The coordinate vector is the most useful idea in a first linear algebra course, because it turns geometry into arithmetic. A basis B is a set of building block vectors, and the coordinate vector [x]B is the recipe for rebuilding x from those blocks: the list of weights, in order, that makes the combination equal x. Find the weights and you have found the coordinate vector. Everything else in this guide is technique for finding weights reliably, plus the verification step that makes the answer certain. One scope note before starting: this is the mathematician's meaning of coordinates, weights relative to a basis, a cousin of the map coordinates elsewhere on this site. The explorer above works with map pairs; use it to notice that a pair is also just two numbers whose meaning depends on the frame they are read in, which is exactly the idea this page formalizes.
Home / Coordinates and Location Formats
How to Find a Coordinate Vector (Linear Algebra): The [x]B Method, Worked

Photo: Monstera Production / Pexels
Coordinate Explorer: A Pair Is a Recipe
Click a point and read its two numbers. In linear algebra those numbers are weights: how much of each basis direction rebuilds the point. This explorer uses the standard basis; the article teaches every other one. It teaches the method and does not solve systems.
Map data © OpenStreetMap contributors. Coordinates are read from the map itself, in the WGS 84 system your phone uses.
The Definition, in One Equation
Let B = {b1, b2} be a basis for the plane (the same idea works with more vectors in higher dimensions). The coordinate vector of x relative to B is the list (c1, c2) such that:
x = c1 b1 + c2 b2
That is the entire definition. The coordinates are not properties of x alone; they are properties of x relative to B, and the same x wears different coordinates in different bases, just as the same place wears different numbers in different map projections. The standard basis {(1,0), (0,1)} is the special case where the recipe is the vector itself: (3,1) = 3 times (1,0) plus 1 times (0,1), so [(3,1)] in the standard basis is (3,1). Every other basis asks you to actually solve something, and the solving is a linear system, which is why this topic sits where it sits in the course: it is the first place systems of equations earn their keep.
The Method: Three Steps
Step one: write the combination. Set the given vector equal to unknown weights times the basis vectors, component by component. Step two: solve the system. Each component gives one equation; solve by substitution, elimination, or an augmented matrix. Step three: verify by recombining. Multiply your weights by the basis vectors and add; the sum must return the original x exactly. The verification is not decoration. Coordinate errors are almost always sign slips and swapped weights, and recombining catches both in seconds. Make it a reflex: no coordinate vector is finished until its recombination has been checked. The worked examples below model all three steps every time, including the check.
Worked Example One: A Symmetric Basis
Let B = {(1, 1), (1, minus 1)} and x = (3, 1). Find [x]B.
Step one: c1 (1, 1) + c2 (1, minus 1) = (3, 1). Component by component: c1 + c2 = 3 (first components), and c1 minus c2 = 1 (second components).
Step two: add the equations: 2 c1 = 4, so c1 = 2. Then from c1 + c2 = 3, c2 = 1.
So [x]B = (2, 1).
Step three, verify: 2 times (1, 1) plus 1 times (1, minus 1) = (2, 2) + (1, minus 1) = (3, 1). That is x, so the answer stands. Notice the elegance the problem setter intended: this basis points along the two diagonals, and the weights came out as small whole numbers. Textbook bases are chosen to behave; the method does not depend on their kindness.

Worked Example Two: Reading a Lecture Style Answer
Lecture notes often present the result first and the archaeology is yours. A standard example runs: with basis B = {(1, 0), (1, 1)} (a common teaching basis), a vector u is found to have [u]B = (2, minus 3). What is u itself? This is the reverse journey, and it uses step three as the whole method: u = 2 times (1, 0) plus (minus 3) times (1, 1) = (2, 0) + (minus 3, minus 3) = (minus 1, minus 3). The exercise matters because it proves you understand what the bracket notation claims: [u]B is not a different vector, it is the same u described in B's dialect. Going from coordinates to the vector is always just the combination; going from the vector to coordinates is the solve. Students who can travel both directions without hesitation own this topic, and the second direction is the one exams actually ask.
A Third Example, With a Trap Built In
Let B = {(2, 0), (1, 1)} and x = (4, 3). Step one: c1 (2, 0) + c2 (1, 1) = (4, 3), giving 2 c1 + c2 = 4 and c2 = 3. The second equation hands you c2 immediately; substituting into the first, 2 c1 = 1, so c1 = one half. [x]B = (0.5, 3). Verify: 0.5 times (2, 0) plus 3 times (1, 1) = (1, 0) + (3, 3) = (4, 3). Correct. The trap this example teaches is the assumption that coordinates should be whole numbers, or that the first weight must be found first. Coordinates are just real numbers, halves and negatives included, and a well built system often solves in whatever order the equations volunteer. Students who expect integers panic at the fraction and redo correct work; students who trust the recombination check simply note the half and move on. That trust, in the check rather than in expectations, is the professional habit this topic is really training.
