On this concept: Formula

Not sure which lesson to do first? Find your level and get a plan →

Math · Advanced Functions · Grade 9-12

Feedback

Feedback is when a system's output loops back to influence its own next input: positive feedback amplifies changes, negative feedback stabilizes them.

xn+1=f(xn) (output feeds back as the next input)

Quick Answer

Feedback is when a system's output loops back to influence its own next input: positive feedback amplifies changes, negative feedback stabilizes them. You model it with a recurrence xn+1=f(xn): the same rule is applied again and again, each result becoming the next starting point. A savings balance earning interest on its interest is positive feedback; a thermostat switching the heat off when the room is warm is negative feedback. Use it to predict where a repeated process heads: off without limit, toward a steady value, or into oscillation.

Why This Matters

Many systems are driven by their own past: populations, bank balances, room temperatures, a microphone near its speaker. Spotting the loop tells you to iterate the rule instead of applying it once, and the type of feedback tells you the long-run outcome before you compute it. Mistake a self-correcting loop for a runaway one and you predict explosions that never come; mistake the reverse and you miss real ones.

Intuitive Explanation

Start with the rule xn+1=1.5xn and x0=10. Feed each output back in: 10, 15, 22.5, 33.75. Every round the change is bigger than the one before, because the rule acts on a bigger number each time. That is positive feedback: a change produces more of the same change.

Now try xn+1=0.5xn+10 from x0=0: 0, 10, 15, 17.5, 18.75. The values climb toward 20 and slow down as they get close. Start above it, at 40, and you get 30, 25, 22.5, falling toward the same 20. This is negative feedback: the farther x is from 20, the bigger the correction pulling it back, from either side. A thermostat does the same thing to a room's temperature.

A loop is not the same as a chain. Converting Celsius to Fahrenheit and then rounding uses two different steps once each, and nothing comes back around. In feedback one rule runs repeatedly on its own results. Ask where each output goes: if it becomes the starting value for the same rule, you have a loop.

Think of it like this. A microphone held near its own speaker: a small sound goes in, comes out louder, is picked up again and comes out louder still, until the squeal. That runaway is positive feedback. A thermostat is the negative kind: too warm switches the heat off, too cold switches it on, so every pass pushes back toward the setting.

Core idea. In feedback, what a system puts out comes back as part of what it takes in next, either amplifying or damping the change.


When to Use

Think feedback when a quantity's next value depends on its current value through the same rule every round: each year a balance earns interest on itself, a population breeds from its current size, a thermostat reacts to the temperature it helped create. Write the rule as xn+1=f(xn) and iterate. If different functions are applied once each in a fixed order, it is a composition chain; if you need to know whether the loop settles at a fixed point, use stability.

Tip. Ask: Does the system's output get fed back in as the input for the next step?

How to Recognize It

Before modelling something as feedback, look at what each new value is computed from, and at whether each round pushes the value further out or back toward a target.

  1. Does a larger current value make the next change larger in the same direction, as interest on a bigger balance does?

    If yes, it is positive feedback: changes amplify and the values run away until something limits them. If a larger value makes the next change pull it back down, it is negative feedback.

  2. Is the system reacting to its distance from a target, like a thermostat or cruise control?

    If yes, it is negative feedback, and the value should settle near the target; whether it settles smoothly or overshoots is a question of stability. With no target, look for runaway growth instead.

  3. Are several different rules applied once each, in a fixed order?

    If yes, that is a composition chain, worked out once from the inside out. Feedback reapplies one rule to its own result, so you compute x1, then x2, then x3, and so on.

  4. Does the question ask where the values end up after many rounds, rather than the next value?

    If yes, look for a fixed point where f(x∗)=x∗ and test its stability. If it asks for the next few values, just iterate the rule step by step.

Feedback compared with Composition chains, Stability of an equilibrium, and Recurrence / iteration

Each row involves applying rules repeatedly or in sequence, so the vocabulary overlaps. Ask whether one rule keeps acting on its own result or different rules run once, and whether you want the next values or the long-run outcome.

Feedback

Meaning
Use this when the output of one step becomes the input of the next through a repeated rule.
Key test
Does the system's output get fed back in as the input for the next step?
Formula
xn+1=f(xn) (output feeds back as the next input)
Example
A savings rule is xn+1=1.1 xn (each year's balance feeds back, growing 10%). Starting at x0=100, find x2 and name the feedback type.

