Vector Products and Geometry

01 / UNDERSTAND & APPLY

1. A vector records a displacement, not a location

A point tells you where something is. A vector tells you a displacement with magnitude and direction. After choosing an origin O, the position vector of a point A is a = OA. The displacement from A to B is AB = b − a. Shifting the origin changes both position vectors but leaves their difference unchanged, so distances and shapes do not depend on that arbitrary choice.

In a right-handed orthonormal coordinate system, write a vector u = (u₁,u₂,u₃). Its length is |u| = √(u₁² + u₂² + u₃²). Addition is the head-to-tail combination of displacements, and multiplication by a real scalar changes length and possibly reverses direction. The zero vector has length zero and no defined direction.

A unit vector along a nonzero u is u/|u|. The opposite unit vector is −u/|u|. If the question asks for a unit vector parallel to a line without specifying orientation, both signs may be valid. Dividing the zero vector by its length is undefined, so always check nonzero conditions before normalizing.

Products of vectors serve different purposes. The dot product is a scalar describing alignment. The cross product in three-dimensional space is a vector describing oriented area and a perpendicular direction. The scalar triple product is a scalar describing oriented volume. Keeping the output type in mind prevents impossible expressions and mistaken substitutions.

02 / UNDERSTAND & APPLY

2. Internal and external division derived from one parameter

Every point P on the line through distinct A and B can be written p = a + t(b − a) = (1 − t)a + tb. The parameter t measures directed progress from A toward B: t = 0 gives A, t = 1 gives B, and 0 < t < 1 gives the interior of the segment. Values outside [0,1] describe external points on the same line.

If AP:PB = m:n internally with m,n > 0, then t/(1 − t) = m/n and t = m/(m + n). Therefore p = (na + mb)/(m + n). The coefficient of b is paired with the distance from A, not with the distance from B. Deriving it from t is safer than memorizing an easily reversed formula.

For external division with positive length ratio AP:PB = m:n and m ≠ n, one obtains p = (mb − na)/(m − n). This point lies outside the segment. The denominator is not m + n, and m = n has no finite external division point for distinct A,B. Directed-ratio conventions vary, so defining the point with a line parameter removes ambiguity.

Weighted averages with positive weights lie inside the appropriate segment or convex region. Weights summing to one preserve affine location under a change of origin. In contrast, a raw sum a + b usually does not represent the midpoint; it must be divided by two. These are statements about locations, while b − a is a displacement.

03 / UNDERSTAND & APPLY

3. Centroids, medians and a useful geometric proof

The centroid of a triangle with vertex vectors a,b,c is g = (a + b + c)/3. To derive it, let M be the midpoint of BC, so m = (b + c)/2. The point two-thirds of the way from A to M is a + (2/3)(m − a) = (a + b + c)/3. The same symmetric expression arises from either other median, proving that all three medians meet there.

The centroid divides each median in the ratio AG:GM = 2:1, with the longer part adjacent to the vertex. This is not a statement that it divides every segment inside the triangle in that ratio. A tetrahedron’s vertex centroid is (a + b + c + d)/4 and lies three-quarters of the way from a vertex to the centroid of the opposite face.

Collinearity is expressed by b − a = λ(c − a) for some real scalar λ, provided the relevant points are distinct when a direction is needed. Coplanarity of four points will later use the scalar triple product of three displacements from one common vertex. Using position vectors themselves can accidentally test a plane through the origin rather than the plane containing the points.

For a diagram, draw one median and label the midpoint of its opposite side. Add the two equal displacement arrows that define the midpoint, then mark the centroid two-thirds from the vertex. The algebraic average and the geometric ratio become the same construction rather than unrelated formulas.

04 / UNDERSTAND & APPLY

4. The dot product measures alignment

For u,v in Euclidean three-space, u·v = u₁v₁ + u₂v₂ + u₃v₃. For nonzero vectors it also equals |u||v|cos θ, where θ is the angle between their chosen directions in [0,π]. This identity follows by comparing the cosine rule for |u − v|² with its coordinate expansion.

