GridRecall

What you see

The screen has two parts: the memory grid (3 by 3 up to 5 by 5) and a figure card beside it (above it on a phone), with Symmetrical and Not symmetrical buttons. Press Start (or Space); after a short 3, 2, 1 countdown the trial runs in beats, one per cell to remember:

1. A cell lights for 800 ms. One cell of the grid turns on, then off. Nothing names it on screen: you remember it as a place. A line under the board counts the cells, for example "Cell 2 of 4", and then the figures, for example "Figure 2 of 4".
2. A figure appears on the card. It is a 6 by 6 block pattern with a dashed line down the middle. Decide whether its left half mirrors its right half exactly, and press Symmetrical (Left arrow) or Not symmetrical (Right arrow). While you judge, the grid is dimmed and shows nothing. From level 5, two figures follow each cell, one after the other.
3. The next cell lights, and so on until the last cell and its figures are done.
4. Recall. The grid comes back with no cells lit. Click every cell that lit up; click a marked cell again to unmark it. You can also move with the arrow keys and mark with Space. On levels where order counts, the question reads "Which cells lit up, in order?" and each marked cell shows the number of its click (1, 2, 3...). Press Check answer (or Enter) when done. The verdict gives the cells you found and the figures you judged right, and the grid marks each cell: green for a lit cell you marked, red for a marked cell that never lit, and a dashed outline for a lit cell you missed.

The figures wait for your answer: there is no clock on them. The lit cells do not wait: each is on for exactly 800 ms.

The figures

Every symmetrical figure is built by mirroring random blocks across the dashed axis. Every non-symmetrical figure starts life as a symmetrical one, then has a few blocks on its left half flipped, which breaks exactly that many mirror pairs. How many depends on the level: four broken pairs at levels 1 and 2 (easy to spot), three at levels 3 and 4, two at levels 5 to 7, and just one at levels 8 to 10. Within each trial, half of the figures are symmetrical (if the count is odd, one more is symmetrical).

How the cells are chosen

• A cell never lights twice in one trial.
• No three lit cells ever sit in a straight, touching line (across, down or diagonal), so you cannot store three cells as one easy stroke.
• On levels 1 to 5, consecutive cells never touch, not even at a corner. From level 6 that restriction is lifted and the next cell may be a neighbour of the last.

The ten levels

Practice starts at level 3. A perfect trial (every cell recalled, in order where order counts, and every figure judged correctly) moves you up one level. Anything less, as long as something was right, holds your level. A trial that scores zero moves you down one level.

Levels 8 and 9 use identical settings. At level 10 you hold seven positions in order through fourteen symmetry judgements.

What it measures

GridRecall is a complex span task, the family of "storage plus processing" tests that began with Daneman and Carpenter's reading span in 1980, where people read sentences aloud while remembering the last word of each. Turner and Engle's operation span swapped reading for arithmetic, and Kane and colleagues' symmetry span made both halves spatial: people made symmetry decisions while trying to remember square locations. GridRecall follows that symmetry span design closely.

Complex span tasks are consistently linked to reasoning, reading comprehension and maths achievement. The interruption is the point. Simple spatial span, as in the Corsi block test, has nothing between the items; putting a processing task there forces you to protect the stored positions against interference, which is the controlled-attention ability that tracks fluid intelligence. The automated operation span task, a standard lab version of this design, showed good internal consistency (.78) and test-retest reliability (.83) in its validation study.

In our taxonomy GridRecall feeds two memory facets at once: span capacity and working under load. Its competency tags are 70% working memory and 30% spatial visualisation.

How it is scored here

Each trial combines two parts, scored out of 10,000 points:

Score = 70% recall + 30% figure accuracy

Figure accuracy is simply the share of figures you judged correctly.

Recall depends on whether order counts.

• Levels 1 to 5 (order does not count): recall = (correct cells − wrong cells) ÷ cells shown, never below zero.
• Levels 6 to 10 (order counts): recall = (the longest run of your marks that follows the true order, even with gaps, − wrong cells) ÷ cells shown, never below zero. A cell marked out of order breaks the run.

