Model Domain Limitations — I can explain why a model may be unrealistic outside a certain domain. - Grade 9 Math

SECTION 1

Historical Context & Motivation

Throughout history, scientists and mathematicians have built mathematical models to describe real-world phenomena—everything from the path of a cannonball to the growth of a population. These models are powerful because they allow us to predict what will happen next. However, every model has boundaries, and pushing a model outside those boundaries can lead to absurd or impossible predictions. The story of how thinkers discovered and grappled with these limitations is central to the development of modern science and math.

1600s • Galileo's Projectile Model

Galileo modeled projectile motion as a parabola. This worked beautifully for short distances but became inaccurate over long ranges where air resistance and the curvature of the Earth matter.

1798 • Malthus and Population Growth

Thomas Malthus proposed that human population grows exponentially. His model matched short-term data but predicted impossibly large populations when extended far into the future, ignoring resource limits.

1900s • Physics Models Break Down

Newtonian mechanics, one of the most successful models in history, was shown to be inaccurate at very high speeds (near the speed of light) and very small scales (atoms), prompting Einstein and others to build new models.

Today • Models Everywhere

Modern data science uses models to forecast weather, stock prices, and disease spread. In every case, analysts must specify the domain—the range of inputs—over which the model is trustworthy.

The key question this lesson addresses is: Why does a model that works perfectly within a certain range of values become unrealistic or nonsensical outside that range? By the end, you will be able to identify the domain of a model, explain why predictions fail outside it, and communicate your reasoning clearly.

SECTION 2

Core Principles & Definitions

Before diving deeper, let's establish the foundational ideas behind model domain limitations. A model is a simplified mathematical representation of a real situation. The domain of a model is the set of input values for which the model produces meaningful, realistic outputs. Outside this domain, the model may give answers that are physically impossible, logically absurd, or simply wrong.

Mathematical Model

An equation, graph, or function that approximates a real-world relationship. Examples include linear, quadratic, and exponential functions used to describe patterns in data.

Domain of a Model

The specific range of input values (usually x-values) over which the model produces realistic and meaningful outputs. Outside this range, predictions may break down.

Extrapolation

Using a model to make predictions beyond the data it was built from. Extrapolation is risky because the real-world pattern may change outside the observed range.

Physical Constraints

Real-world limits that a mathematical model may ignore—such as the fact that distance, time, and population cannot be negative, or that nothing can travel faster than light.

Reasonableness

The quality of a model's output being sensible in context. A height model that predicts a person is −3 feet tall or 50 feet tall is not reasonable and signals a domain issue.

✦

KEY TAKEAWAY

Think of a model like a map of your town. The map is accurate for navigating your neighborhood, but if you try to use a town map to navigate across the ocean, it becomes useless—it simply wasn't designed for that territory. In the same way, a mathematical model is designed for a specific domain, and using it outside that domain is like driving off the edge of the map.

SECTION 3

Visual Explanation — A Model That Breaks Down

Imagine a ball is thrown upward, and its height in feet after t seconds is modeled by the equation h(t) = −16t² + 48t + 4. This quadratic model works well while the ball is in the air, but what happens if we plug in values of t that are very large or negative? The diagram below shows the full parabola alongside the realistic domain highlighted in green.

The green-shaded region shows the valid domain (0 ≤ t ≤ 3.08 seconds) where the height model gives realistic values. In the red-shaded regions, the model predicts heights for negative time (before the throw) or after the ball has already hit the ground—both are physically meaningless.

Notice how the parabola extends in both directions as a dashed curve. Mathematically, the equation is defined for all values of t, but the real-world scenario only makes sense between the moment the ball is thrown (t = 0) and the moment it hits the ground (t ≈ 3.08 s). For t < 0, we'd be talking about time before the throw happened. For t > 3.08, the ball is on the ground—it can't go to a negative height. The model's domain must be restricted to match the physical reality.

SECTION 4

Mathematical Framework

Let's formalize the idea of restricting a model's domain. When you have a function that models a situation, you need to determine which input values produce reasonable outputs. This usually involves identifying physical or contextual constraints and translating them into inequalities.