A positive dot product corresponds to an acute angle, zero to a right angle, and negative to an obtuse angle. The zero-vector exception matters: u·0 = 0 algebraically, but there is no defined angle with the zero vector. When claiming perpendicular directions, require both vectors to be nonzero.

The dot product is commutative and distributive, and u·u = |u|². Expanding |u + v|² gives |u|² + 2u·v + |v|². This is often the fastest way to infer an angle or compare diagonal lengths. It is not true that |u + v| = |u| + |v| in general; that equality needs compatible directions.

Cauchy–Schwarz gives |u·v| ≤ |u||v|, so a computed cosine must lie between −1 and 1. Equality holds for linearly dependent vectors. If an angle calculation produces a value outside that range, recheck the dot product, lengths and whether the vectors represent the intended directions.

05 / UNDERSTAND & APPLY

5. A projection splits a vector into parallel and perpendicular parts

To project u onto a nonzero v, seek a vector λv such that u − λv is perpendicular to v. The dot-product condition (u − λv)·v = 0 gives λ = (u·v)/(v·v). Therefore projᵥu = [(u·v)/|v|²]v. The remaining component u − projᵥu is perpendicular to v by construction.

The scalar component along the oriented unit direction v/|v| is (u·v)/|v|. It may be negative. The length of the projected vector is its absolute value. Distinguish scalar component, projection vector and projection length, because all three can appear in similar wording.

Pythagoras applies to this perpendicular decomposition: |u|² = |projᵥu|² + |u − projᵥu|². This gives a convenient numerical check. Replacing v by a nonzero scalar multiple changes neither the projection vector nor its length, though a negative multiplier reverses the chosen oriented direction and changes the sign of the scalar component.

06 / UNDERSTAND & APPLY

6. The cross product is an oriented area vector

For u = (u₁,u₂,u₃) and v = (v₁,v₂,v₃), define u×v = (u₂v₃ − u₃v₂, u₃v₁ − u₁v₃, u₁v₂ − u₂v₁). It is perpendicular to both vectors. In a right-handed coordinate system, its direction follows the right-hand rule as fingers turn from u toward v through the smaller angle.

Its magnitude is |u×v| = |u||v|sin θ, the area of the parallelogram with adjacent sides u and v. Reversing the order reverses the vector but preserves the area: v×u = −u×v. A zero cross product means the two vectors are linearly dependent, including the possibility that one is zero.

The determinant mnemonic has a minus sign hidden in the middle component. Expanding directly as the coordinate formula can prevent a sign slip. A strong check is to dot the computed cross product with each original vector; both results should be zero. This catches many arithmetic mistakes before an area or volume is calculated.

The identity |u×v|² = |u|²|v|² − (u·v)² connects dot and cross products. It follows from sin²θ + cos²θ = 1 and also holds in degenerate cases. It can compute an area without finding the full cross-product vector when only lengths and a dot product are known.

07 / UNDERSTAND & APPLY

7. Areas use adjacent edges, and products do not associate

The triangle with vertices A,B,C has area (1/2)|(b − a)×(c − a)|. Both edge vectors must start from the same vertex. Using a×b without translating to a common vertex generally measures a triangle involving the origin instead of the requested triangle.

The vector triple-product identity is u×(v×w) = v(u·w) − w(u·v). The result lies in the plane spanned by v and w, which fits the fact that it is perpendicular to v×w. The order matters: (u×v)×w = v(u·w) − u(v·w), a different expression.

Cross products are not associative. With basis vectors i,j,k, i×(i×j) = i×k = −j, while (i×i)×j = 0. Never omit parentheses in a triple cross product. Also do not confuse a vector triple product with the scalar u·(v×w), which measures oriented volume and is treated next.

