Binomial Distribution (AI HL)

A binomial distribution counts the number of successes in nn independent trials, where each trial has the same probability of success pp, and is written X∼B(n,p)X\sim B(n,p). Its mean is npnp and its variance is np(1−p)np(1-p). Probabilities such as P(X=k)P(X=k) and P(X≤k)P(X\le k) are found with the GDC, once the conditions for the model have been checked. It's part of the broader Probability & Distributions topic.

16 questions on this sub-topic.

Syllabus

CodeSyllabus content
SL4.8Binomial distribution. Mean and variance of the binomial distribution. In examinations, binomial probabilities should be found using available technology. Formal proof of mean and variance is not required.

Official IB source: Mathematics: applications and interpretation subject brief (PDF, ibo.org).

Skills practised on this page:

Mean and variance

Neither formula is in the formula booklet, so both are worth memorising.

Mean

X∼B(n,p)X\sim B(n,p): E(X)=npE(X)=np

Not in the formula booklet. The average number of successes over nn trials.

Variance

Var(X)=np(1−p)\text{Var}(X)=np(1-p)

Not in the formula booklet. Note the (1−p)(1-p) factor - it's easy to drop under time pressure.

Need the full syllabus wording and formula-booklet reference table? See Probability & Distributions. For calculator routines, see the parent topic's GDC guidance.

Key terms

What conditions must hold for a binomial model?

There must be a fixed number of trials, each with only two outcomes (success or failure), the same probability of success each time, and independent trials.

e.g. Counting sixes in 1010 rolls of a fair die fits B(10,16)B\left(10,\dfrac16\right).

How do you find P(X≥k)P(X\ge k)?

Use the complement: P(X≥k)=1−P(X≤k−1)P(X\ge k)=1-P(X\le k-1). This works because XX takes whole-number values only.

e.g. For X∼B(10,0.5)X\sim B(10,0.5), P(X≥3)=1−P(X≤2)=1−561024≈0.945P(X\ge3)=1-P(X\le2)=1-\dfrac{56}{1024}\approx0.945.

Why is P(X<k)P(X<k) not the same as P(X≤k)P(X\le k)?

Because XX is discrete, P(X<k)P(X<k) excludes the value kk, so P(X<k)=P(X≤k−1)P(X<k)=P(X\le k-1). Misreading the inequality symbol is a common source of errors.

e.g. P(X<5)=P(X≤4)P(X<5)=P(X\le4), but P(X≤5)P(X\le5) includes the extra value X=5X=5.

How do you find the number of trials needed for at least one success?

Write 1−(1−p)n1-(1-p)^n greater than the target probability, then find the smallest whole number nn that satisfies it using the GDC.

e.g. With p=0.5p=0.5 and a target above 0.90.9, 1−0.5n>0.91-0.5^n>0.9 gives 0.5n<0.10.5^n<0.1, so the least nn is 44.

Worked examples

1
Easy
GDC
[4 marks]

X∼B(40,0.3)X \sim B(40, 0.3).

(a) Find E(X)E(X).

(b) Find Var(X)\text{Var}(X).

Worked solution

(a) E(X)=np=40(0.3)E(X) = np = 40(0.3) M1
E(X)=12.E(X) = 12. A1

(b) Var(X)=np(1−p)=40(0.3)(0.7)\text{Var}(X) = np(1-p) = 40(0.3)(0.7) M1
Var(X)=8.4.\text{Var}(X) = 8.4. A1On the GDC. (a) Find E(X). (b) Find Var(X).

TI-84 Plus CE: (a) 2nd mode then type 40*0.3 enter Then (b) 2nd mode then type 40*0.3*0.7 enterCasio fx-CG50: (a) MENU Run-Matrix then type 40*0.3 EXE Then (b) MENU Run-Matrix then type 40*0.3*0.7 EXECasio fx-CG100: (a) HOME Calculate then type 40*0.3 EXE Then (b) HOME Calculate then type 40*0.3*0.7 EXETI-Nspire CX: (a) home Calculator then type 40*0.3 enter Then (b) home Calculator then type 40*0.3*0.7 enterFull GDC guide: Enter a calculation on the home screen Topic page

M1 npnp A1 Correct answer of 1212 M1 np(1−p)np(1-p) A1 Correct answer of 8.48.4
2
Medium
GDC
[4 marks]

In a factory, 4% of items are defective. A sample of 50 is taken. Let XX be the number defective, modelled as binomial.

