How to Think Like a Mathematician: Problem Solving Strategies

What do you do when you're staring at a problem you've genuinely never seen before?

Not a repeat of yesterday's homework. Not the same question with the numbers swapped. An actual dead end, the kind where flipping through your notes looking for "the formula" gets you nowhere.

Here's something professional mathematicians don't always say out loud: they hit that wall constantly. What gets them through isn't a bigger formula cabinet. It's a small set of moves, the same handful, that turn an unfamiliar problem into a familiar one.

We're going to learn four of them. Each one is a way of looking at a problem sideways until it cracks open.

You don't need more formulas, you need better moves

Think about the last hard problem you actually solved, in math or anywhere else. Chances are you didn't just recall the answer. You did something to the problem. You drew it out, you tried a smaller version of it, you ran it backward, or you noticed something about it that refused to change no matter what you tried.

Those four things, drawing it (re-representing), shrinking it (small cases), reversing it (working backward), and watching for what stays fixed (invariants), aren't tricks reserved for geniuses. They're habits. Ordinary moves you can practice on any problem, before you even know which branch of math it belongs to.

Let's build each one from scratch, with something you can actually play with.

Move one: turn the algebra into a picture

Here's a classic claim: add up the first $n$ odd numbers, $1 + 3 + 5 + \dots + (2n-1)$, and you always land on a perfect square, $n^2$.

You could check this with algebra, maybe induction, maybe a summation identity. But here's the thing, there's a way to see that it's true without proving anything symbolically.

Picture a grid. The first odd number, 1, is a single cell, a 1×1 square. Wrap the next odd number, 3, around it as an L-shape, and you get a 2×2 square. Wrap 5 more around that, and you get a 3×3 square.

Each new odd number is exactly the right size to wrap the previous square into the next one up. That's not a coincidence, it's forced by the geometry. The picture is the proof.

Drag $n$ up and watch the L-shaped layers, mathematicians call them gnomons, stack into a bigger square every time. Hit play and watch it build itself. The running sum and $n^2$ never disagree, because they're the same picture, counted two different ways.

Once the picture convinces you, the notation is just bookkeeping:

That's move one. When an equation feels abstract, ask what it would look like drawn out. Numbers, sets, even probabilities usually have a picture hiding inside them.

Move two: when the problem feels too big, shrink it

Try this one: a room has $n$ people in it, and everyone shakes hands with everyone else exactly once. How many handshakes happen in total?

Don't try to answer that for, say, 40 people, straight in your head. That's the mistake. Shrink the problem first.

With 2 people, there's 1 handshake. Add a third person, they shake hands with both people already there, so now it's $1 + 2 = 3$. Add a fourth, they shake hands with all 3 already in the room, so it's $3 + 3 = 6$.

...notice the pattern? Every time you add one more person, they shake hands with everyone already there. So the total for $n$ people is $1 + 2 + 3 + \dots + (n-1)$.

Slide $n$ up and watch the lines multiply. Hover any single dot and you'll see exactly $n - 1$ lines coming out of it, because that person shakes hands with everyone but themselves. Multiply that by $n$ people and you've counted every handshake twice, once from each end, so we divide by 2:

That's move two. A big, unfamiliar problem is usually a small, familiar one wearing a trench coat. Work out $n = 2, 3, 4$ by hand, watch what changes from one to the next, and the general rule is often already sitting there.

Move three: start at the finish line and walk backward

Some problems are hard forward and trivial backward. Here's one: I'm thinking of a number. I multiply it by 2, add 9, divide by 3, then subtract 4. I end up with 11. What number did I start with?

You could set up an equation and solve for $x$. But watch what happens if you just run the whole chain in reverse.

The last thing that happened was "subtract 4," so undo that first: add 4 back. The step before that was "divide by 3," so undo that too: multiply by 3. Keep walking backward, undoing one operation at a time, and you land on the starting number with zero algebra.

Set a target and hit play. Watch the value travel right to left through the chain, inverting whatever operation sent it forward at every box: $\times$ becomes $\div$, $+$ becomes $-$. By the time it reaches the far left, you're holding the original number, not because you solved for it, but because you retraced the steps.

This works anywhere the forward process is a sequence of clean, reversible steps: unscrambling a cipher, backing out of a maze, even debugging code by running the failure backward to its cause. If forward is a chain of steps, backward is that same chain, flipped.

Move four: find the one thing that refuses to change

Last one, and it's the strangest of the four. Take a 6×6 board, 36 squares colored like a chessboard, and cut off two opposite corners. Can you cover what's left using dominoes, where each domino covers exactly two neighboring squares?

Before placing a single one, notice something about the coloring: it alternates. Every domino you place, anywhere on the board, covers exactly one black square and one white square, because neighboring squares are never the same color.

Now count. The two corners we cut off are the same color (opposite corners on an even board always are). So we didn't remove one black and one white, we removed two of the same color. The board now carries an imbalance that no arrangement of dominoes can ever undo, because every domino keeps that imbalance exactly the same.

Go ahead, try to tile it. Click two neighboring empty squares to place a domino. Watch the black-remaining and white-remaining counts as you go, their difference never moves. That gap of 2 was baked in the moment we cut the corners, and it survives no matter what you try.

That unshakeable quantity, the color difference, is called an invariant. Whenever a problem asks "is this even possible?", hunting for something that stays fixed no matter what moves you make is often faster than trying every move by hand.

Stacking the moves

Real problems rarely announce which move to use. Part of thinking like a mathematician is just trying one, and if the picture doesn't help, shrinking the problem, and if that doesn't help, reversing it, and if that doesn't help either, hunting for an invariant.

None of these four moves needs advanced machinery. They're available to you on the very first hard problem you ever meet, and professional mathematicians are still reaching for exactly the same four on problems that have resisted solution for decades.

The short version

Thinking like a mathematician isn't about knowing more formulas. It's about having reusable moves you can try on any unfamiliar problem: turn the algebra into a picture, shrink it to a small case and look for the pattern, walk backward from the goal, or find the thing that never changes. Master those four, and "I've never seen this problem before" stops being scary. It's just a problem you haven't picked the right move for yet.

All four visualizers above are interactive React components rendered entirely in SVG, no charting library involved. Drag the sliders, hover the dots, click the board, it's all running live in your browser.