The Matrix Shortcut: Multiply by the Inverse
When the basis has three or more vectors, hand solving gets long, and the matrix form takes over. Assemble the basis vectors as the columns of a matrix P. Then the defining equation becomes x = P [x]B, and solving for the coordinate vector is one multiplication:
[x]B = P inverse times x
Check it on example one: P has columns (1,1) and (1, minus 1). Its inverse is one half times the matrix with rows (1, 1) and (1, minus 1). Multiply by x = (3, 1): first entry, one half times (3 + 1) = 2; second entry, one half times (3 minus 1) = 1. Result (2, 1), matching the hand solve, as it must. The augmented matrix version is the same idea without computing an inverse: write P with x attached as the last column and row reduce until the left side is the identity; the last column is then [x]B. Use the inverse when you must convert many vectors into the same basis (compute it once, multiply repeatedly); use the augmented matrix for a single conversion. Both routes are the three step method wearing machinery, and both still end with the recombination check.
Changing Between Two Nonstandard Bases
The natural next question: given [x]B, find [x]C for a second basis C. The bridge is the standard basis, or a single change of basis matrix. Conceptually: first rebuild x from its B coordinates (the combination, using B's vectors), then solve for x in C, exactly the two journeys you already know, back to back. In matrix form, if PB and PC hold the two bases as columns, then [x]C = PC inverse times PB times [x]B, and the product PC inverse PB is the change of basis matrix from B to C. The order matters and is the classic exam trap: the matrix converts coordinates from the basis whose matrix stands on the right. Write the direction in words beside the symbols every time ("from B to C") until the convention is automatic. This is also where the map analogy repays its visit: converting coordinates between datums and projections is the same intellectual move, a change of frame with a matrix (or its grown up equivalents) doing the work, which is why convert latitude and longitude tasks feel familiar to linear algebra students.
The Mistakes That Cost Points
Four errors account for nearly every lost mark on this topic. Swapped order: the weights must be listed in the basis's given order; (1, 2) and (2, 1) are different coordinate vectors, and graders are not merciful. Rows for columns: the basis matrix P takes the basis vectors as columns; building it from rows computes coordinates for a different problem. Skipping the check: a sign slip in step two produces a confident wrong answer that ten seconds of recombination would have caught. Assuming the standard basis: if the problem names a basis, the given components of x are in the standard basis unless stated otherwise, and your answer must be in the named one; mixing the two frames mid problem is the deepest version of the swapped order error. Keep the definition in front of you, x equals the weighted sum, and every one of these errors becomes visible before it becomes permanent. For the geometric cousins of this topic on this site, how to find cartesian coordinates and how to find coordinate direction angles of a vector develop the standard frame fluency this method builds on.
Related Guides
This guide does not stand alone. The coordinates guides below take up the questions it leads into:
- How to Find Direction Angles of a Vector: Formula, Worked Examples, and the Cos Squared Check: The direction angles of a vector are the three angles it makes with the positive x, y, and z axes, found by dividing each component by the...
- How to Find the Centre Coordinates of a Circle: Four Methods, Worked in Full: The centre of a circle hides in plain sight in every form of its equation, and each representation gives a different way to read it off.
- How to Find the Coordinate Direction Angles of a Moment: Cross Product First, Then the Angles: The direction angles of a moment hide a trap that the direction angles of an ordinary vector do not: nobody hands you the moment.
Frequently Asked Questions
What is a coordinate vector in linear algebra?
It is the list of weights that rebuilds a vector from a given basis. If B = {b1, b2} and x = c1 b1 + c2 b2, then the coordinate vector [x]B is (c1, c2), in the basis's order. The same vector has different coordinate vectors in different bases.
How do I find the coordinate vector step by step?
Write x as an unknown combination of the basis vectors, turn each component into an equation, and solve the system for the weights. Then verify by multiplying your weights by the basis vectors and adding: the sum must equal the original x exactly.
What is the matrix formula for [x]B?
Build the matrix P whose columns are the basis vectors. Then x = P [x]B, so [x]B = P inverse times x. For a single vector, row reducing the augmented matrix [P | x] gives the same result without computing the inverse explicitly.
Can a vector have different coordinates in different bases?
Yes, and it always does unless the bases coincide. Coordinates describe a vector relative to a frame. The vector itself never changes; only its description does, exactly as one place has different latitude longitude and grid reference numbers.
Why does my answer not match the book's?
The usual causes are swapped weight order, a basis matrix built from rows instead of columns, or a sign slip in solving. Recombine your answer with the basis and compare against the original vector; the component that mismatches points at the line where the error entered.
Photo credits: Monstera Production (Pexels); Bernice Chan (Pexels). Photos are served from this site; follow the credit to the library page for the original.