DILR Without Drawing a Table
Building a table is the usual default DILR move, but not every set is grid-shaped. This piece introduces the Container Question: checking whether a set is really a grid, a chain, or a web before choosing a notation.
DILR Without Drawing a Table
When building a table is actually the wrong first move for a DILR set
Not every CAT DILR set is a grid. Some are strict sequences that move left to right, step by step. Others are tangles of pairings that never touch a row or a column in the first place.
A table earns its place on the first shape. Force it onto the other two, and you lose time drawing something the set never asked for.
We've watched this happen in review sessions more times than we can count. An aspirant draws a clean grid, fills two rows with confidence, then stalls. The third clue simply refuses to fit into any cell.
The set was never a grid to begin with. It was four people connected by who trusts whom, better drawn as dots and lines than as boxes.
- A table only earns its place when a set is genuinely grid-shaped, several attributes intersecting.
- Chain-shaped sets (a strict order) and web-shaped sets (pairings and connections) waste time inside a table.
- The Container Question checks a set's actual shape, grid, chain, or web, before you pick a notation.
- This is about matching the shape, not about skipping tables altogether or timing when you fill one in.
Why the table becomes the reflex move
Most DILR prep starts the same way. Read the set, draw a grid, fill it in as the clues arrive.
It's taught early because it's simple to teach, and because a good share of introductory CAT DILR previous year questions really are grid-shaped: several attributes belonging to several entities. So the habit sticks, and it doesn't notice once the sets stop being grids.
By the time real CAT-level sets show up, plenty of them are built around a strict sequence or a web of relationships instead. The table habit doesn't check for that. It reaches for the same six-by-six grid no matter what's actually in front of it, and you end up with empty cells, crossed-out guesses, and eight minutes gone before you've placed a single fact correctly.
Sound familiar? It should. It's less a skill gap and more a reflex nobody ever asked you to check.
The Container Question: three shapes to check
This is a different question from going gridless on principle. One related post, the DILR Gridless Method, argues for skipping tables altogether. Another, the Empty Grid Advantage, argues for delaying when you fill a table in.
Both make good points. Neither is what's happening here.
The Container Question isn't about avoiding tables, or about timing when you fill them in. It's about checking what shape a set actually is, grid, chain, or web, before you pick any notation at all.
The Container Question
A table is one shape. Not every set is that shape.
- Is It a Grid?: multiple independent attributes across multiple entities that genuinely intersect. A table earns its place here.
- Is It a Chain?: a strict order or sequence relationship, where a simple left-to-right line captures everything a table would, minus the empty cells.
- Is It a Web?: relationships between entities that don't reduce to rows and columns at all, who's paired with whom, who knows whom, better drawn as connected points than as a grid.
These three checks run in parallel, not in order. You're not working through the list top to bottom during the exam. You're glancing at the clues for a few seconds, asking all three at once, and going with whichever one clearly matches what you just read.
Grid, chain, or web: telling them apart fast
The fastest tell is what the clues sound like, not what the passage is about. Two sets built around the exact same topic, say five colleagues on a project, can end up shaped completely differently. It comes down to whether the clues describe attributes, order, or connections.
| Shape | What the clues sound like | Best notation |
|---|---|---|
| Grid | Several separate facts about several people or items: age, city, rank, score. | A table, entities as rows, attributes as columns. |
| Chain | Mostly "before," "after," "immediately following," "between." | One left-to-right line of entities in order. |
| Web | Mostly who is paired with, connected to, or opposite whom. | Dots for entities, lines for the stated connections. |
A chain-shaped set, worked through
Take a set built around six events, A through F, with clues like these: C comes right after A. B happens somewhere before D. E is the last event. F never sits immediately next to D.
Every one of those clues is about relative position. None of them hands you a second attribute to track, no color, no city, no score. That alone is the tell that this is a chain, not a grid.
Skip the table. Draw a single line with six blanks instead.
Anchor E at the last blank first, since that's fixed outright. Lock A and C together as one pair, A immediately followed by C, and slide that pair into positions three and four. Put B ahead of D, and keep F out of the slot right next to D.
One arrangement satisfies every clue at once, and the whole line clicks into place: B, F, A, C, D, E.
Four clues, one line, and every clue lands exactly where it's stated. No invented columns for an attribute this set never had in the first place.
Practice shape recognition on real sets
Run the Container Question on full-length, timed CAT DILR sets, not just the example above.
Practice CAT DILR SetsBottom line
None of this makes the table a bad tool. Give it a truly grid-shaped set, several attributes intersecting across several entities, and it earns its place completely. Hand it a chain or a web instead, and the same table that would have saved you time spends it instead.
Before you draw anything on your next DILR set, run the Container Question. Is it a grid, several attributes truly intersecting? Is it a chain, one clean order end to end? Or is it a web, pairings that were never going to sit inside rows and columns? Only one of those three answers earns you a table.
Test this on your next DILR set
Before you draw a single grid, run the Container Question on a real, timed CAT DILR set.
Start Practicing DILRFrequently Asked Questions
Should I stop using tables for DILR sets altogether?
No. A genuinely grid-shaped set, several attributes intersecting across several entities, is still best solved with a table. The point is checking the shape first, not avoiding tables on principle.
How can I tell a chain-shaped set from a grid-shaped one quickly?
Read the clues once and ask whether they mostly describe order and sequence, chain, or mostly describe attributes belonging to entities, grid. A set built almost entirely from 'before,' 'after,' and 'between' clues is usually a chain.
What does a web-shaped set actually look like on paper?
Instead of rows and columns, it's dots representing entities with lines connecting the ones that are related, similar to a simple relationship diagram. Trying to force this into a table usually creates rows and columns that don't correspond to anything real.
Is this different from the gridless method already covered on this blog?
Yes. That post argues for skipping grids generally. This one is narrower: identify whether a set is actually a grid, a chain, or a web first, then pick the notation that fits, which sometimes is a table and sometimes genuinely isn't.
Solve real CAT DILR sets timed
Hand-picked LR puzzles and DI caselets with timer + solution breakdown.