GMAT Focus Problem Solving items are the longest-standing question family in the exam's quantitative section, and on the current GMAT Focus edition they still carry the bulk of the arithmetic, algebra, and word-problem weight that a 60+ scaled score demands. A Problem Solving prompt is, in plain terms, a multiple-choice question with five answer options and exactly one correct value, drawn from a pool of content that spans arithmetic, number properties, algebra, word problems, rate and work, geometry, and counting. The candidate who treats every Problem Solving item as a single uniform type usually runs out of time, because the family contains a small set of recognisable sub-types, and each sub-type asks for a different first move. What follows is a working method: how to read a stem, which family the question belongs to, where the trap answers live, and how the answer choices should shape the calculation path. The goal is not to memorise a hundred tricks but to install a five-step discipline that survives the moment the test goes faster than the practice set did.
What a GMAT Focus Problem Solving item actually asks of you
A Problem Solving item presents a complete piece of information, then closes with a question that requires a single numerical or algebraic value. Five answer options sit beneath the prompt, labelled A through E, and exactly one of them is correct. There is no partial credit, no 'none of the above' as a default, and no credit for the method used; the score rewards the right letter. The unwritten contract is that the test-writer must construct five plausible options, which means the wrong four answers are themselves a piece of information. The classic move is to ignore the choices, compute the value, then scan; the smarter move is to read the choices early, because their shape tells you whether the question is testing an integer, a fraction, a sum of two terms, or a ratio.
Three features separate a GMAT Focus Problem Solving item from a school-level word problem. First, the question is always solvable in well under two minutes by a prepared test-taker, but only if the test-taker commits to a method before reaching for the calculator. Second, the wrong answers are not random; they are usually the result of a tempting shortcut that breaks at one step, such as cancelling before factoring, or treating a percentage increase as additive. Third, the answer choices are almost always ordered, either ascending or descending, which is a gift for the elimination method: once you know the answer must be larger than 165 and smaller than 175, you can rule out three options without finishing the calculation.
For most candidates, the first failure mode is reading the stem as a single block of text. The second is launching into arithmetic before deciding what the question is actually asking. The third is using the on-screen calculator on a step that the test-writer designed to be done by hand, which costs ten to fifteen seconds per item and compounds across a section. The remedy for all three is a structured read, and the read starts with the question sentence, not the story.
The five-step method that I install in every Problem Solving tutoring block
For most candidates, the difference between a 47 and a 60+ on Quant is not raw knowledge; it is the discipline of doing the same five steps on every item, in the same order, regardless of how the prompt feels. The method is short, and that is the point. It survives pressure because it does not require thought under pressure.
Step one is to read the last sentence first. The question almost always lives there. Reading it first turns the rest of the stem into a known-purpose data block instead of a story to interpret. Step two is to identify the family. The five recurring families are rate and work, mixture and weighted average, ratio and proportion, integer and divisibility, and algebraic expression. If you can name the family within five seconds, the rest of the method chooses itself. Step three is to set up variables on the scratch pad before computing. A two-letter table for rate problems, a single variable for algebraic expressions, and a fraction line for ratio questions are enough. Step four is to compute only what the question asks, no more. The trap answer that adds an extra step exists because the test-writer expected the candidate to compute one more line than was needed. Step five is to use the answer choices to back-solve when the direct path feels heavy, which is a feature of the format, not a compromise.
Why the first step is the last sentence, not the first
Most candidates read a Problem Solving stem top to bottom because that is how school teaches them to read. On the GMAT Focus this habit is expensive. The first two sentences of a word problem often contain background numbers that are not used; the question sentence carries the verb that defines the calculation. Reading the verb first reframes the entire block. A stem that begins with 'A certain mixture of acid and water' becomes a known-purpose data set the moment you know the verb is 'what fraction of the mixture is acid'. This is one of those tactical adjustments that adds back perhaps eight seconds per item, which across 21 items in a Quant section becomes the difference between finishing and bubbling in a guess at minute 40.
How to recognise the six recurring Problem Solving families
The pool of Problem Solving content is not infinite, and the test-writer draws from a small set of recognisable templates. Recognising the family within the first ten seconds is the single highest-leverage skill a candidate can install, because the family dictates the variable, the equation, and the trap. The six families below cover roughly 90% of the items you will meet.
Rate and work problems
Rate items describe machines, pipes, workers, or people completing a task at a stated rate, and they ask either for a combined rate, a completion time, or a comparison between two scenarios. The first move is to convert every rate into a single unit, usually 'per hour' or 'per minute', and then to add rates when work is in parallel, or to invert them when work is sequential. A common trap is the candidate who adds rates for a sequential task. A second trap is the 'start the second machine halfway through' variation, which forces a time-weighted split rather than a simple rate average. The answer choices usually include the value obtained by adding rates naively, which is the test-writer's way of penalising candidates who skip the variable step.
Mixture and weighted average problems
Mixture items present two solutions of different concentrations being combined, and they ask for the concentration of the mixture, or for one of the two original concentrations. The lever is the weighted average formula, which says the distance from the mixture to each original is inversely proportional to the volume of that original. The trap is to set up a single equation where the test-writer expected two unknowns; the cleanest version of the item is solvable by inspection once the weighted-average lever is in place. When the answer choices are far apart, the candidate can use the alligation method and avoid algebra altogether.
Ratio and proportion problems
Ratio items often hide a proportion inside a word problem. The reliable move is to assign a single variable to one part of the ratio and write the rest of the ratio in terms of that variable; the stem usually gives a sum or a product that closes the system. The trap is to assume the ratio is preserved after a change, when the change typically alters one term independently. For example, 'the ratio of boys to girls is 3 to 5, then 6 boys leave' produces a new ratio, and the candidate must track the change to the boys term only.
Integer and divisibility problems
Integer items ask for the largest value, the smallest count, or the remainder when a number is divided by another. The reliable move is to factor the divisor first, then test each answer choice for the condition. A second reliable move is to scan the answer choices for parity, because the test-writer will include at least one even answer when the correct answer is odd, and vice versa. The trap is to commit to long division on a divisor like 36 when the question only needs the remainder modulo 4.
Algebraic expression problems
Algebra items usually present a function, a relation between variables, and a question that asks for a specific value when a condition is met. The reliable move is to substitute simple values and compute, then re-substitute to confirm. The trap is to chase the algebraic expansion, which is the path the test-writer designed for candidates who do not see the substitution shortcut. For most candidates reading this, substitution is faster and more accurate than symbolic manipulation, and the only time to flip that default is when the substitution produces messy arithmetic.
Counting and probability problems
Counting items ask for the number of ways an event can occur, often with a constraint. The reliable move is to separate the constraint from the unconstrained count, then to subtract the forbidden cases. The trap is double-counting, which happens when the candidate applies two constraints independently. Probability items follow the same pattern: count the favourable, count the total, and divide, with care to avoid complementary-counting errors.