Analysing DILR in CAT mocks: a worked review of one section

DILR is the section where two mocks a week apart can give you wildly different scores with no change in ability. The reason is structural. You are not answering 22 separate questions; you are choosing a few sets out of several, and one wrong choice can eat a quarter of your 40 minutes. So the most useful thing a mock tells you in DILR is not "you scored 20". It is which decisions cost you marks.

This guide gives a four-step review you can run after every mock, then walks through one section in full with made-up numbers so you can see the method at work.

Why a DILR review is different from a QA review

In QA, a wrong answer usually points to a concept or a careless step. In DILR, marks are lost at three different levels, and each needs a different fix:

Most aspirants review only the third level, because it is the one the score report shows. The first level is usually the biggest.

The four-step review

1. Rank every set by true difficulty. After the mock, give each set a fresh untimed attempt, including the ones you never touched. Time each one. The order of those untimed times is the true difficulty ranking.
2. Compare with what you chose. Write down the order in which you actually attempted sets, and how you rated each one during your scan. Where your scan rating and the true ranking disagree, you have a selection lesson.
3. Compare methods. For every set, read the official solution and ask one question: was there a faster representation? A grid instead of a list, a table of cases instead of trial and error, a single variable instead of four.
4. Classify every wrong answer. Label each one as selection, solving or accuracy, and write a one-line cause in plain words.

Budget roughly as much time for the review as the section took, or more. A mock you do not review is mostly a timing drill.

A worked example (illustrative data, made up)

Suppose an aspirant has just finished a mock DILR section of 22 questions in five sets. Here is what happened in the exam hall. All numbers below are invented for teaching.

The scan took 2 minutes, so the section adds up to 40.

Score. 7 correct × 3 = 21. One wrong MCQ costs 1. The wrong TITA costs nothing. Net score 20, from 9 attempts at 78% accuracy.

Step 1: the untimed re-solve

The next day, the aspirant re-solves all five sets without a clock and notes the time each takes to finish completely:

Step 2: selection

The aspirant started with the hardest set in the paper and never opened the third-easiest. The truth-teller set was rated "hard" in the scan only because it was wordy.

Now put a number on it. The 15 minutes on games produced one correct and one wrong answer: 3 − 1 = 2 net marks. The truth-teller set took 12 minutes untimed. Even four correct answers out of its five would have been worth 12 marks. That single choice cost about 10 marks, half of the final score.

Step 3: methods

Reading the solutions shows two things:

• In the games set, the official solution fixed the total points first (matches × points per match) and worked backwards. The aspirant had tried to fill a results grid match by match.
• In the table set, the solution added a "difference" column before answering. The aspirant recomputed each difference inside each question.

Both are reusable ideas, so both go into the aspirant's notebook of methods, not just the answers.

Step 4: classify the wrong answers

Notice that neither wrong answer was a knowledge gap. Both were reading slips.

The lesson for this mock

Three lines, written at the top of the mock sheet:

• Wordy is not the same as hard. Rate a set by how many fixed facts it gives, not by its length.
• In tournament sets, find the total points first.
• Before entering any percentage, say the base out loud.

Keep a set log across mocks

One mock is an anecdote. Five mocks are a pattern. Keep a running log with one row per set you saw:

After four or five mocks, sort the log by set type. Here is what a made-up log might show after five mocks:

This aspirant's weekly practice writes itself: five binary-logic sets (the type they avoid), three tournament sets solved with the total-points method, and a percentage-base drill for tables. The arrangements need only one set a week.

Questions to ask about every set

• Did I see this set during the scan, and what did I rate it?
• If I skipped it, was the skip right? Check against the untimed time.
• If I attempted it, at what minute did I know whether it would crack?
• Which condition unlocked the set, and how long did I take to use it?
• Was any single question answerable without solving the whole set?

The last question matters more than it seems. Some questions in a hard set, such as the total number of matches in a tournament, need only one clue. Spotting these in review trains you to harvest them in the exam.

What to do next

• Re-solve your most recent DILR section untimed, all sets, and fill in the true-rank table.
• Put a mark value on your worst selection decision.
• Start the set log today and keep it for at least five mocks.
• Read DILR set selection to sharpen the scan, and time management in DILR for the in-section clock.
• For the full mock routine across all three sections, see mock tests and analysis for CAT.