sanity check. Solve for x, glance at the choices, and check that your x falls in the same neighbourhood. If your algebra says x = 47 and the choices are 0.47, 4.7, 47, 470, and 4,700, the unit-decimal placement is your real risk, not the algebra. A 5-second glance saves a 5-minute re-score.
Common pitfalls and how to avoid them
Most word-problem errors on the GMAT Focus come from one of four traps, and once you have seen the trap once, you can usually see it in the answer choices before you finish the prompt. The traps are sign errors, unit mismatches, double-counting, and the silent extra condition.
Sign errors happen when a verb hides the direction of change. 'Loses 20% of its value' and 'is reduced to 20% of its value' sound similar but produce different multipliers (0.80 and 0.20). The candidate who reads both as 20% off and applies the same multiplier loses the point. The fix is to convert the prose to a multiplier and then to check that the multiplier is between 0 and 1 for decreases and greater than 1 for increases. If your multiplier for a decrease is 1.20, you read the prose wrong.
Unit mismatches happen when the prompt mixes hours and minutes, dollars and cents, or miles and feet. The fix is to pick a unit at the start of the prompt and convert every quantity to that unit before you write the equation. A 90-second problem with three unit conversions is still a 90-second problem if you batch the conversions at the start; it becomes a 4-minute problem if you convert as you go.
Double-counting happens when a quantity appears twice in the equation because the prompt mentions it in two different sentences. The fix is to assign each piece of data to a single line in a two-column setup, one column for 'given' and one for 'to find', before you start writing equations.
The silent extra condition is the most dangerous trap, because it is not a maths error. The prompt states, almost in passing, that 'all the tickets were sold' or 'no apples were left over' or 'the trains met exactly once'. That extra condition is what makes the problem solvable, and the candidate who misses it sets up an equation with two unknowns and one equation. The fix is to reread the prompt, this time underlining every word that constrains the answer. There will always be at least one constraint that is not signalled by a number.
Comparing the seven families at a glance
Below is a compact reference table for the seven word-problem families. Use it as a study card during the first three weeks of preparation, and as a checklist during the final two weeks. If a prompt does not fit any of the rows, you have either misread the prompt or you are looking at a Data Sufficiency question, which uses word problems in a different way and is outside the scope of this article.
| Family | Default setup | Default trap | Default solve | Time budget (sec) |
|---|---|---|---|---|
| Combined work | Sum of fractional rates | Adding a drain instead of subtracting | Common denominator, then reciprocal | 75 |
| Distance-rate-time | d = r·t for each object | Misaligned start times or directions | Express one distance two ways | 90 |
| Mixture / concentration | Total solute = Σ (c · v) | Final volume treated as original volume | Weighted-average equation | 100 |
| Profit / revenue / cost | Revenue − cost = profit | Double-counting a discount or tax | Base value of 100, multipliers | 90 |
| Ratio / proportion | Let ratio be k·a, k·b | Solving for absolute values when only ratio asked | Substitute k·a, k·b into the second equation | 85 |
| Overlapping sets | Two-circle or three-circle Venn | Missing triple overlap or 'neither' | Label every region, then sum | 95 |
| Sequence / pattern | First three or four terms | Closed form when recursive count would do | Verify against 1st, 2nd, 3rd; generalise | 80 |
The time budgets in the rightmost column are realistic averages for a candidate scoring 81+ on Quant. Candidates scoring in the 75–78 band typically run 15–20 seconds over each budget, and that overflow is what forces the last two questions of the section to be guessed. The 90-second average is not a suggestion; it is the structural pacing target that the adaptive section assumes.
Practising word problems: a six-week micro-cycle
Knowing the families is not the same as owning them under timer pressure. The two skills need different practice methods. Family recognition is built by untimed problem sets of 12 to 15 questions, grouped by family, with the prompt covered until you have written the English sentence that the equation encodes. Timed execution is built by mixed sets of 10 questions, drawn randomly from three families, with a hard 90-second-per-question stopwatch and a rule that any unfinished question counts as wrong.
Week 1 and week 2 should be untimed and family-grouped. Spend two sessions on combined work and distance-rate-time, two on mixture and profit, two on ratio and overlapping sets, and one on sequences. For each session, do 12 problems, read the English sentence aloud before you write the equation, and mark the trap that the wrong answer was built from. By the end of week 2, you should be able to name the family of any prompt within 10 seconds.
Week 3 and week 4 should be timed and mixed. Do one 10-question mixed set per day, hard 90-second limit, and after each set spend 15 minutes on the two or three problems you missed. The review matters more than the set. For each missed problem, write one sentence: 'I missed this because I treated X as additive when it was subtractive.' That sentence is your personal error log, and it will repeat itself across the next 50 problems until you fix it.
Week 5 should be a single full-length Quant section under test conditions, then a re-group of the misses into families. If the misses cluster in two families, spend week 6 on those two families only, with a fresh 15-problem untimed set for each, followed by another full Quant section. The cycle ends when your miss distribution is roughly even across families, which is the signature of a candidate who is reading the prose, not pattern-matching the numbers.
How word problems interact with the rest of the GMAT Focus Quant section
The GMAT Focus Quantitative section is not a pure word-problem test. It also contains pure-arithmetic, algebra, and geometry questions. Word problems are usually the most time-consuming family, which means a candidate who lets word problems eat 3 minutes each will not have time for the data-interpretation sets at the end of the section. The section is adaptive, and the difficulty of the second module is calibrated to your performance in the first 12 to 14 questions. Slow word-problem execution in the first module pushes you into an easier second module, which lowers the ceiling on your Quant score.
Two tactical moves help. The first is to do the data-interpretation sets before the word problems in module one, because the data-interpretation sets reward speed and accuracy in equal measure and they tend to be shorter, which gives you a confidence buffer. The second is to skip-and-return on any word problem that you cannot name within 15 seconds. Flag it, do the other questions, and return with 3 to 4 minutes of banked time. A skipped word problem is not a lost word problem; it is a deferred word problem. A word problem that you have spent 4 minutes on and still cannot set up is a lost word problem and two lost minutes.
For most candidates, the score lift from a disciplined word-problem routine is between 15 and 25 scaled points on the GMAT Focus Quant section. That lift is not visible in any single practice question. It shows up over the course of a full section, when the cumulative 30-second savings per problem add up to three extra minutes, which is enough time to convert a guessed last question into a solved last question. That is where the move from a 655 to a 705 usually lives, and it almost never comes from learning a new formula. It comes from learning to read the prose.
What to do on test day
On test day, the value of this entire routine is to free your attention for the prose. If you have practised the seven families, you will not be deciding which family the prompt belongs to. You will be deciding what the actors and the relationship are. That decision is the only one that has to be made fresh, every time, on every prompt. Everything else is pattern, and pattern is fast.