The Irreversible Move in Every DILR Set

Most of what you do in a DILR set is reversible. You try a value, it fails, you back out, and the set is exactly where it was. Cheap, and it is why exploring feels safe.

Then there is the irreversible kind. You commit to a case and start building on it. You fill the grid in pen instead of in your head. You throw away a branch because it looked wrong. These do not back out cleanly. If they were wrong, you do not return to where you were; you return to something worse, with four minutes gone and a table you can no longer trust.

A turnstile is the honest picture. It lets you through, and then it does not let you back. Every DILR set has one, and the difference between a set you finish and a set you abandon is usually whether you noticed which move it was.

Why Some DILR Moves Cannot Be Undone

Reversibility in a caselet is not about the logic, it is about the paper. A deduction is reversible because it was never optional: it followed from the clues, so it stays true whatever else happens. A choice is different. Everything you build on it inherits it, and unwinding means identifying every downstream entry and removing exactly those.

The cost of unwinding grows in a particular way:

• One step down, you simply rub out the last entry.
• Three steps down, you have to reason about which entries were independent.
• Six steps down, you cannot reliably tell, so the safe move is to redo the whole branch.
• Past that, redoing costs more than the set is worth, which is the point at which aspirants abandon.

That is straightforward for two steps and impossible for eight, because by then you cannot remember which entries came from the assumption and which were independent.

The moves that turn out to be doors are consistent across sets:

• Committing to a case without recording where the branch started.
• Erasing eliminations to tidy the grid, destroying the reasoning that produced them.
• Discarding a branch on a feeling rather than a contradiction.
• Merging two grids into one when a case split should have kept them separate.

The One-Way Door: Three Tests Before You Commit

Each test takes a couple of seconds and they run in order. The first one that answers badly means you are at a door.

The thirty-second test is the most practical, because it is about your notation rather than the logic:

• If every entry from this point carries a branch marker, unwinding is trivial and this is not a door.
• If entries will be indistinguishable from earlier ones, it is a door regardless of how confident you are.
• If you would have to redraw the grid to undo it, it is a door and an expensive one.

Overwriting versus adding catches the doors that do not feel like decisions at all:

• Erasing is always overwriting. Striking through is adding.
• Replacing an option list with a single value overwrites the alternatives you eliminated.
• Tidying a messy grid destroys state, which is why the tidy-up so often precedes the collapse.

Choosing versus deducing is the test aspirants fail most often, because a strong likelihood feels like a conclusion:

• Ask which clue forces it. If you cannot name one, you are choosing.
• "It has to be this, otherwise nothing works" is a choice until you have actually shown nothing works.
• The most dangerous choices are the ones that turn out to be right, because they teach you the habit is safe.

How to Walk Through a Door Safely

The answer is not to avoid case splits. Many CAT sets cannot be solved without one. The answer is to make the door two-way before you step through it.

Every entry in the right-hand column costs a few seconds and converts a permanent decision into a temporary one. That conversion is the whole skill, and it is entirely mechanical.

When to Open the Door Early

There is a timing question underneath all of this. Splitting into cases too early multiplies work; too late and you stall. A few reliable guides:

• Split when the two-option slots stop producing. If no further deduction follows from the constraints, a case split is the only way forward.
• Split on the most constrained slot, not the first one you notice. Fewer branches, and each one resolves faster.
• Split on two options, never three. If your candidate has three, look for a different slot; three-way splits rarely finish inside the time.
• Split early rather than late once you have decided to. A split at minute four with six minutes to test is worth far more than one at minute eight.
• Do not split while the untouched list is non-empty. Unused information is cheaper than assumed information, every time.

Recognising when a set has genuinely stopped yielding, rather than when you have stopped looking, is covered in why every impossible DILR set is actually waiting for one observation.

Common Mistakes Around Irreversible Moves

The habit shows up in several forms:

• Erasing rather than striking through. Destroying eliminations you will have to redo.
• Unmarked case splits. Assuming without recording where the assumption began.
• Three-way splits. Opening more branches than the clock can test.
• Abandoning a branch on feel. Dropping a case without a contradiction, then being unable to justify returning to it.
• Redrawing on a fresh page. Copying the conclusions and leaving the reasoning behind.

A Practice Drill for Spotting Doors

The drill trains recognition of the commitment point, which is the part that happens too fast to notice in real time.

1. Take eight DILR sets and solve them normally, but draw a line every time you make a move that is not forced by a clue.
2. Afterwards, count the lines. Most aspirants are surprised by how many there are.
3. For each line, check whether you could have unwound it in thirty seconds. Mark the ones you could not.
4. On the next eight sets, convert every marked case into a recoverable one before proceeding.

What usually emerges:

• Aspirants make two to four unforced moves per set and consciously notice about one.
• Most doors are opened in the first four minutes, when confidence is highest and the grid still looks manageable.
• Sets that ran out of time almost always contain an unmarked branch made before minute five.
• Once branches are marked, the number of sets abandoned drops before accuracy on individual questions moves.

The Bottom Line

DILR punishes commitment without a record, not commitment itself. Every set asks you to choose something eventually. The only question that matters is whether the choice you made is one you can still walk back from at minute nine, and that is decided entirely by how you wrote it down.

The One-Way Door, Recap

• Thirty-second test: if unwinding takes longer, it is a door.
• Overwrite versus add: adding is safe, overwriting destroys state.
• Choosing versus deducing: name the clue that forces it, or you are choosing.

Build the Habit on Timed CAT DILR

Marking branches is the first discipline to go when the clock is running, so it has to be built against one. Work through CAT previous year questions in timed blocks, or sit full CAT mock tests and past papers so the habit survives real fatigue. More DILR method sits in the CAT DILR blog archive, and if sets keep becoming unrecoverable it is worth having your approach reviewed honestly.

The Optima Learn Editorial Team builds CAT preparation content from exam-pattern analysis and Optima Learn's adaptive practice data. This guide is part of our CAT DILR preparation series.