One detail catches many dMAT candidates off guard: you are not given rough paper. The three Core subtests — Figure Sequences, Mathematical Equations and Latin Squares — are designed to be solved entirely in your head, and that single constraint changes how you should prepare. Knowing the method for each question type is necessary but not sufficient; you also need the mental capacity to hold several pieces of information at once while you work on them. That capacity is your working memory, and it is easy to overlook precisely because it never appears as a topic on any syllabus. This article explains what working memory does on the dMAT and how to train it deliberately in the weeks before your sitting.
Why working memory matters on the Core module
Working memory is the small mental workspace where you hold and manipulate information for a few seconds — the numbers you are carrying, the rule you have half-worked-out, the option you have provisionally ruled out. The Core subtests lean on it heavily, and the no-scratch-paper rule described on the syllabus is what makes it a limiting factor rather than a convenience.
Consider each subtest in turn. In Mathematical Equations you often have to keep two or three partial results in mind while you substitute one into another; lose one and you start again. In Latin Squares you are tracking which letters remain possible in several cells at once, updating that picture every time you place a new one. And in Figure Sequences you have to hold a candidate transformation rule in mind while you test it against the next figure in the matrix. In all three, the method is only as good as your ability to keep its moving parts from slipping away mid-solve.
This is why two candidates who both "know how" to solve a question type can perform very differently under exam conditions. The one whose working memory is stretched loses accuracy not because the method failed, but because a piece of it fell out of their head at the wrong moment. Training that capacity is therefore not a separate skill bolted on for its own sake — it is what lets your methods actually hold together when you cannot write anything down.
How to train it without gimmicks
Be wary of the promise that a single "brain-training" game will transform your test scores. The evidence for such transfer is modest, and the official sources make no such claims. The reliable approach is far more direct: practise the actual thing you need to do, which is to reason through Core-style problems entirely in your head.
A few habits do most of the work. First, practise under the real constraint from the start — put the pen down. It is tempting to jot intermediate steps while you learn a method, but if you always lean on paper you never build the capacity to work without it. Once you understand a method, drill it mentally so that the exam feels like more of the same, not a sudden new restriction.
Second, build the habit of chunking. Working memory holds only a handful of items, so the trick is to make each item carry more. Grouping a messy calculation into a couple of familiar landmarks, or seeing a figure transformation as one named move rather than three separate changes, frees space for everything else. Much of what looks like a better memory is really better organisation of what you are trying to remember.
Third, train little and often. Short, frequent sessions of mental problem-solving build this capacity more effectively than occasional marathons, and they are less draining. Five focused minutes of head-only arithmetic on most days does more than an hour once a week. If sustaining concentration across a full sitting is your real limit, the separate matter of building study stamina is worth reading alongside this.
Fold it into practice you are already doing
The good news is that you do not need a separate training programme. Every time you sit down to work on a Core subtest, you can also be training the working memory it depends on — simply by refusing the crutch of paper and noticing when your head feels overloaded. Those moments of overload are useful data: they point to the exact step where your method has too many loose parts, which is usually a sign to chunk more aggressively or to rehearse that step until it becomes automatic.
Weave this into your normal routine rather than treating it as an extra chore. When you follow a broader plan such as the Core module strategy guide, let "solve it in my head" be the default mode for every drill. And when you take a timed run on the free dMAT practice test, pay attention not only to what you got wrong but to where your mental workspace felt full — that is where a little targeted training will pay off most.
A quiet advantage, built gradually
Working memory is not something you fix in a weekend, and it is not something the exam ever names. But because the Core subtests are solved mentally, the candidate who has trained to hold and manipulate information cleanly has a real, quiet advantage over one who has only memorised methods. Practise head-only, organise what you hold into fewer, larger chunks, and keep the sessions short and regular. Do that consistently, and by test day the constraint that once caught you off guard will feel like just another part of how you work.
