Testing Conjectures — I can use counterexamples and boundary cases to test conjectures and refine claims. - Grade 11 Math

SECTION 1

Historical Context & Motivation

Mathematics has always been a discipline built on claims and proofs. For thousands of years, mathematicians have proposed bold statements—called conjectures—and then worked to determine whether those statements are always true or sometimes false. The practice of testing conjectures by searching for counterexamples and boundary cases is one of the oldest and most powerful tools in mathematical reasoning. When even a single counterexample is found, the conjecture is disproved—no matter how many cases supported it before.

~300 BCE • Euclid's Elements

Euclid established the standard of deductive proof in geometry. His work showed that claims must be proven, not merely observed from examples—a principle that naturally raises the question of how to disprove false claims.

1637 • Fermat's Last Theorem Conjectured

Pierre de Fermat wrote in a margin that no three positive integers satisfy aⁿ + bⁿ = cⁿ for n > 2. For centuries, mathematicians searched for a counterexample without finding one.

1769 • Euler's Conjecture Disproved (Eventually)

Leonhard Euler generalized Fermat's conjecture, claiming similar rules for higher powers. In 1966, a counterexample was finally found using computers: 27⁵ + 84⁵ + 110⁵ + 133⁵ = 144⁵. One counterexample destroyed a 200-year-old conjecture.

1848 • Polya's Conjecture & Boundary Thinking

George Pólya later formalized heuristic strategies for mathematical discovery, emphasizing the importance of testing claims at extreme and special values—what we now call boundary cases.

Modern Era • Computational Conjecture Testing

Today, computers test billions of cases per second, but the logical framework remains the same: a single counterexample disproves a universal claim, while no amount of confirming examples can fully prove one.

The core question this lesson addresses is straightforward but deeply important: How do we systematically determine whether a mathematical claim is true, false, or in need of revision? Rather than relying on intuition alone, you will learn to deploy counterexamples and boundary cases as precision tools for evaluating conjectures.

SECTION 2

Core Principles & Definitions

Before diving into techniques, you need to understand the key vocabulary and ideas that underpin conjecture testing. A conjecture is a mathematical statement believed to be true based on observation or pattern recognition, but not yet proven. A counterexample is a specific case that satisfies the conditions of the conjecture but violates the conclusion, thereby disproving it. A boundary case (also called an edge case) is an input at the extreme end of the conjecture's domain—such as zero, one, negative numbers, or very large values—where claims are most likely to fail.

Conjecture

A statement proposed as true based on patterns or examples. It remains a conjecture until it is either proven or disproven.

Counterexample

A single specific example that meets the conjecture's conditions but shows the conclusion is false. One counterexample is enough to disprove any universal claim.

Boundary Case

An extreme or special input value (like 0, 1, −1, or very large numbers) where a conjecture is most vulnerable to failure. Always test boundaries first.

Refining a Claim

When a counterexample exposes a flaw, you can often salvage the conjecture by adding conditions or narrowing its scope. This process turns a broken conjecture into a stronger, more precise claim.

Universal vs. Existential Claims

Universal claims ("for all x...") need just one counterexample to disprove. Existential claims ("there exists an x...") need just one confirming example to prove. Knowing the type of claim tells you what kind of evidence matters.

✦

KEY TAKEAWAY

Think of a conjecture like a "No Leaks" guarantee on a water balloon. You can fill it successfully a hundred times, but if it bursts just once, the guarantee is broken. A counterexample is that one burst. Boundary cases are like squeezing the balloon at its thinnest spot—that's where leaks are most likely to appear.

SECTION 3

Visual Explanation — The Conjecture Testing Process

The diagram below illustrates the logical flowchart you should follow whenever you encounter a conjecture. Notice that the process is cyclical: when a counterexample is found, you don't simply throw away the conjecture—you refine it and loop back through the testing process with a stronger version of the claim.

The flowchart shows the five-stage process: state the conjecture, identify boundary cases, test cases systematically, determine whether a counterexample exists, and either confirm the conjecture survives or refine the claim and loop back.

The diamond-shaped decision point in the middle is the critical juncture. If you answer "Yes"—a counterexample was found—the conjecture is immediately disproved. However, this is not the end of the story. The dashed arrow shows that you can refine the original claim by adding restrictions or conditions that exclude the problematic case, then restart the testing cycle. If you answer "No" after exhaustive testing, the conjecture survives—but remember, surviving is not the same as being proven. A formal proof would still be needed for certainty.

