Did you know? A straight line is the shortest route between two points, but that simple statement has an algebraic twin. The triangle inequality tells us that one combined movement cannot be longer than the two movements taken separately. Cauchy–Schwarz expresses another kind of limit: how large can the interaction between two lists of numbers be?
These two ideas turn unfamiliar expressions into statements you can prove. For a JC2 H3 Mathematics student in Bukit Timah, the goal is not to memorise two names. It is to see hidden sums of squares, choose the correct inequality, and explain when equality happens.
Quick answer: For real a,b,c,d, the Cauchy–Schwarz inequality is (ac+bd)² ≤ (a²+b²)(c²+d²). For real x,y, the triangle inequality is |x+y| ≤ |x|+|y|. Both generalise to vectors. Each bound has an equality condition that must be checked if you are solving a minimum or maximum problem.
Jump to a clear answer
- Cauchy–Schwarz in one line
- Why the square proof works
- A worked minimum with a fixed sum
- The useful fraction form
- Triangle and reverse triangle inequalities
- A complete distance problem
- Nine short learning clinics
- Mistakes and equality cases
- JC2 revision plan
- FAQs and official sources
Cauchy–Schwarz: a ceiling for an inner product
Take two ordered pairs of real numbers, (a,b) and (c,d). Their dot product is ac+bd, while their squared lengths are a²+b² and c²+d². Cauchy–Schwarz states that the square of their dot product cannot exceed the product of their squared lengths.
In familiar form, (ac+bd)² ≤ (a²+b²)(c²+d²). For real lists a₁,…,aₙ and b₁,…,bₙ, the extension is (Σaᵢbᵢ)² ≤ (Σaᵢ²)(Σbᵢ²). The point of that notation is to recognise two sums of squares and a matching sum of products.
Equality holds when the two nonzero vectors are scalar multiples of one another (and also when one vector is zero). In a numerical optimisation problem, say which vectors are being compared, then solve their proportionality condition.
For example, (1,2) and (2,4) are proportional. Their dot product is 10; the squared dot product is 100. Their squared lengths are 5 and 20, giving 100. The bound is attained exactly.
The 2026 and 2027 H3 Mathematics 9820 content outlines explicitly include Cauchy–Schwarz under additional inequalities. The algebra may look compact, but the writing challenge is recognising the right quantities and proving the relevant equality case.
Where does the formula come from?
Expand the difference between the two sides: (a²+b²)(c²+d²)−(ac+bd)². Multiplying out and cancelling gives a²d²−2abcd+b²c²=(ad−bc)².
Every real square is nonnegative. Therefore (a²+b²)(c²+d²)≥(ac+bd)². Equality occurs when ad−bc=0. This elementary proof is useful because it exposes the inequality as a guaranteed nonnegative square, not a mysterious rule.
For more than two coordinates, the vector form extends the same principle. A JC2 learner should practise moving between the symbols, the geometry and a concrete numerical example.
Worked question: least sum of squares with fixed total
Question: Real numbers x and y satisfy x+y=10. What is the smallest possible value of x²+y²?
Apply Cauchy–Schwarz to vectors (x,y) and (1,1). We get (x+y)²≤(x²+y²)(1²+1²)=2(x²+y²). Hence 100≤2(x²+y²), so x²+y²≥50.
Equality requires (x,y) proportional to (1,1), so x=y. The condition x+y=10 then gives x=y=5. Thus the minimum value is 50, attained at (5,5).
The important move was recognising the sum x+y as a dot product with (1,1). In three variables a+b+c=12, the same strategy gives (a+b+c)²≤3(a²+b²+c²), hence a²+b²+c²≥48, with equality at a=b=c=4.
The fraction form many students overlook
For positive a,b and real x,y, another form is x²/a + y²/b ≥ (x+y)²/(a+b). It is not a separate magic rule: apply Cauchy–Schwarz to (x/√a, y/√b) and (√a,√b).
Here the dot product is x+y, while the squared lengths are x²/a+y²/b and a+b. Rearrange the resulting inequality. Equality requires x/a=y/b.
Take the expression x²/2+(4−x)²/2 for real x. The numerators sum to x+(4−x)=4, and the denominators sum to 4. So the value is at least 4²/4=4.
Equality requires x/2=(4−x)/2, giving x=2. The minimum is 4. A student who simply writes the inequality without solving x=2 has not shown that the bound is attainable.
This form is helpful when several squared expressions are divided by positive weights. Always check that every denominator used in this transformation is positive.
Triangle inequality: why detours cannot shorten a journey
For real x and y, |x+y| ≤ |x|+|y|. Imagine moving x units and then y units on a number line. The net displacement cannot exceed the total distance travelled. If the movements have opposite signs, some of the journey cancels.
For vectors u and v, the corresponding statement is ||u+v||≤||u||+||v||, where || || denotes Euclidean length. Cauchy–Schwarz gives u·v≤||u||||v||. Squaring the norm of u+v then proves the triangle bound.
Complex numbers obey the analogous modulus inequality |z+w|≤|z|+|w|. For instance, z=3+4i and w=6+8i point in the same direction: |z+w|=15 and |z|+|w|=5+10=15.
When can equality occur for real numbers? When x and y have the same sign or one is zero. For nonzero complex numbers or Euclidean vectors, equality means they point in the same direction, one a positive scalar multiple of the other.