Composition chains

Meaning
Different functions applied in sequence, not the same one looping.
Key test
Use when distinct functions f,g,h are chained once, not iterated.
Formula
f(g(h(x)))
Example
Convert units then tax then round

Stability of an equilibrium

Meaning
Whether a feedback loop settles at or flees from a fixed point.
Key test
Use to judge what a feedback system does long-term near x∗.
Formula
∣f′(x∗)∣<1 stable
Example
Ball returning to a bowl's center

Recurrence / iteration

Meaning
The mathematical machinery of repeatedly applying f — the how, not the loop concept.
Key test
Use to actually compute successive values x1,x2,…
Formula
xn+1=f(xn)
Example
xn+1=2xn from x0=1

Formula & Notation

xn+1=f(xn) (output feeds back as the next input)
xn+1=f(xn); positive feedback: ∣f′(x∗)∣>1 (amplifies perturbations); negative feedback: ∣f′(x∗)∣<1 (dampens perturbations near equilibrium x∗)

How to read it: xn+1=f(xn) denotes a recurrence where the output of step n becomes the input of step n+1.

Worked Examples

Example 1 — Iterate a loop

Easy

A savings rule is xn+1=1.1 xn (each year's balance feeds back, growing 10%). Starting at x0=100, find x2 and name the feedback type.
Show the steps and the answer
  1. Apply the rule once: x1=1.1⋅100=110.

    The starting balance is the first input.

  2. Feed the output back in: x2=1.1⋅110=121.

    Year 2 starts from year 1's result, which is what makes this a loop rather than a single calculation.

  3. Compare the gains: 10 in year 1, 11 in year 2.

    The second gain is bigger because it is earned on a bigger balance: the change feeds more change.

  4. Classify: the multiplier 1.1>1 amplifies, so this is positive feedback.

    Positive feedback keeps pushing the balance further in the direction it is already moving.

Answer: x2=121, positive feedback

Takeaway: Iterate by feeding each output into the same rule; changes that grow from round to round signal positive feedback.

Example 2 — A chain, not a loop

Standard

You convert Celsius to Fahrenheit, then round to the nearest degree, once. Is that feedback?
Show the steps and the answer
  1. Trace one value: C=21.7 gives F=1.8(21.7)+32=71.06, which rounds to 71.

    Following an actual number shows where each output goes.

  2. Ask where the 71 goes next: nowhere, since it is the final answer.

    In a feedback loop the output would return as the next input to the same rule.

  3. Write the process as round(F(C)).

    Two different functions, each used once in a fixed order, form a composition.

Answer: No — it's composition

Takeaway: A chain of different steps used once is composition; feedback needs one rule to receive its own output.

Example 3 — Find the mistake: which kind of feedback?

Application

A heater follows Tn+1=Tn+0.5(20−Tn) each minute, where Tn is the room temperature in °C. Starting at T0=12, Ana writes: "The temperature goes up every minute, so this is positive feedback and the room will keep heating forever." Find the mistake and correct it.
Show the steps and the answer
  1. Iterate: T1=12+0.5(8)=16, T2=16+0.5(4)=18, T3=18+0.5(2)=19.

    The rises are 4, 2, 1, each half the last, so the warming is slowing as the room nears 20.

  2. Start above the setting instead: T0=26 gives T1=26+0.5(−6)=23.

    Above 20 the same rule cools the room, so the correction always points toward 20.

  3. Name the error: Ana judged the loop by the direction the numbers moved, not by how the change responds to the gap 20−Tn.

    Negative feedback opposes the deviation; here each change is half the gap, pointing back toward 20 from either side.

  4. Find where it settles: T=T+0.5(20−T) forces 20−T=0, so T=20.

    At 20 the correction is zero, so the temperature stops changing.

Answer: Negative feedback: the temperature rises 16, 18, 19 and levels off at 20 °C instead of heating forever.

Takeaway: Classify feedback by whether it pushes the system away from a level or back toward it, not by whether the numbers happen to be rising.

Where You'll See It

  • Microphone squeal

    If a microphone picks up its own speaker and each trip around the loop makes the sound 1.2 times louder, a tiny noise becomes 1.220≈38 times louder after 20 trips, a fraction of a second. Sound engineers stop it by lowering the gain below 1 or moving the microphone.