SECTION 4

Mathematical Framework — Logic of Counterexamples

Understanding the logical structure behind conjectures helps you know exactly what kind of evidence is needed to test them. Most conjectures take the form of universal statements: "For all x in some set, property P(x) holds." The negation of this statement is an existential statement: "There exists at least one x such that P(x) does not hold." Finding that one x is what we call a counterexample.

UNIVERSAL CONJECTURE

∀x ∈ S : P(x)

"For all x in set S, property P(x) is true." This claims that every single element in the domain satisfies the property.

NEGATION (COUNTEREXAMPLE)

∃x ∈ S : ¬P(x)

"There exists at least one x in set S such that P(x) is not true." Finding this x disproves the conjecture completely.

BOUNDARY CASE STRATEGY

Test x ∈ {0, 1, −1, n_max, edge values}

When testing a conjecture, always prioritize values at the extremes of the domain. These include zero, one, negative one, the smallest and largest allowed values, and any values where formulas might produce division by zero or sign changes.

Here is an important asymmetry to internalize: a thousand confirming examples do not prove a universal conjecture, but a single counterexample disproves it instantly. This is why mathematicians often spend more effort looking for counterexamples than collecting supporting evidence. If the conjecture survives every test—including all boundary cases—you gain confidence in it, but only a formal proof provides certainty.

⚠

Why Boundary Cases Matter

Many false conjectures work perfectly for "typical" values but break at the edges. For instance, the claim "n² > n for all integers" works for n = 2, 3, 100, but fails at the boundary values n = 0 (where 0² = 0, not > 0) and n = 1 (where 1² = 1, not > 1). Always test 0, 1, −1, and the edges of the domain first.

SECTION 5

Detailed Breakdown — Testing Strategies & Common Traps

Now that you understand the logical foundation, let's explore the specific strategies you should use when testing a conjecture. The diagram below categorizes the types of test values you should always consider, organized from the most common sources of counterexamples to the least common.

The Boundary Case Testing Map shows three categories of values you should always check—zero and identity values, sign changes, and extreme values—along with a worked example demonstrating how boundary testing reveals counterexamples and leads to a refined conjecture.

As the example in the diagram shows, the conjecture "n² > 2n for all integers n" seems plausible if you only test values like 5 or 10. But when you test the boundary cases of n = 0 and n = 1, the conjecture falls apart. The refined version—"n² > 2n for all integers n > 2"—is a stronger, more accurate claim. This cycle of conjecture, counterexample, and refinement is the engine of mathematical discovery.

✓ Strategy 1: Test zero first. Zero is the most common source of counterexamples because multiplication by zero, division by zero, and zero exponents all produce surprising results.
✓ Strategy 2: Try negative values. If the conjecture doesn't specify positive numbers, negative inputs can reverse inequalities and change the sign of products.
✓ Strategy 3: Test fractions between 0 and 1. Squaring a fraction between 0 and 1 makes it smaller, not larger—the opposite of what happens with integers greater than 1.
✓ Strategy 4: Check the domain boundaries. If the conjecture says "for all real numbers," make sure you test negative reals, irrational numbers, and values approaching infinity.

SECTION 6

Worked Example — Testing and Refining a Conjecture

Let's work through a complete example from start to finish. Consider the conjecture: "For all real numbers x, the expression x² + 1 is greater than 2x." We will test this conjecture, identify whether it holds, and refine it if necessary.

Testing the Conjecture: x² + 1 > 2x for all real numbers x

1
Step 1 — Rewrite the Conjecture Algebraically

Start by rearranging the inequality x² + 1 > 2x into a form that's easier to analyze. Subtract 2x from both sides to get x² − 2x + 1 > 0. Notice that the left side factors as (x − 1)² > 0.

Equivalent form: (x − 1)² > 0
2
Step 2 — Test Typical Values

Try x = 3: (3 − 1)² = 4 > 0. ✓ True. Try x = −2: (−2 − 1)² = 9 > 0. ✓ True. Try x = 0: (0 − 1)² = 1 > 0. ✓ True. So far the conjecture holds for every test.

All typical values pass the test.
3
Step 3 — Test the Key Boundary Case

Now test the boundary case x = 1, which is the value that makes the squared expression equal to zero. We get (1 − 1)² = 0² = 0. The conjecture claims this must be greater than zero, but 0 is not greater than 0. We have found a counterexample!

Counterexample found: x = 1 gives (1 − 1)² = 0, which is NOT > 0.
4
Step 4 — Analyze the Counterexample

The expression (x − 1)² is a perfect square, so it is always ≥ 0 for all real numbers. It equals zero only when x = 1. This means the original conjecture is almost true—it fails at exactly one point.