An important companion is the reverse triangle inequality: ||x|−|y||≤|x−y| for real x,y. It follows by applying the triangle inequality to x=(x−y)+y and reversing the roles. It helps bound differences in lengths when direct expressions are awkward.
Worked question: find the minimum distance sum
Question: For real x, find the minimum of |x−3|+|x+2| and all x where it is attained.
Interpret the terms as distances from x to 3 and from x to −2 on the number line. By the triangle inequality, |x−3|+|x+2|≥|(x−3)−(x+2)|=5.
To test equality, place x anywhere between the two fixed points: −2≤x≤3. Then |x−3|=3−x and |x+2|=x+2, so their sum is exactly five.
The global minimum is therefore 5 for every x in [−2,3]. This teaches a valuable lesson: an equality condition can describe an entire interval, not just one ‘optimal point’.
Nine original practice clinics
1. Easy dot product
For (1,2) and (3,4), (ac+bd)²=11²=121, while (a²+b²)(c²+d²)=5×25=125. The difference is (1×4−2×3)²=4.
2. Equality test
Vectors (2,−4) and (−1,2) are proportional, so the Cauchy–Schwarz bound is attained. The constant of proportionality may be negative, unlike the triangle-equality direction.
3. Sum of three squares
If a+b+c=6 for real a,b,c, then a²+b²+c²≥6²/3=12. Equality at a=b=c=2.
4. Weighted fractions
For a,b>0, 1/a+1/b≥4/(a+b). Set x=y=1 in the fraction form. Equality requires a=b.
5. A two-term distance
For real x, |x|+|x−4|≥4. Equality holds exactly for 0≤x≤4.
6. Reverse triangle
For real x,y, ||x|−|y||≤|x−y|. Set x=7, y=−2: the lower bound is 5, and the actual difference is 9.
7. Strict triangle case
If z=3+4i and w=−3−4i, then |z+w|=0, while |z|+|w|=10. The inequality is strict because the vectors oppose.
8. Equality is not automatic
The bound x²+y²≥50 when x+y=10 becomes equality only when x=y=5. Reporting 50 without finding a permitted equality case is incomplete.
9. Watch the denominator
A fraction-form proof using 1/a is invalid for a=0 and requires additional caution for negative a. Confirm the domain before invoking it.
Common errors and an easy self-check
One mistake is taking a positive square root of an inequality without checking that both sides are nonnegative. Another is confusing dot-product equality with triangle equality. Cauchy–Schwarz permits vectors proportional with a negative factor; equality in the triangle bound for nonzero vectors requires a nonnegative one.
A third mistake is applying the fraction form with a denominator that might be zero or negative. A fourth is writing a final number without demonstrating that some permissible values achieve it. A fifth is forgetting that an absolute-value minimum might hold over an interval.
Before submitting a solution, write three notes: the vectors or terms chosen, the inequality being used, and the equality condition. These three parts make the underlying reasoning visible to an examiner or tutor.
Finally, avoid forcing every problem through the same formula. A square-completion proof, AM–GM or direct distance argument may be shorter. Good H3 problem solving is choosing a justified route, not showcasing the most impressive-looking name.
A practical JC2 revision routine
Start with three dot products and identify whether Cauchy–Schwarz equality holds. Then solve two fixed-sum minimum problems in two variables and one in three variables. Move on to two weighted fraction problems, making the denominator restrictions explicit.
Practise triangle inequalities on the real line before tackling vectors or complex numbers. Draw two points and a variable point, then turn the picture into an absolute-value inequality and find its equality set.
Ask a tutor to challenge one step: ‘Why is that inequality applicable here?’ and one conclusion: ‘Have you proved the bound can be reached?’ The purpose of individual feedback is to expose the exact missing explanation, not to increase the number of drills endlessly.
The eduKateSG Mathematics reference near Sixth Avenue MRT documents a three-pupil Secondary 1 small-group model. It does not establish a JC2 H3 timetable or availability. Families can ask about current JC2 H3 support before making arrangements.
Frequently asked questions
Are Cauchy–Schwarz and triangle inequality covered in H3 Mathematics 9820?
Yes. Both are listed under additional inequalities in the official 2026 and 2027 H3 Mathematics syllabus outlines.
Is Cauchy–Schwarz only for vectors?
No. It works for finite lists of real numbers. Writing a sum as a dot product is often the simplest way to recognise it.
What is the equality condition for Cauchy–Schwarz?
The two vectors are linearly dependent, including the case that one is the zero vector.
Can triangle inequality produce many minimisers?
Yes. On a real line, the sum of distances to two fixed points is minimal at every point between them.
How are these inequalities different from AM–GM?
AM–GM compares arithmetic and geometric means of nonnegative terms. Cauchy–Schwarz compares dot products and lengths, while triangle inequality compares the size of a sum with individual sizes.
Official syllabus and connected H3 reading
The 2027 Singapore–Cambridge H3 Mathematics (9820) syllabus lists the inequality topics and clarifies that H2 Mathematics knowledge is assumed. Current JC2 candidates can consult the 2026 H3 syllabus for their examination year.
Review the AM–GM inequality and equality conditions for a complementary optimisation technique. For writing mathematically rigorous explanations, see precise H3 statements and proof language.