GENERAL MODEL FORM

y = f(x), where x ∈ [a, b]

Here, f(x) is the model equation, and [a, b] is the restricted domain. The values a and b are determined by real-world constraints, not by the math alone.

PROJECTILE HEIGHT EXAMPLE

h(t) = −16t² + 48t + 4, where 0 ≤ t ≤ 3.08

The constraint t ≥ 0 comes from the fact that time cannot be negative. The upper bound t ≤ 3.08 comes from solving h(t) = 0, which is when the ball reaches the ground.

LINEAR EARNINGS EXAMPLE

E(h) = 12h, where 0 ≤ h ≤ 40

If E is weekly earnings and h is hours worked at $12/hour, the model only applies from 0 to 40 hours (a standard work week). At h = 100, you'd predict $1,200 in a week—but there are only 168 hours total, and overtime pay changes the rate.

The process for finding a model's realistic domain involves three checks. First, identify mathematical restrictions (e.g., you can't divide by zero or take the square root of a negative number). Second, apply physical constraints (e.g., distance can't be negative, population must be a whole number). Third, consider contextual limits (e.g., a model built from data collected between 2010 and 2020 may not apply to 1950 or 2050).

SECTION 5

How Models Break Down — Types of Domain Violations

Not all domain violations look the same. Sometimes a model predicts a negative quantity that can't be negative in real life, while other times it predicts explosive growth that clearly can't continue forever. The diagram below categorizes the most common ways a model can become unrealistic outside its domain.

The four most common ways a model's domain can be violated: negative impossibility, unbounded growth, wrong trend direction, and undefined / discontinuous values. Each type requires you to restrict the domain differently.

When you encounter any of these situations, it's a signal that the model's domain needs to be restricted. The right approach is to ask: "What range of inputs makes physical and logical sense in this context?" That range becomes your model's realistic domain.

SECTION 6

Worked Example — Restricting a Model's Domain

A small coffee shop tracks the number of customers per hour, C, based on the number of hours after opening at 6 AM. They fit a model: C(t) = −2t² + 20t − 18. Let's determine the realistic domain and explain why the model fails outside it.

Finding the Realistic Domain of a Coffee Shop Model

1
Step 1 — Understand the Context

The variable t represents hours after the shop opens at 6 AM. The shop closes at 6 PM, so t can range from 0 to 12. Also, the number of customers C must be zero or greater—you can't have a negative number of customers.

Contextual limits: 0 ≤ t ≤ 12 and C(t) ≥ 0
2
Step 2 — Find When C(t) = 0

Set the model equal to zero: −2t² + 20t − 18 = 0. Divide both sides by −2: t² − 10t + 9 = 0. Factor: (t − 1)(t − 9) = 0. So t = 1 or t = 9.

C(t) = 0 at t = 1 hour (7 AM) and t = 9 hours (3 PM)
3
Step 3 — Check Where C(t) Is Positive

Since the parabola opens downward (the coefficient of t² is negative), C(t) > 0 between the two zeros: 1 < t < 9. For t < 1 or t > 9, the model predicts negative customers, which is impossible.

C(t) > 0 for 1 ≤ t ≤ 9
4
Step 4 — Combine All Constraints

The shop is open from t = 0 to t = 12, and the model gives non-negative values from t = 1 to t = 9. The realistic domain is the intersection of these constraints.

Realistic domain: 1 ≤ t ≤ 9 (from 7 AM to 3 PM)
5
Step 5 — Explain Why It Fails Outside

At t = 0 (6 AM), the model gives C(0) = −18, which means negative 18 customers—clearly nonsensical. At t = 11 (5 PM), C(11) = −2(121) + 220 − 18 = −40, also negative. Even though the shop is open at these times, the quadratic model doesn't capture what's really happening early and late in the day. The real pattern is likely more complex than a single parabola.

Outside the domain, the model predicts negative customers, violating physical reality.