A wrong cell is a marked cell that never lit. On every level, every wrong cell cancels a right one.

A trial counts as correct, and moves you up, only when both parts are perfect. Anything less is capped at 9,999.

Some worked numbers:

Notice the second and third rows, and the last one. On every level, adding a cell you are unsure of only helps if you are more likely right than wrong; a wrong guess wipes out a correct cell.

From trials to your 0–100 score

Time is not scored. Each trial is weighted by its measured difficulty, which counts the number of cells, plus extra weight for a bigger grid, for two figures per cell and for order counting. On our 0–100 scale, 50 means you scored what the difficulty you work at predicts.

Your report shows:

In the trial-by-trial list, a trial with under 85% of figures right is marked as not a clean reading of memory.

Where you will meet it

This is the memory game you are most likely to meet in a real assessment, because it is the design behind the grid challenges in several game-based batteries.

• Aon lists gridChallenge as a nine-minute task that "measures your ability to keep something in mind while simultaneously trying to process other information". Preparation sites describe its mechanics as remembering the positions of dots on a grid while answering short spatial questions (symmetry, rotation, or whether two grids combine into a third) between them, then recalling all the positions.
• Capgemini: Indian preparation sites report a Grid Challenge among the games in its game-based aptitude round.
• Procter and Gamble: preparation sites publish practice material for a "P&G Grid Challenge" with the same alternating memory and spatial questions.

The vendor versions differ in detail: dots rather than cells, rotation questions as well as symmetry, a total time limit. GridRecall is a close cousin, not a copy. The habits it builds (fast, accurate interruption answers and a stored path you rehearse between them) are the ones those games reward.

The method

1. Decide your code before you press Start

   Choose how you will hold positions: as a path (a shape you draw through the cells in order) or as names (row and column, such as "3-1" for row 3, column 1, which you work out yourself). A path is quick on small grids; if paths start to blur on 5 by 5, switch to names, or use both.

2. Grab each cell in its 800 ms

   When a cell lights, look straight at it and add it to your path, or say its short name once. Link it to the previous cell: "then down-right to the corner".

3. Judge the figure quickly and correctly

   Fold the figure along the dashed line in your mind and compare mirror pairs row by row: outermost columns first, then middle, then innermost. One mismatched pair means Not. Answer as soon as you are sure; every second on the figure is a second your positions are fading.

4. Rehearse once between jobs

   After answering, before the next cell appears, replay your path or names once from the start. On two-figure levels, do this after the second figure. Keep it to a single fast pass.

5. Recall in the order you stored

   Mark cells in the order they lit, even on levels where order does not count; your memory is organised that way. On order levels, the numbers on your marked cells let you check the sequence before you press Check answer.

6. Decide about uncertain cells by level

   On every level, leave out a cell you think is more likely wrong than right, since a wrong cell cancels a right one. Enter a guess only when you think it is more likely right than wrong; on order levels it must also go in the right place in the order to add anything.

Worked example 1: level 4, order does not count

Four cells on a 4 by 4 grid, one figure after each. Picture the positions below as they light, one at a time, with a figure between each.

Show the solution

Cells, using the caption's naming (row first): 1-2, 3-4, 4-1, 2-1. As a path: top edge near the left, across and down to the right edge, sweep to the bottom-left corner, then back up the left edge. Notice that no two consecutive cells touch, as level 4 guarantees.

●●●●

Scoring: marking all four with every figure right is 10,000 points and a level up. Marking three of them plus one wrong cell gives recall of (3 − 1) ÷ 4 = 0.5, and 6,500 points.

Worked example 2: level 6, order counts

Five cells on a 4 by 4 grid, two figures after each. From this level consecutive cells may touch.

Show the solution

Names: 1-1, 2-2, 1-4, 4-3, 3-1. As a path: start top-left corner, step diagonally in, jump to the top-right corner, drop to the bottom row third column, then back to the left edge, third row. A short story helps too: "corner, step in, far corner, floor, back left".