(x − 1)² ≥ 0 always, but (x − 1)² = 0 when x = 1.
5
Step 5 — Refine the Conjecture

We can fix the conjecture in two ways. Option A: change the strict inequality to a non-strict one: x² + 1 ≥ 2x for all real x (this is true because (x − 1)² ≥ 0). Option B: add a restriction to the domain: x² + 1 > 2x for all real x ≠ 1. Both versions are now correct, robust claims.

Refined conjecture: x² + 1 ≥ 2x for all real numbers x. ✓ TRUE

💡

Notice the Pattern

This example demonstrates a crucial lesson: the counterexample at x = 1 didn't destroy the conjecture—it improved it. Changing > to ≥ was a small but precise fix that made the claim universally true. Counterexamples are tools for refinement, not just rejection.

SECTION 7

Strengths & Limitations of Conjecture Testing

Like any method, conjecture testing with counterexamples and boundary cases has clear strengths and important limitations. Understanding both will help you know when this approach is the right tool and when you need more.

Aspect	Strengths	Limitations
Power of Disproof	A single counterexample conclusively disproves a universal claim—no ambiguity.	Cannot prove a conjecture is true; absence of counterexamples ≠ proof.
Efficiency	Boundary cases narrow the search dramatically—you don't need to test millions of values.	Some counterexamples exist only at very large values or obscure numbers, making them hard to find.
Refinement	Failed conjectures can be salvaged by adding conditions, leading to stronger claims.	Over-restricting conditions can make a conjecture trivially true and uninteresting.
Accessibility	Requires no advanced proof techniques—just careful substitution and logic.	Does not develop the formal reasoning skills needed for deductive proof.
Scope	Works across all branches of math: algebra, geometry, number theory, and beyond.	Some conjectures involve continuous domains (all real numbers), making exhaustive testing impossible.

Comparison of strengths and limitations of counterexample-based testing

✦

KEY TAKEAWAY

Think of counterexample testing like a quality control inspector sampling products from a factory. If the inspector finds one defective item, the batch fails—that's the power of a counterexample. But if the inspector checks 100 items and finds no defects, it doesn't guarantee the next 1,000 are perfect—that's why we still need formal proof for absolute certainty.

SECTION 8

Connection to Formal Proof & Advanced Reasoning

Counterexample testing is a powerful first step, but mathematics ultimately demands proof for claims that survive testing. Understanding how testing relates to proof gives you a complete picture of mathematical reasoning. The table below compares the testing approach you've learned in this lesson with the formal proof techniques you'll encounter in more advanced courses.

Feature	Counterexample Testing	Formal Proof
Purpose	Disprove false claims; build confidence in true ones	Establish absolute certainty that a claim is true
Approach	Test specific inputs, especially boundary cases	Use deductive logic from axioms and definitions
Outcome	Conjecture disproved or survives (not proven)	Theorem established for all cases
Difficulty	Accessible—requires substitution and logic	Advanced—requires techniques like induction, contradiction
Relationship	Often the first step before attempting a proof	Completes the process that testing begins

Counterexample testing vs. formal proof

In later courses, you'll learn proof techniques such as mathematical induction (proving a claim for all natural numbers by showing it works for n = 1 and then showing that if it works for n = k, it also works for n = k + 1), proof by contradiction (assuming the claim is false and deriving an impossible result), and direct proof (building a chain of logical steps from assumptions to conclusion). The skill of testing conjectures you're building now directly prepares you for these methods—before you attempt to prove something, you should always test it first to make sure it's worth proving.

SECTION 9

Practice Problems

Test your understanding with the following five problems, which progress from conceptual reasoning to critical thinking. For each problem, consider what types of test cases would be most effective, and remember to check boundary values before drawing conclusions.

Question 1: CONCEPTUAL

A student claims: "All prime numbers are odd." Is this conjecture true or false? If false, provide a counterexample and explain why it disproves the claim.

Show answer

The conjecture is false. The number 2 is a counterexample: it is a prime number (its only factors are 1 and itself), but it is even, not odd. This single counterexample disproves the universal claim because the conjecture says all primes are odd, and 2 is a prime that is not odd. A refined version would be: "All prime numbers greater than 2 are odd."

Question 2: BASIC CALCULATION

Test the conjecture: "For all positive integers n, the expression n² − n + 41 always produces a prime number." Test n = 1, 2, 3, and then find a value of n for which the conjecture fails.

