Home · Resources · Visual search

Visual search: dot sequences and letter columns

Two search papers that look mechanical and are not, and then a third part that runs both together. The third part is the one being read most closely.

standard Stage 2 6 min read

Two tasks, then both at once:

The dot sequence. A dot steps around the corners of a square, one position at a time, square after square. You mark every square where the sequence breaks — a skipped corner, a step backwards, or no movement at all.
The letter columns. Each box holds two short columns of letters. You mark the box if the columns contain different letters. Order does not matter, and a letter can appear twice.
Both. Pairs presented together, auto-advancing, with the dot sequence as the primary task.

Dot sequence: About a second a square
Letter columns: About two seconds a box
Combined: About four seconds a pair, auto-advancing
Scoring: Hits, misses and false marks, all three

The dot sequence is a chain, not a list

This is the thing to understand and most people never state it: each square can only be judged against the one before it. The task is not "look at this square" but "is this square the correct successor of the last one?"

Three consequences follow, and they are the whole technique:

Carry exactly one thing forward: the last corner. Not the sequence so far, not the direction, not a count. One position, refreshed every square.
Know which way round this run goes, and confirm it on the first two squares. Runs alternate direction. Half the errors people make in the opening ten squares are the previous run's direction still running in their head.
After a break, re-anchor on the square you are looking at. The next square is judged against the broken one, not against where the sequence "should" have been. Candidates who keep tracking the ideal sequence produce a cascade of false marks after every genuine break.

The letter columns are a comparison, not a reading

Compare as sets, and do not read the letters as words. Look for a letter in one column that has no partner in the other. That is the only question.
Order does not matter, so ignore position entirely. People lose items by matching top-to-top and bottom-to-bottom out of habit.
Duplicates are legal. The same letter twice in one column is not, by itself, a difference.
Lookalikes are the trap. Some runs use letters that look similar on purpose. When a run feels harder than the last one, it is, and the answer is to slow by a fraction rather than to squint harder.

What the combined part really measures

Everybody's score drops in the third part. That is expected, and the drop itself is the measurement: how much accuracy survives when a second stream arrives, and whether you shed it gracefully or collapse.

The plan that works is the same as on TEA-Occ: protect the fragile task. The dot sequence is the chain, so it breaks badly if you drop it and cannot be rebuilt. The letter comparison is independent box to box, so a slower rate there costs you a little and nothing more. Give the sequence the priority and let the comparison run at whatever speed is left.

Drill it: the three visual search tasks

Each one separately, then the combined paper at four seconds a pair. Exam mode takes the paper away when the clock hits zero, which is the part that changes how you work.

Open the drills

Where people lose it

Losing the chain and hiding it. If the dot sequence gets away from you, take the next square as your new anchor immediately rather than trying to reconstruct.
Marking every square after a break. The classic cascade. One break is one mark.
Reading letters instead of matching them. Slower channel, same answer.
Practising on a phone. These two tests in particular are flattered by a small bright screen held close. Sit them on a laptop or a desktop, at arm's length, like the real one.
Treating the combined part as a warm-up. It is the finale, and it is where the informative variance is.

How to practise it

Drill the dot sequence alone until the "carry one corner" habit is automatic. Everything else depends on it.
Then the letter columns alone, until you have stopped reading them.
Then the combined paper, timed, twice a week. Compare your dot-sequence accuracy alone against your dot-sequence accuracy in the combined part — closing that gap is the actual goal.