13254

Tips and traps

Levelling up

• Levels 1 to 4. Build the path habit and a reliable symmetry scan. Grids grow from 3 by 3 to 4 by 4.
• Level 5. Two figures after each cell doubles the interruption. Rehearse after the second figure, not between them.
• Level 6. Order starts counting, five cells, neighbours allowed. Paths become essential.
• Level 7. The 5 by 5 grid arrives. If your picture of the path blurs, try names ("3-1") instead.
• Levels 8 to 10. Six, then seven cells in order, and foils with a single broken pair. The symmetry scan must be systematic, and the rehearsal pass must be quick.

1. Week 1
   • Two short blocks per session
   • Pick path or names and use it on every trial
   • Practise the row-by-row symmetry scan until you never miss a four- or three-pair foil
2. Week 2
   • Aim to hold level 5 with two figures per cell
   • Check your report: if the neglect gap is large, slow down on the figures
   • Add one Digit Span block for ordered recall
3. Week 3
   • Work at level 6 and above: order counts
   • Switch to names on 5 by 5 grids if paths blur
   • Finish each session with one Flashback block to train updating

Check yourself

5 practice questions

1. Level 4, order does not count. Four cells lit. You mark three correct cells and one cell that never lit, and judge every figure correctly. What is the trial score?
2. Level 7. The true order is A B C D E and you mark A C B D E with every figure right. What happens?
3. At level 9, how does a non-symmetrical figure differ from a symmetrical one?

and 2 more

Create a free account to answerSign in

Sources

1. Daneman, M., and Carpenter, P. A. (1980). Individual differences in working memory and reading. Journal of Verbal Learning and Verbal Behavior, 19, 450–466. doi:10.1016/S0022-5371(80)90312-6
2. Turner, M. L., and Engle, R. W. (1989). Is working memory capacity task dependent? Journal of Memory and Language, 28, 127–154. doi:10.1016/0749-596X(89)90040-5
3. Kane, M. J., Hambrick, D. Z., Tuholski, S. W., Wilhelm, O., Payne, T. W., and Engle, R. W. (2004). The generality of working memory capacity: a latent-variable approach to verbal and visuospatial memory span and reasoning. Journal of Experimental Psychology: General, 133(2), 189–217. doi:10.1037/0096-3445.133.2.189
4. Redick, T. S., and Lindsey, D. R. (2013). Complex span and n-back measures of working memory: a meta-analysis. Psychonomic Bulletin and Review, 20(6), 1102–1113. doi:10.3758/s13423-013-0453-9
5. Unsworth, N., Heitz, R. P., Schrock, J. C., and Engle, R. W. (2005). An automated version of the operation span task. Behavior Research Methods, 37(3), 498–505. doi:10.3758/BF03192720
6. Conway, A. R. A., Kane, M. J., Bunting, M. F., Hambrick, D. Z., Wilhelm, O., and Engle, R. W. (2005). Working memory span tasks: a methodological review and user's guide. Psychonomic Bulletin and Review, 12(5), 769–786. doi:10.3758/BF03196772
7. Đokić, R., Koso-Drljević, M., and Đapo, N. (2018). Working memory span tasks: group administration and omitting accuracy criterion do not change metric characteristics. PLOS ONE, 13(10), e0205169. doi:10.1371/journal.pone.0205169
8. Engle, R. W., Tuholski, S. W., Laughlin, J. E., and Conway, A. R. A. (1999). Working memory, short-term memory, and general fluid intelligence: a latent-variable approach. Journal of Experimental Psychology: General, 128(3), 309–331. doi:10.1037/0096-3445.128.3.309
9. Berch, D. B., Krikorian, R., and Huha, E. M. (1998). The Corsi block-tapping task: methodological and theoretical considerations. Brain and Cognition, 38(3), 317–338. doi:10.1006/brcg.1998.1039
10. Aon. Prepare for your online assessment (gridChallenge). aon.com
11. JobTestPrep. P&G Grid Challenge practice guide (candidate-reported mechanics). jobtestprep.com
12. PrepInsta. Capgemini Exceller game-based aptitude test questions (candidate-reported). prepinsta.com

Put it into practice: GridRecallSign in to try it