(a) Find P(X=0).P(X = 0).

(b) Find P(X≤3).P(X \le 3).

Worked solution

(a) P(X=0)=(0.96)50P(X=0) = (0.96)^{50} M1
≈0.130.\approx 0.130. A1

(b) P(X≤3)≈0.861.P(X \le 3) \approx 0.861. M1 A1On the GDC. (a) Find P(X=0)P(X=0) for X∼B(50, 0.04)X\sim B(50,\ 0.04). (b) Find P(X≤3)P(X\le 3) for X∼B(50, 0.04)X\sim B(50,\ 0.04).

TI-84 Plus CE: (a) 2nd vars A:binompdf( trials 50 p 0.04 x 0 enter Then (b) 2nd vars B:binomcdf( trials 50 p 0.04 x 3 enterCasio fx-CG50: (a) MENU Statistics F5 F5 F1 F2 ▼ x 0 EXE Numtrial 50 EXE p 0.04 EXE ▼ F1 Then (b) MENU Statistics F5 F5 F2 F2 ▼ x 3 EXE Numtrial 50 EXE p 0.04 EXE ▼ F1Casio fx-CG100: (a) HOME Distribution Binomial OK Tail (X=) x 0 OK Numtrial 50 OK p 0.04 OK Execute OK Then (b) HOME Distribution Binomial OK Tail LX (x or lower) x 3 OK Numtrial 50 OK p 0.04 OK Execute OKTI-Nspire CX: (a) menu Probability Distributions Binomial Pdf Num Trials 50 tab Prob Success 0.04 tab X Value 0 enter Then (b) menu Probability Distributions Binomial Cdf Num Trials 50 tab Prob Success 0.04 tab Lower Bound 0 tab Upper Bound 3 enterFull GDC guide: Binomial distribution Topic page

M1 P(X=0)P(X=0) A1 Correct answer of ≈0.130\approx0.130 M1 Binomcdf A1 Correct answer of ≈0.861\approx0.861
3
Hard
GDC
[5 marks]

A seed germinates with probability 0.8. A gardener wants the probability of at least one germinating to exceed 0.999.

(a) Write an inequality for the number of seeds nn.

(b) Find the least nn.

Worked solution

(a) P(≥1)=1−(0.2)n>0.999⇒(0.2)n<0.001.P(\ge 1) = 1 - (0.2)^n > 0.999 \Rightarrow (0.2)^n < 0.001. M1 A1

(b) n>ln⁡0.001ln⁡0.2≈4.29n > \dfrac{\ln 0.001}{\ln 0.2} \approx 4.29, so nn M1
n=5.n = 5. A1 A1On the GDC. Find the least whole number x for which 1−(1−0.8)x1-(1-0.8)^{x} is more than 0.9990.999. The calculator shows about 5. The calculator steps below are for (b).

TI-84 Plus CE: Y= Y1= 1-(1-0.8)^x 2nd window TblStart 1 enter ΔTbl 1 enter 2nd graph scroll down to the first x where the value passes 0.999 and read that xCasio fx-CG50: MENU Table Y1 1-(1-0.8)^x EXE F5 Start 1 EXE End 11 EXE Step 1 EXE EXIT F6 scroll down to the first x where the value passes 0.999 and read that xCasio fx-CG100: HOME Graph&Table y1 1-(1-0.8)^x EXE ▲ OK Set Table Domain Range Start 1 EXE End 11 EXE Step 1 EXE Display Table OK scroll down to the first x where the value passes 0.999 and read that xTI-Nspire CX: Graphs page, f1(x) = 1-(1-0.8)^x enter menu Table Split-screen Table menu Table Edit Table Settings Table Start 1 Table Step 1 OK, scroll down to the first x where the value passes 0.999 and read that xFull GDC guide: Make a table of values Topic page

