Is CAT DILR Testing Reasoning Ability or Just Speed?

DILR

The question gets asked as a complaint. If I can solve every set given enough time, isn't DILR just a speed test? And underneath it is a genuine grievance: a candidate who reasons well feels the section is measuring something cruder than their ability.

It is measuring something, and it is neither of the two options in the question. DILR tests reasoning under a constraint, and the constraint changes what counts as good reasoning rather than merely compressing it. This piece explains what the section actually rewards, why "I could do it with more time" is both true and irrelevant, and what to train if you are strong at one half and weak at the other.

Neither Reasoning Nor Speed

The question offers two answers and the correct one is not on the list.

If DILR were a pure reasoning test it would be untimed, or generously timed, and would separate candidates by whether they can construct valid inferences at all. If it were a speed test the sets would be mechanical and would separate candidates by execution rate.

It is neither. There is a hard 40 minute sectional limit with no movement between sections, several sets, and no obligation to attempt any particular one. That combination creates a different task: allocate a scarce resource across options of unknown cost, then execute the ones you chose. Both halves are reasoning. The second half is the part people call reasoning and the first half is the part they mistake for administration.

On "I Could Solve It With More Time"

This deserves a direct answer rather than a dismissal, because it is usually true.

Most candidates preparing seriously can solve most DILR sets given unlimited time. That is a real statement about capability and it is not worth nothing, because it means the reasoning machinery is present. What it is not is evidence that the section is unfair.

Every test measures a capability under conditions, and the conditions are part of what is being measured rather than an obstacle placed in front of it. A management programme is selecting for people who make good decisions with incomplete information and limited time, which is a defensible thing to test for and is exactly what the format produces. Whether you agree with that choice is a separate argument from whether the section is coherent.

What Speed in DILR Actually Consists Of

The word speed suggests moving faster, and that is almost never where the time comes from.

None of these is hurrying. They are all reasoning decisions made early, and they are why two candidates with similar reasoning ability can finish the same set eight minutes apart with neither one rushing.

Recognition Is the Cheapest Win

Of the four, recognition returns the most per hour of practice, and it is the one that volume genuinely builds.

The recurring families are stable enough to learn: arrangements, games and tournaments, distribution and selection, tables and caselets, networks, scheduling. A candidate who has worked through enough of each recognises the shape on sight and arrives with a default structure already in mind, which removes the most expensive part of a cold start.

This is why practising by type first and mixed sets second is the right order. Type-wise work builds the recognition, and mixed practice then trains you to identify which type you are looking at under pressure.

What feels like the problem	What it usually is	What to train
Not enough time	Slow cold start on unfamiliar shapes	Type-wise volume for recognition
Getting stuck midway	Wrong representation chosen early	Deliberate practice choosing structures
Carrying many possibilities	Constraints applied in a weak order	Identify the most restrictive condition first
Rebuilding for each question	Structure not built for reuse	Read all questions before building
Entering bad sets	Selection treated as administration	Estimation drills without solving

Why Selection Matters More Here Than in Quant

DILR is more unpredictable between years than Quant, and that has a direct consequence for how to prepare.

Set types and counts swing, so the frame of "this chapter carries this many marks" does not apply the way it does to Quant. What transfers across papers is not coverage of a fixed syllabus but the ability to recognise a recurring set type and to judge quickly whether this particular instance is worth your minutes.

That is why set selection is a trainable skill that almost nobody trains, and why it is worth disproportionate attention. In Quant a weak selection round costs you some questions. In DILR, where a single set can absorb a third of the section, it costs you the section.

Working Out Which Half Is Yours

The two failure profiles need different work, and mistaking one for the other wastes months.

If you attempt few sets and get most of them right, your reasoning is fine and your cost is in execution speed and cold starts. The work is type-wise volume for recognition and deliberate practice on representation.

If you attempt many sets and get a poor proportion right, speed is not your constraint. You are entering sets you should not enter, or your structures are breaking down under pressure, and more practice at going faster will make it worse rather than better.

If you attempt few sets and also get them wrong, the honest answer is that the reasoning fundamentals need building first, and selection work on top of shaky fundamentals produces nothing. That is a longer job and it is better to know it than to keep drilling selection.

A Diagnostic That Settles It

There is a simple test that separates a reasoning problem from a constraint problem, and it takes one sitting.

Take four sets you failed in a timed mock and attempt them untimed, properly, without looking at solutions. If you solve all four comfortably, your reasoning is intact and everything you need to fix lives in selection, recognition and representation. If you cannot solve two of them even untimed, you have found a genuine reasoning gap and it needs addressing directly before any of the speed work will pay.

Run this once rather than repeatedly. Its value is in telling you which programme you are on, and after that the timed work is where the improvement comes from.

What Good Selection Actually Looks Like

Since selection is where most of the argument lands, it is worth being concrete about what a good selection round contains, because "pick the easy sets" is not an instruction anyone can execute.

A selection round is a sweep across every set in the section, spending under a minute on each, producing three judgements per set rather than one. Can I see what structure this wants. Do the constraints look like they will close quickly or branch widely. Do the questions ask for one thing from the structure or several different things. Those three answers together give you a rough cost, and cost is what you are ranking by.

Then you order the sets and start with the cheapest that you are confident about, not the most interesting. The most interesting set in a section is frequently the most expensive, because what makes it interesting is usually that its structure is unusual, and unusual means no recognition to draw on and a cold start you will pay for in minutes.

What makes this hard is that it requires holding back. Candidates who have read a set they can clearly do find it almost physically difficult not to start, and the sweep collapses at the second set. The discipline is not analytical, it is behavioural, which is why it has to be rehearsed in mocks until it stops requiring a decision.

The Summary

DILR is not a reasoning test and not a speed test. It is reasoning under a constraint, and the constraint changes what good reasoning looks like rather than simply compressing it.

The claim that you could solve every set with more time is usually true, applies to most of the serious field, and does not describe an injustice. What the section actually selects for is the ability to choose well among options of unknown cost and then execute the ones you chose, and the choosing is reasoning rather than administration.

Speed here is almost never hurrying. It is recognition of the set family, choosing the right representation immediately, applying the most restrictive constraint first, and building a structure the later questions can reuse. Diagnose which half is yours with an untimed attempt at sets you failed, and then train the half that is actually costing you, because the two programmes look nothing alike.

If you cannot say whether your problem is selection or execution, that is the first thing to settle, because the two need opposite work. A CAT preparation strategy review will read that split from your mock data, and a personalised CAT preparation plan puts the right half into your week.