Show answer

For n = 1: 1 − 1 + 41 = 41 (prime ✓). For n = 2: 4 − 2 + 41 = 43 (prime ✓). For n = 3: 9 − 3 + 41 = 47 (prime ✓). The formula produces primes for many small values, which is why this conjecture is famous. However, it fails at n = 41: 41² − 41 + 41 = 41² = 1681, which equals 41 × 41 and is clearly not prime. This boundary case works because plugging in n = 41 makes the expression factor as 41(41 − 1 + 1) = 41 × 41. This is a classic example of how a pattern can hold for dozens of cases but still be false.

Question 3: INTERMEDIATE

Consider the conjecture: "If a quadrilateral has two pairs of equal sides, then it is a parallelogram." Determine whether this is true or false. If false, describe or sketch a counterexample and then refine the conjecture to make it true.

Show answer

The conjecture is false. A kite is a counterexample: it has two pairs of consecutive equal sides (for example, sides of lengths 3, 3, 5, 5 arranged so equal sides are adjacent), but it is not a parallelogram because its opposite sides are not parallel. The conjecture confuses "two pairs of equal sides" with "two pairs of opposite equal sides." A refined version is: "If a quadrilateral has two pairs of equal and parallel opposite sides, then it is a parallelogram." This correction adds the necessary condition about which sides are equal.

Question 4: APPLIED

A tutor tells a student: "Whenever you take the square root of a sum, you can split it up: √(a + b) = √a + √b." The student suspects this is wrong. Identify at least two specific counterexamples using boundary cases, explain why the conjecture fails, and state a correct relationship involving √(a + b).

Show answer

Counterexample 1: Let a = 9 and b = 16. Then √(9 + 16) = √25 = 5, but √9 + √16 = 3 + 4 = 7. Since 5 ≠ 7, the conjecture fails. Counterexample 2 (boundary): Let a = 0 and b = 4. Then √(0 + 4) = √4 = 2, and √0 + √4 = 0 + 2 = 2. This boundary case actually works! So the conjecture isn't always false—it holds when one of the values is zero. Counterexample 3: Let a = 1 and b = 1. Then √2 ≈ 1.414, but √1 + √1 = 2. The conjecture fails because the square root function is concave, meaning √(a + b) ≤ √a + √b always holds (by the triangle inequality), but equality requires at least one of a or b to be zero. The correct relationship is the inequality √(a + b) ≤ √a + √b for non-negative a and b.

Question 5: CRITICAL THINKING

Consider the conjecture: "For all real numbers x, if x³ > x², then x > 1." Test this conjecture systematically using boundary cases and values from different parts of the number line. Is the conjecture true? If not, disprove it, explain why it fails, and provide the most precise refined version you can.

Show answer

Let's test systematically. For x = 2: 8 > 4 and 2 > 1—consistent. For x = 0.5: 0.125 > 0.25 is false, so the hypothesis (x³ > x²) is not met—this case doesn't apply. For x = −2: (−2)³ = −8 and (−2)² = 4, so −8 > 4 is false—hypothesis not met. Now the key boundary case: try x = −0.5. We get (−0.5)³ = −0.125 and (−0.5)² = 0.25, so −0.125 > 0.25 is false—still doesn't apply. However, consider a very negative number: x = −3. Then x³ = −27 and x² = 9, so −27 > 9 is false. What about negative values between −1 and 0? None of those work either because x³ is negative and x² is positive. Actually, this conjecture IS true. If x³ > x², we can write x³ − x² > 0, which factors as x²(x − 1) > 0. Since x² ≥ 0 for all real x, and x² = 0 only when x = 0 (but then x²(x − 1) = 0, not > 0), we need x² > 0 AND x − 1 > 0, meaning x ≠ 0 and x > 1. Therefore x > 1, and the conjecture is proven true. This problem illustrates that sometimes exhaustive boundary testing builds the confidence needed to attempt—and succeed at—a proof.

SUMMARY

Lesson Summary

Summary

A conjecture is a mathematical claim proposed as true based on patterns, but not yet proven. To test a conjecture, you search for counterexamples—specific cases that satisfy the conditions but violate the conclusion. A single counterexample is sufficient to disprove any universal claim, no matter how many confirming examples exist. The most efficient strategy is to test boundary cases first: values like 0, 1, −1, fractions between 0 and 1, negative numbers, and extreme values at the edges of the domain.

When a counterexample is found, the conjecture is disproved—but not discarded. Instead, you refine the claim by adding conditions, narrowing the domain, or adjusting the inequality. This process of conjecture → test → refine is central to mathematical reasoning and prepares you for formal proof techniques like induction and contradiction. Remember: no amount of confirming examples proves a conjecture, but one counterexample disproves it. Always test boundary cases before drawing conclusions.