SECTION 7

Strengths and Limitations of Common Models

Different types of models have different characteristic ways of breaking down outside their domains. Understanding these patterns helps you anticipate problems before they arise. The table below compares three common model types and where each tends to become unrealistic.

Model Type	Strengths Within Domain	How It Fails Outside Domain
Linear (y = mx + b)	Simple, easy to interpret. Works well for short-range predictions when data shows a steady rate of change.	Predicts unlimited growth or decline forever. A linear model of tree height would eventually predict a tree miles tall or with negative height.
Quadratic (y = ax² + bx + c)	Captures rise-and-fall or fall-and-rise patterns. Useful for projectiles, profit curves, and area problems.	Predicts negative values beyond the zeros. Also diverges toward ±∞, which rarely matches reality over large domains.
Exponential (y = a × bˣ)	Models rapid growth or decay accurately in early stages (bacteria, radioactive decay, compound interest).	Predicts astronomically large values. An exponential bacteria model might predict more bacteria than atoms in the universe after enough time.

Common model types and their domain limitations

✦

KEY TAKEAWAY

Think of a model like a pair of shoes that fits you perfectly right now. A linear model is like shoes that are comfortable for a short walk but painful for a marathon. A quadratic model is like shoes that work great going up a hill and back down, but would be useless on flat terrain. An exponential model is like rocket-powered shoes—amazing at first, but eventually they'll send you to the moon when you just wanted to cross the street. Every model has a "comfort zone" where it fits reality, and stepping outside that zone leads to problems.

SECTION 8

Connection to Advanced Modeling

The idea of domain limitations doesn't go away as math gets more advanced—it becomes even more important. In higher-level courses, you'll encounter piecewise functions that use different equations for different parts of the domain, logistic models that cap growth at a carrying capacity, and regression analysis where statisticians quantify exactly how far you can trust a model's predictions. The foundational skill you're building now—recognizing when a model stops being realistic—is the same skill that data scientists and engineers use every day.

Concept Now (Grade 9)	Advanced Version (Later Courses)
Restrict domain to avoid negative or impossible outputs	Use piecewise functions that switch equations at boundary points
Recognize that exponential growth can't continue forever	Apply logistic growth models with carrying capacity: y = L / (1 + e⁻ᵏˣ)
Explain in words why extrapolation is risky	Calculate confidence intervals that widen as you extrapolate further from the data
Check if a model's output makes sense in context	Perform residual analysis and goodness-of-fit tests to evaluate model accuracy

How domain-limitation ideas evolve in later courses

The good news is that the critical thinking skill at the heart of this lesson—asking "Does this answer make sense in the real world?"—is something you can apply right now, in every class, and it will serve you throughout your academic and professional life.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A student creates a linear model that predicts the temperature of a cup of coffee over time: T(t) = −5t + 180, where T is in °F and t is in minutes. The student uses the model to predict the temperature at t = 40 minutes and gets T = −20°F. Explain why this prediction is unrealistic and identify the type of domain violation.

Show answer

The model predicts −20°F, which is far below room temperature and physically impossible for a cup of coffee sitting in a room. This is a negative impossibility / wrong trend violation. In reality, the coffee's temperature will decrease and then level off near room temperature (around 70°F), not continue dropping below zero. The linear model is only realistic while the coffee is still significantly warmer than the room—probably for the first 20–25 minutes at most. After that, the cooling rate slows down, and a linear model can't capture this behavior.

Question 2: BASIC CALCULATION

A model for the height of a rocket (in meters) is h(t) = −4.9t² + 49t, where t is time in seconds after launch. Find when the rocket returns to ground level, and state the realistic domain of this model.

Show answer

Set h(t) = 0: −4.9t² + 49t = 0. Factor out t: t(−4.9t + 49) = 0. So t = 0 or −4.9t + 49 = 0, giving t = 49 ÷ 4.9 = 10 seconds. The rocket launches at t = 0 and returns to the ground at t = 10. The realistic domain is 0 ≤ t ≤ 10 seconds. For t < 0 (before launch) or t > 10 (after landing), the model predicts negative heights, which would mean the rocket is underground—clearly unrealistic.

