How DILR Scoring Works: Is There Partial Marking Here?

DILR

Two beliefs circulate about DILR scoring and both are wrong in opposite directions. One says a set is all or nothing, so an unresolved set is a total loss. The other says there is partial credit for partial work, so effort on a hard set is never wasted.

Neither describes what actually happens. There is no partial marking for a set as a unit, and questions within a set are marked individually, which means a set you never fully resolved can still produce marks. That distinction sounds academic and it changes how you should behave in the last ten minutes of a section.

What the Rules Actually Are

The marking scheme is the same across all three sections and it is worth stating exactly.

A correct answer is worth three marks. An incorrect multiple choice answer deducts one mark. A question left unattempted scores zero. Type-in-the-answer questions, where you enter a value rather than selecting an option, carry no negative marking at all.

Nothing in that scheme references sets. Marks are assigned per question, and the set is a presentational device: a shared scenario with several questions attached to it. The exam does not award or withhold anything at the level of the set.

The section runs to a hard forty minute limit, and in the 2022 paper DILR carried 20 questions worth 60 marks, though the composition has moved between years.

Why Both Common Beliefs Are Wrong

Each belief leads to a specific behaviour, and both cost marks.

Effort Is Never Directly Rewarded

This is the harder half to accept. Twelve minutes on a set that did not resolve produces exactly the same score as not opening it, unless some of its questions happened to become answerable along the way. There is no credit for having tried.

But Some Questions Are Reachable Early

Sets frequently include questions that do not require full resolution. Questions asking whether a particular arrangement is possible, or what must be true, can often be settled from the constraints alone. Questions asking for a specific value, position or count generally cannot.

Which Is Why Reading Order Matters

If you read a set's questions before building the structure, you know which ones are cheap and which require everything. That knowledge changes both your commitment decision and your exit strategy, and it costs about thirty seconds that you were going to spend reading the questions anyway.

Question type within a set	Needs full resolution?	Reachable from partial work?
Which of these must be true	Often not	Frequently
Is this arrangement possible	Often not	Frequently
What is the value for a given item	Usually yes	Rarely
How many items satisfy a condition	Usually yes	Rarely

The Type In Question Inside a Set

DILR uses type-in questions regularly, and the scoring consequence is worth isolating.

A wrong type-in answer costs nothing. That means a type-in question attached to a set you partially built should always be attempted if you have any reasoned value, because the downside is zero and a blank is worth exactly the same as a wrong answer.

This is the cheapest available recovery from a set that went badly. A candidate who has spent eight minutes on a structure that did not fully resolve, and who then leaves a type-in question blank because they are not certain, has paid the cost and declined the only free part of the compensation.

The behaviour to build is to check the format of each question in a set before deciding what to leave. One blanket policy applied to both formats is either too cautious or too aggressive, and usually both in the same section.

What This Changes About Commitment

The scoring rules make the commitment decision sharper rather than softer.

Because effort earns nothing, a set that will not resolve is a pure loss of minutes, and the correct response is to leave it early. The sunk cost at minute eight is genuinely sunk, and the only live question is what the next eight minutes should buy.

Because questions are marked individually, leaving is not necessarily a total loss, which removes the argument that you may as well continue since you have already invested. Check what is reachable, collect it, and go.

And because a wrong multiple choice answer costs a mark, the questions in an abandoned set that you cannot reach should be left blank rather than guessed. The exception is any type-in among them, where guessing is free.

The Arithmetic of a Bad Set

Running the numbers makes the strategy concrete rather than abstract.

Suppose a set carries five questions. You spend nine minutes and your structure is incomplete. Two questions are of the must-be-true kind and your constraints settle both; one is a type-in you can reason a value for; two require full resolution.

Leaving immediately and blankly scores zero from nine minutes. Spending thirty more seconds to answer the two you can settle, plus entering the type-in, converts the same nine minutes into a meaningful return with no additional risk, since the type-in carries no penalty.

That is not a strategy for salvaging disasters. It is the reason to read questions first, because the information that makes this possible arrives before the solving rather than after it.

What Not to Do With the Rest

Guessing the two questions that need full resolution is a bet you are paying to place, since each wrong multiple choice answer deducts a mark. Leave them, and take the loss on those specifically rather than converting a partial recovery into a net negative.

How to Practise the Habit

In untimed practice, read every set's questions before touching the structure, and note which ones look reachable early. Running that habit across a block of caselet based sets is a good place to start, since caselets often mix cheap and expensive questions. Doing it where there is no time pressure is what makes it survive under pressure. Our guide on the five signs a DILR set is a trap covers the features that predict a set will not resolve at all.

Why the Set Feels Like the Unit Anyway

Even once the rule is understood, the section continues to feel all or nothing, and that feeling is not irrational.

The reason is that the economics of a set really are lumpy, even though the marking is not. You pay a fixed setup cost to build the structure, and most of the returns arrive together once it resolves. So a set behaves like a single investment with a threshold, and the marking rule does not change that; it only means the threshold is softer than candidates assume, with a couple of questions sitting below it.

Holding both facts at once is what produces the right behaviour. Treat the commitment decision as lumpy, because it is, and treat the exit as granular, because the marking is. Candidates who apply one of those framings to both moments either overcommit or leave marks behind.

The Answer

Is there partial marking for DILR sets? No, not as a unit. There is no mechanism that awards credit for having partly solved a set, and effort spent on a set that does not resolve earns nothing by itself.

What exists instead is that every question is marked on its own answer, three marks for correct, minus one for a wrong multiple choice, zero for a blank, and no penalty on type-in questions. So a set that never fully came together can still produce marks from the questions your partial work happened to settle.

The behaviour that follows is small and specific. Read the questions before you build, so you know which are cheap. Before abandoning anything, spend thirty seconds collecting what is reachable and enter every type-in you can reason. Leave the rest blank rather than guessing. That sequence turns the most demoralising event in the section into a contained one.

If your DILR sections contain sets you spent minutes on and scored nothing from, the exit habit is where the recoverable marks are. A CAT preparation strategy review will show how much, and a personalised CAT preparation plan can build the reading-first habit into your practice.