Neither of the two: probability through the complement

12 min · 1 drill

In the textbook: Ch. 5.1 Grundbegriffe

“Weder … noch”, “keines”: this is how a German task asks about what does not happen. The task almost never gives this share. You get it by subtracting: first work out “at least one”, then take it away from the whole. The point is not to subtract the common part of the two events twice.

For example: what share of customers buys neither bread nor milk, how many residents neither read the local paper nor listen to local radio.

Back to the dice game from the chapter's first article: the token moves forward on a 2, 4, 5 or 6, that is, on an even number or a number greater than 3. Now the other case matters: it is neither, and the token stays put.

With what probability does the token stay put?

Here a look at the faces is enough: the token stays put only on a 1 or a 3, probability 26. Subtracting both chances from one, 1−36−36=0, does not work: faces 4 and 6 are subtracted twice. In a task with percentages there are no faces to look at, and a way is needed that works there too.

Check yourself. A task at exam level: in a group, 40 % of the students go to the gym, 30 % to the pool and 10 % to both. What share of the students goes neither to the gym nor to the pool?

The basics: “at least one” and “not”

“Neither … nor” is computed from two things covered in the article “Events: and, or, not”. Here they are on the same die:

  • Events are sets of faces: A “even”, A={2,4,6}, and B “greater than 3”, B={4,5,6}.
  • “Not”: the complement A′ is every outcome except those of A: A′={1,3,5}.
  • “At least one” is the union A∪B={2,4,5,6}, with probability 46.
  • Why not 36+36: faces 4 and 6 lie in both A and B, so they are subtracted once. This is the addition rule: P(A∪B)=P(A)+P(B)−P(A∩B)=36+36−26.

The token stays put when a face comes up that lies neither in A nor in B: that is the event A′∩B′. What is left is to find its probability without subtracting the common part twice.

Neither A nor B: through the complement

On the die the faces could be counted one by one. In the exam task there are no faces, only percentages, and “neither … nor” is put together from them in two steps:

Exam-style task

Von den Kundinnen und Kunden einer Bank besitzen 45 % eine Kreditkarte (Ereignis A) und 30 % ein Wertpapierdepot (Ereignis B); 12 % besitzen beides.

Mit welcher Wahrscheinlichkeit besitzt eine zufällig gewählte Person weder eine Kreditkarte noch ein Wertpapierdepot?

  1. a)0.63
  2. b)0.37
  3. c)0.41
  4. d)0.46
  5. e)0.25

The question “neither … nor” (“weder … noch”) and three given shares: one per event and “both”. First compute “at least one” with the addition rule, then subtract it from 1.

Back to the die.

A dieCompare how many faces lie outside circle A and how many lie outside both circles.

no even number came up: everything outside circle A

A die. Outcomes as dots in a box, events as regions over them. A holds 2, 4, 6; B holds 4, 5, 6. Lit: not A.ΩAB
Step 1 of 4Not A

The same six outcomes of the throw, together called Ω, and the same circles: A “an even number comes up”, B “a 4, 5 or 6 comes up”.

“Not even” is 1, 3, 5: everything left of Ω without A. That is 6−3=3 outcomes, and as shares

P(A′)=6−36=1−36=1−P(A)

The picture shows 100 bank customers, all together Ω: circle A the card, circle B the account.

A: credit card, P(A)=0.45. B: securities account, P(B)=0.30. “Have both” is A∩B, 0.12.

“Neither a credit card nor a securities account” is A′∩B′, everything outside A∪B.

0.45+0.30−0.12=0.63. That is option (a), but it answers “at least one”, not the task's question.

The complement of the union: 1−0.63=0.37, option (b).

The numbers of the parts in the picture: card only 33, account only 18, both 12. Inside the circles 33+18+12=63, outside 37. It matches.

Why is 1−0.45−0.30=0.25 wrong, although the number looks plausible? Hint: which two numbers contain the 12 %?

In an office 70 % of the staff drink coffee, 40 % tea, and 25 % neither coffee nor tea.

What share of the staff drinks coffee and tea?

The picture shows 100 employees, all together Ω: circle A drinks coffee, circle B tea.

“Neither coffee nor tea” is the complement of “at least one”: 100−25=75 %.

75=70+40−P(both), so P(both)=110−75=35 %.

The numbers in the picture: coffee only 70−35=35, tea only 40−35=5, both 35, outside 25. Together 35+5+35+25=100.

A village has 200 residents. 120 read the local paper, 90 listen to the local radio, 60 do both. How many residents neither read the paper nor listen to the radio? At least one: 120+90−60=150. Everyone else does neither: 200−150=50 residents, a share of 0.25. You cannot subtract both numbers from 200: 200−120−90=−10 residents, because the 60 are subtracted twice.

Your turn1 of 1

In a town 55 % of families have a car, 35 % a bike and 20 % both. What share of families has neither a car nor a bike?

Traps. The wrong options of the exam task and the mistakes in the “neither … nor” question come about like this:

  • Stopping at the union, option (a), 0.63: this is “at least one of the two”. “Neither … nor” asks for the complement, 1−0.63.
  • Subtracting both events from the whole, option (e), 0.25: 1−0.45−0.30=0.25. The number is plausible and does not give the mistake away. The intersection was subtracted twice; add it back once: 0.25+0.12=0.37. With large shares the same mistake gives a negative number, and then you see it at once.
  • Multiplying the probabilities: taking 0.45⋅0.30=0.135 instead of the given 12 % gives 1−(0.45+0.30−0.135)=0.385. That option does not exist, and the intersection need not be computed at all: it is in the task.
  • “Not both” instead of “none”: 1−0.12=0.88 are all who lack at least one of the two. “Weder … noch” leaves only those who have neither.
  • The other options, (c) 0.41 and (d) 0.46, come from no sensible calculation: the check with the groups gives 0.37.

On the formula sheet: P(A′)=1−P(A) (complement). “Neither … nor” is 1−P(A∪B), and the union is computed first by the addition rule.

At a glance: neither of the two

The topic in a minute

  • Complement: A′ is everything except A, and P(A′)=1−P(A).
  • Neither … nor (“weder … noch”, “keines”): P(A′∩B′)=1−P(A∪B). The union first by the addition rule, then the rest up to one.
  • Do not subtract both at once: 1−P(A)−P(B) subtracts the intersection twice.
  • The other way round: if “neither … nor” is given, the union equals 1−P(A′∩B′), and the addition rule gives the intersection: P(A∩B)=P(A)+P(B)−P(A∪B).
  • Neither A nor B through the complement

    The question asks “neither … nor”, “weder … noch”, “keines”.
    P(A′∩B′)=1−P(A∪B)

    Trap: subtracting both events from the whole and not adding the intersection back.

The six faces of the die were counted one by one. Hundreds of cases cannot be counted that way; you count selections: “Probability as a share: how many cases fit”.

Now practise it

Neither of the two: probability through the complement

Trainer · Probability

Practise
Next: Probability as a share: how many cases fit
0/1

Was this article helpful?