CAT PYQs: How Many Years Should You Practice and How

"How many years of PYQs should I do?" is one of the most common questions in CAT prep groups, and it usually gets answered with a number, "do the last 5 years" or "do the last 10," without much reasoning behind the number itself. The honest answer is that the right amount depends less on a fixed count and more on what you're trying to learn at each stage of prep.

Why CAT PYQs are different from generic practice questions

CAT's pattern, difficulty calibration, and question style shift gradually year over year, sometimes within a single year across different slots. A 2017 DILR set and a 2023 DILR set can feel like different exams in terms of data complexity and the number of variables in play. This is exactly why previous year questions are worth more than an equivalent volume of generic practice material. They're not just more questions, they're a record of how the exam actually behaves, which generic question banks can only approximate.

The three things PYQs are actually good for

One, calibration. Working through real exam level difficulty tells you, honestly, where you stand against the actual bar, not against a curated practice set designed to feel achievable.

Two, pattern recognition. After enough exposure, you start recognizing question archetypes, the VARC inference trap that shows up in slightly different clothing every year, the DILR set structure that always has one piece of obvious information that's actually a red herring.

Three, time pressure benchmarking. A drill question answered untimed and the same difficulty question answered with the clock running are functionally different tests. PYQs, attempted as full timed papers, are the only honest way to measure this gap.

So, how many years?

For most aspirants with a four to six month prep window, working through four to six years of full papers, attempted properly under timed conditions, gives a strong and representative sample of the exam's range. Beyond that, the marginal value of an additional year drops, not because older papers are useless, but because you're better served spending that time on focused, topic wise PYQ practice rather than one more full paper attempted casually.

This is the reasoning behind topic wise PYQ practice. Instead of working through entire papers chronologically, you isolate a specific weak area, say Para Summary or Logical Reasoning sets, and pull questions on that exact topic across multiple years in one sitting. This compresses what would otherwise take you through six separate full papers into a single, focused session.

The order that actually works

Start with section wise practice, not full papers, in the first few weeks. A full timed paper is a high stakes, high fatigue exercise. Running one before you've built section specific comfort tends to produce a discouraging score that doesn't reflect your actual trajectory, just your unfamiliarity with the full paper format.

Once each section feels reasonably stable individually, move to full timed papers, ideally one every seven to ten days, alternating between different years so you're not just getting used to one specific paper's quirks.

In the final six to eight weeks, shift toward smart mix practice, randomized questions on a given topic pulled across multiple years, rather than working through one year at a time. By this stage you've likely seen most individual papers at least once. Smart mix forces you to apply pattern recognition fresh, without the benefit of half remembering a specific paper's answer key.

A note on re-attempts

It's tempting to re-attempt a paper you did poorly on, hoping a second pass will feel more reassuring. There's some value here, but only if enough time has passed that you're genuinely re-solving the questions rather than half recalling answers. As a rough rule, don't re-attempt the same paper within three to four weeks. The score improvement you'll see is mostly memory, not skill.

PYQ volume matters less than most aspirants assume. What matters is whether each attempt is structured, timed, reviewed, and followed by an honest accuracy breakdown by topic, rather than treated as a box to check off a list.