08 / UNDERSTAND & APPLY

8. A scalar triple product is base area times signed height

For three vectors u,v,w, the scalar triple product [u,v,w] = u·(v×w) is the determinant of their coordinate rows. The magnitude |v×w| is the base parallelogram area, and the component of u along its unit normal is the signed height. Their product is therefore an oriented parallelepiped volume.

The geometric volume is |u·(v×w)|. Cyclic permutations preserve the scalar triple product, while swapping two vectors changes its sign. If the product is zero, the three vectors are linearly dependent and lie in some common plane through the origin. A zero product can also arise from a zero vector, which is a degenerate case rather than a three-dimensional solid.

For a tetrahedron with adjacent edges u,v,w from one vertex, the volume is |u·(v×w)|/6. One way to see the factor is to take a triangular base, whose area is half the parallelogram base, and use the pyramid formula one-third times base area times height. The factors 1/2 and 1/3 combine to 1/6.

To test whether four points A,B,C,D are coplanar, form AB,AC,AD and set their scalar triple product to zero. The common starting point is essential. The test includes collinear and repeated-point degeneracies, so if the problem asks for a unique plane, also verify that at least three of the points are noncollinear.

09 / UNDERSTAND & APPLY

9. Surface area and volume are different combinations of edges

A parallelepiped with adjacent edges u,v,w has three pairs of congruent opposite faces. Its surface area is 2(|u×v| + |v×w| + |w×u|), while its volume is |u·(v×w)|. Multiplying the three edge lengths gives the volume only when the edges are mutually perpendicular. In a skew solid, some of that length lies sideways rather than contributing to height.

A prism has volume base area times perpendicular height, and a pyramid or cone has one-third of that product. For a right circular cylinder of radius r and height h, volume is πr²h; a right circular cone has πr²h/3. The slant height belongs to lateral surface area, not to the perpendicular-height volume formula. For a sphere, surface area is 4πr² and volume is 4πr³/3.

Uniformly scaling every length by a positive factor λ multiplies areas by λ² and volumes by λ³. This gives a quick check on any coordinate formula: a cross product of two edge vectors scales quadratically, and a scalar triple product scales cubically. If only one edge changes or an angle changes, uniform-scaling rules no longer apply without further analysis.

When the task gives several faces, draw the solid and label adjacent edges from one corner. The same vertex may have three edge vectors, three face-area vectors and one triple-product volume. This organizes the computation and prevents substituting a face diagonal where the formula requires an edge.

10 / UNDERSTAND & APPLY

10. Original practice: positions, angles and areas

Use coordinates only after identifying the correct geometric vectors. Check every normalized or projected direction is nonzero.

11 / UNDERSTAND & APPLY

11. Original practice: volume, degeneracy and scaling

Keep signed triple products until the orientation question is settled, then take absolute values for physical volume. All problems here are original practice.

12 / UNDERSTAND & APPLY

12. A geometric audit of vector calculations

Before computing, label the geometric meaning and output type of each expression. A position vector locates a point; a difference gives an edge; a dot product is a scalar; a cross product is a perpendicular vector; its magnitude is an area; a scalar triple product is a signed volume. This type-check is a powerful safeguard against plausible but meaningless algebra.

For a section point, test whether it lies inside or outside the segment as requested and verify the ratio using displacements. For a dot product, check the resulting cosine range. For a cross product, verify perpendicularity to both inputs. For a volume, verify common-vertex edges and compare with base area times height when the geometry allows it.

The common errors are reversing section weights, treating an external ratio as an internal one, forgetting a nonzero condition before normalization, losing the middle cross-product sign, omitting the triangle factor 1/2 or tetrahedron factor 1/6, and multiplying skew edge lengths as though they were perpendicular.

Finally, a zero area or volume signals degeneracy rather than a mysterious negative-dimensional object. Collinear triangle vertices give zero cross-product area; coplanar tetrahedron vertices give zero triple-product volume. If the question assumes a genuine triangle, plane or solid, check the nondegeneracy conditions explicitly.