  • Body temperature control

    When your body warms above about 37 °C you sweat, and evaporation cools you; when it cools you shiver, which makes heat. Either way the response opposes the change, a negative feedback loop that holds your temperature within about a degree.

  • Melting sea ice

    White ice reflects most sunlight, while dark open water absorbs it. When ice melts, the exposed water absorbs more heat, which melts more ice. Each change produces more of the same change: positive feedback that speeds up warming in the Arctic.

Common Mistakes

  1. The slip: Swapping positive and negative feedback

    Instead: positive amplifies/destabilizes, negative corrects/stabilizes.

  2. The slip: Treating a feedback loop like a one-shot function

    Instead: the rule is applied repeatedly, each output re-entering as input.

  3. The slip: Assuming positive feedback always grows without limit

    Instead: real loops often hit saturation or get checked by negative feedback.


Mini Practice

Try these on your own. Tap Reveal when you want to check.

  1. A loop uses xn+1=2xn−1 with x0=3. Find x1, x2 and x3.

    Hint: Each output becomes the next input.

  2. A thermostat uses Tn+1=Tn+0.5(20−Tn) with T0=12°C. Find T1, T2 and T3. Is this positive or negative feedback?

    Hint: Each step closes part of the gap to 20.

  3. Iterate xn+1=0.8xn+10 from x0=0 to find x1, x2 and x3. What value is it heading toward?

    Hint: For the long-run value, find x with x=0.8x+10.

  4. A microphone picks up its own speaker, and each loop amplifies the sound by a factor of 1.5. Starting at 2 units, how loud is it after 4 loops? Name the feedback type.

    Hint: Multiply by 1.5 four times.

  5. You apply f(x)=x+3, then g(x)=2x, then h(x)=x2, once each, to the input 1. Is that feedback?

    Hint: Is the same rule used again on its own output?

  6. A student says 'The loop xn+1=xn−0.3(xn−100) is positive feedback, because it keeps changing x.' Find and correct the mistake. Try x0=200.

    Hint: Does each step move x toward 100 or away from it?

Want the full set?

50 practice questions for this concept — free to try, every one with a complete worked solution showing the why, not just the answer.

Frequently Asked Questions

Why do I end up at the same number if I keep pressing cos on a calculator?

Each press feeds the last output back in, so you are running xn+1=cos⁡xnx_{n+1} = \cos x_n. In radian mode, any starting number ends up near 0.739085, the one number whose cosine is itself. Near that value the cosine graph has a slope of about −0.67-0.67, smaller than 1 in size, so each press shrinks the error and the values zigzag inward. It is negative feedback found by accident, and stability explains why it settles.

What does feedback look like in the climate?

One well-known loop involves sea ice. Ice reflects most sunlight, while open water absorbs most of it. When warming melts some ice, more dark water is exposed, it absorbs more heat, and more ice melts: positive feedback. A loop pushing the other way is radiation to space: a warmer Earth gives off more heat, which pulls the temperature back down. Much of climate science is about how strong these opposing loops are compared with each other.

Can negative feedback make a system swing back and forth?

Yes, when the correction is too strong or comes too late. A shower is the familiar case: the water is cold, you turn the tap far toward hot, the hot water arrives seconds later and scalds you, so you swing the tap back too far. In the rule xn+1=x∗+c(xn−x∗)x_{n+1} = x^* + c(x_n - x^*), a cc between −1-1 and 0 makes the error flip sign and shrink each round; a cc below −1-1 makes it flip and grow.

How is a feedback rule different from a formula like y=2x?

A direct formula jumps straight to any step: 2102^{10} gives the tenth value at once. A feedback rule such as xn+1=2xnx_{n+1} = 2x_n only says how to get from one step to the next, so you need a start, x0x_0, and you walk forward. Some loops can be turned into a direct formula, as this one becomes xn=2nx0x_n = 2^n x_0, but many, including most population models with crowding, have no simple direct formula and must be iterated.

Learning Path

Feedback

You are here

You need Exponential Function, because the simplest positive loop, xn+1=bxn, produces xn=x0bn. Feedback adds the idea of a rule acting on its own output, and the split into amplifying and correcting loops. Next, Introduction to Differential Equations describes the same loops in continuous time, as in dPdt=kP, where the rate of change depends on the current amount.

See Also