Question 3: INTERMEDIATE

A biologist models a bacteria population with P(t) = 100 × 2ᵗ, where t is time in hours and P is the number of bacteria. The petri dish can hold a maximum of 50,000 bacteria. Find the domain of this model considering this constraint, and explain what happens to the model beyond that domain.

Show answer

We need P(t) ≤ 50,000: 100 × 2ᵗ ≤ 50,000. Divide both sides by 100: 2ᵗ ≤ 500. Taking log base 2: t ≤ log₂(500) ≈ 8.97 hours. Since t also can't be negative (we start timing at t = 0), the realistic domain is 0 ≤ t ≤ 8.97 hours. Beyond t ≈ 9 hours, the model predicts the population exceeds the dish's capacity. In reality, the bacteria would run out of space and nutrients, and growth would slow dramatically. This is an unbounded growth violation—the exponential model ignores the physical limits of the environment.

Question 4: APPLIED

A ride-share company models its profit (in thousands of dollars) based on the number of drivers, d, using the equation P(d) = −0.5d² + 30d − 200. The company currently has between 10 and 50 drivers. Determine the realistic domain of this model by finding where profits are non-negative, and explain what the model's predictions mean outside that domain.

Show answer

Set P(d) = 0: −0.5d² + 30d − 200 = 0. Multiply by −2: d² − 60d + 400 = 0. Using the quadratic formula: d = (60 ± √(3600 − 1600)) / 2 = (60 ± √2000) / 2 = (60 ± 44.72) / 2. So d ≈ 7.64 or d ≈ 52.36. Profits are non-negative between approximately d = 8 and d = 52 drivers. Intersecting with the company's range of 10 to 50, the realistic domain is 10 ≤ d ≤ 50. With fewer than about 8 drivers, the company loses money (too few to cover overhead). With more than about 52, costs outweigh revenue (too many drivers competing for limited riders). Beyond d = 52, the model also predicts increasingly negative profit, which could represent real losses—but the quadratic shape may not accurately capture the business dynamics at that scale.

Question 5: CRITICAL THINKING

Two students each create a model for the same data set showing how a plant's height (in cm) changes over 8 weeks. Student A uses a linear model: H(w) = 3w + 2. Student B uses a quadratic model: H(w) = −0.25w² + 5w − 1. Both models fit the data well from week 1 to week 8. Compare the predictions of both models at week 0, week 12, and week 20. Which model is more realistic at each point, and what does this tell you about choosing models?

Show answer

At week 0: Linear gives H = 2 cm, Quadratic gives H = −1 cm. The linear model is more realistic because a plant can start at 2 cm (seed sprouting), but −1 cm has no physical meaning. At week 12: Linear gives H = 38 cm, Quadratic gives H = −0.25(144) + 60 − 1 = 23 cm. Both are plausible, but the quadratic suggests growth is slowing, which is often more realistic for plants. At week 20: Linear gives H = 62 cm, Quadratic gives H = −0.25(400) + 100 − 1 = −1 cm. Now the quadratic predicts negative height, which is impossible, while the linear prediction of 62 cm might be too large but is at least possible. This comparison reveals that no single simple model may be realistic across the entire range. The linear model fails in the long run (unbounded growth), and the quadratic fails both before and after its useful zone (negative values). Choosing a model requires matching the model type to the domain you care about, and being honest about where each model breaks down.

SUMMARY

Lesson Summary

Summary

A mathematical model is only as useful as the domain over which it produces realistic results. Every model is built from assumptions, and those assumptions only hold within a certain range of input values. When you extrapolate beyond the domain, you may encounter negative impossibilities, unbounded growth, wrong trend directions, or undefined values—all signs that the model has left its zone of reliability.

To find a model's realistic domain, check for mathematical restrictions (like division by zero), physical constraints (like non-negative distance or population), and contextual limits (like the time frame of the data). The most important skill is asking: "Does this answer make sense in the real world?" If not, you've likely stepped outside the model's domain.