Getting faster without starting to guess
Nobody ever got faster at these by trying to go faster.
Every reasoning test is timed, so the obvious way to improve is to work more quickly. It is also the one approach that reliably does not work. Rushing converts careful solving into guessing, and guessing has a known ceiling — on a five-option paper it is twenty per cent.
Speed on these puzzles is real and it is trainable. It just does not come from where people look for it.
Where the time actually goes
Watch someone solve a matrix puzzle slowly and the time divides into three very unequal parts:
- Figuring out what kind of rule this is — usually most of the time
- Working out the specifics of that rule — much less
- Checking the options — least of all
Fast solvers are not faster at the second and third parts. They are faster at the first, and often skip it entirely, because they recognise the rule family on sight rather than deriving it. That recognition is the thing worth training, and it is built by naming rules, not by hurrying.
Name the rule on every puzzle, including the ones you get right
This is the single highest-value habit and almost nobody does it, because when you get a puzzle right there is no obvious reason to go back.
But there is a large difference between two students who both answered correctly: one can say 'progression on count, constancy on shape' and one cannot. The first has added a recognisable pattern to memory. The second has had a nice feeling and learned nothing, and will spend the same thirty seconds on the same rule next week.
If you got it right but cannot name the rule, you guessed well. Guessing well does not compound.
Practise slowly, on purpose
Counter-intuitive and well supported: to get faster at a recognition task, practise it slowly and correctly. Thirty puzzles done carefully, each rule named and each error diagnosed, produce more improvement than two hundred done at speed.
The reason is that speed practice rehearses whatever you already do, including the bad habits. Slow practice with naming builds the rule library that makes speed possible later. You cannot recognise a rule family you have never explicitly identified.
This is the same shape as learning to read. Nobody gets faster at reading by being told to hurry. They get faster because words that once had to be decoded become words that are simply recognised — and that only happens by encountering them properly first.
The checklist gets faster, it does not get abandoned
Running the attribute checklist — count, position, orientation, shading, shape, size — feels slow for the first few sessions. It is slower than staring, initially.
Within a couple of weeks it stops being a list you work through and becomes a single sweep. You are still checking all six things; you are no longer conscious of doing it one at a time. That is what expertise in this kind of task looks like, and it is why abandoning the checklist because it feels slow is exactly the wrong move at exactly the wrong moment.
Triage is a skill of its own
A large part of finishing a paper on time is not solving faster but choosing better. Three rules that consistently pay:
- Set a per-question limit before you start — ninety seconds is typical — and obey it. When you hit it, eliminate what you can, choose, and move on.
- Mark what you abandoned. Coming back with fresh eyes to a puzzle you already half-understand is very efficient. Coming back to a puzzle you have to re-read from scratch is not.
- Do not save the easy ones for later. Later may not arrive. Marks are usually equal, so take the cheap ones first.
What to do in the last two minutes
Answer everything. On a paper with no penalty for a wrong answer, a blank is strictly worse than a guess — it has a zero per cent chance instead of twenty.
If you have eliminated two options, the remaining guess is a one-in-three rather than one-in-five, which over a whole paper is a meaningful number of marks. Elimination is worth doing even when you cannot finish the reasoning.
A practice session that works
- Do ten puzzles with no timer at all.
- For each, write the rule in a sentence before checking the answer.
- For the ones you got wrong, find the exact step where your rule broke. Not the right answer — the point of failure.
- For the ones you got right, check your named rule matches the real one. Sometimes a right answer comes from a wrong rule, and that is worth catching.
- Only once that is comfortable, start timing. The speed will already be there.