M1 Complement inequality A1 (0.2)n<0.001(0.2)^n<0.001 M1 Take logs A1 Correct answer of ≈4.29\approx4.29 A1 n=5n=5

Common mistakes

  • Assuming a variable is binomial without checking independence. The binomial model needs a fixed number of independent trials with a constant success probability - sampling without replacement from a small population breaks this and the model no longer applies exactly.
  • Dropping the (1−p)(1-p) factor in the variance. Var(X)=np(1−p)\text{Var}(X)=np(1-p), not npnp - it's easy to write down the mean formula twice under exam pressure and forget the extra factor.
  • Using P(X≤k)P(X\le k) when the question asks for P(X<k)P(X<k) or P(X≥k)P(X\ge k). Translate the wording into the correct inequality on paper before reaching for binomcdf - "at most 3" is X≤3X\le3, "fewer than 3" is X≤2X\le2, and "at least 3" needs 1−P(X≤2)1-P(X\le2).

Using your GDC

Every step below is a real button sequence, not a vague "use your calculator" hint - covering the TI-84 Plus, TI-Nspire, and Casio fx-9860/fx-CG50. Pick your model to filter down to just the steps that apply to you.

Show steps for:
Inverse normal (find the value for a given probability)

The reverse of a normal probability: given a percentage, find the cut-off value (e.g. the mark for the top 10%).

  1. Work out the area to the LEFT of the value you want.
  2. 2nd → VARS (DISTR) → invNorm(area, μ, σ). Newer OS lets you pick the tail.TI-84
  3. menu → Probability → Distributions → Inverse Normal; enter the area, μ and σ.Nspire
  4. Main menu → Statistics → DIST → NORM → InvN; set the tail and enter area, σ, μ.Casio

Tip: invNorm needs the area to the LEFT. For "top 10%", use area = 0.90; for "bottom 25%", use area = 0.25.

Find an unknown mean or standard deviation (normal)

Given a probability and one parameter, work back to the missing μ\mu or σ\sigma - a standard exam twist.

  1. Turn the probability into a zz-value using the inverse normal with μ=0\mu = 0, σ=1\sigma = 1.
  2. invNorm(area-to-left, 0, 1) gives zz; then solve z=(x−μ)/σz = (x-\mu)/\sigma for the unknown.TI-84
  3. Use Inverse Normal with μ=0\mu=0, σ=1\sigma=1 to get zz, then solve the standardising equation for μ\mu or σ\sigma.Nspire
  4. DIST → NORM → InvN with μ=0\mu=0, σ=1\sigma=1 gives zz; substitute into z=(x−μ)/σz=(x-\mu)/\sigma.Casio
  5. If two probabilities are given, form two equations and solve them simultaneously for μ\mu and σ\sigma.

Tip: Always sketch the curve first so you can tell which side of the mean each value should fall on.

See the full GDC guide for more calculator models and topics.

Quick questions

What is the mean of a binomial distribution?

If X∼B(n,p)X \sim B(n,p), the mean is E(X)=npE(X) = np.

What is the variance of a binomial distribution?

If X∼B(n,p)X \sim B(n,p), the variance is Var(X)=np(1−p)\text{Var}(X) = np(1-p).

How do I find the probability of no successes?

With no successes, every trial must fail, so P(X=0)=(1−p)nP(X=0)=(1-p)^n. For X∼B(5,0.2)X\sim B(5,0.2), P(X=0)=0.85≈0.328P(X=0)=0.8^5\approx0.328.

Can I use a binomial model if the trials are not independent?

No, because the probability of success changes from trial to trial. For example, drawing cards from a pack without replacement changes the chance that the next card is a heart.

Is the variance of a binomial distribution always less than the mean?

Yes, when 0<p<10<p<1, because np(1−p)np(1-p) is npnp multiplied by 1−p1-p, which is less than 11. For X∼B(20,0.4)X\sim B(20,0.4), the mean is 88 and the variance is 4.84.8.

Ready to practise properly?

16 binomial-distribution questions, marked instantly like the real exam.