Hersi Maths WhatsApp me

Understand · explore · practise

Probability mass functions

Check probability tables, find normalising constants, handle piecewise formulas and calculate events from a discrete probability mass function.

Before you startDiscrete random variables, fractions and simple algebra.

01 / Check both requirements

A total of one is necessary, but not sufficient.

For every supported value x, P(X=x) ≥ 0. The sum over all supported values is 1.

These two conditions together also ensure every individual mass is at most 1.

A probability mass function, or PMF, assigns a probability to each possible numerical value. Always state the values for which the formula applies; outside them the mass is zero.

Is this a probability distribution?Explore

Each mass must be nonnegative and all masses together must sum to 1.

02 / Turn weights into probabilities

Divide each nonnegative weight by the total weight.

Watch: normalise unequal weights

Pause, replay or seek freely. The notes explain the same idea and stay in view.

Let P(X=x)=k(x+1) for x in {0,2,4}, and zero otherwise.Worked example

The weights are 1,3,5, with total 9

The probability total is k+3k+5k=9k.

9k=1, so k=1/9

Normalising uses every supported value once.

Masses: 1/9,1/3,5/9

They are nonnegative and sum to 1.

01 · Read the formula

For this PMF, find P(X=2) and P(X=3).

Hint

Use the formula only on its stated support.

Worked solution

P(X=2)=3/9=1/3. P(X=3)=0 because 3 is outside the support.

02 · An upper event

Find P(X≥2).

Hint

Include 2 and 4.

Worked solution

1/3+5/9=8/9.

03 / Missing values in a table

Account for every occurrence of a parameter.

03 · One missing mass

A table on x=0,1,2,3 has masses 0.15,0.25,a,0.30. Find a.

Hint

Subtract the known total from 1.

Worked solution

a=1−0.70=0.30. The resulting masses are all nonnegative.

04 · A repeated mass

A table on x=−1,0,2,5 has masses b,0.2,b,0.3. Find b.

Hint

There are two entries equal to b.

Worked solution

2b+0.5=1, giving b=0.25.

05 · Formula as weights

P(Y=y)=cy² on y=1,2,3. Find c and P(Y>1).

Hint

The weights are 1,4,9.

Worked solution

14c=1, so c=1/14. P(Y>1)=(4+9)/14=13/14.

04 / Use the correct branch

A piecewise PMF still describes one distribution.

P(Z=z)=kz for z=1 or4; P(Z=2)=2k; zero elsewhere.Worked example

Write the table before summing

At1 the mass is k, at 2 it is 2k, at 4 it is 4k.

k+2k+4k=1

So k=1/7.

P(Z<4)=1/7+2/7=3/7

The condition selects values 1 and 2 only.

06 · Inclusive endpoint

For Z above, find P(1<Z≤4).

Hint

Include 2 and 4 but not 1.

Worked solution

2/7+4/7=6/7.

07 · Missing support value

What is P(Z=3)?

Hint

Which branch includes 3?

Worked solution

0. A formula for selected values does not automatically extend to intermediate integers.

05 / A parameter can have a range

Normalisation may hold for many parameter values.

If masses are a,0.3,0.7−a, their sum is 1 for every a.

Nonnegativity then requires 0≤a≤0.7.

At an endpoint a mass can be zero. A table may list such a value, but it is then not in the positive-probability support.

08 · Test a proposed parameter

Is a=0.8 valid for those masses?

Hint

Check the final entry.

Worked solution

No. The final mass is −0.1. A total of 1 cannot repair a negative probability.

09 · Valid interval

Masses are 2t,0.4,0.6−2t. Find all valid t.

Hint

Check both expressions involving t are nonnegative.

Worked solution

t≥0 and t≤0.3, so 0≤t≤0.3. Their sum is already 1.

06 / Select supported values first

Strict and inclusive signs can change the answer.

Use this distribution for the next questions: values −3,0,2,6 have respective masses 0.1,0.2,0.4,0.3.

10 · Strict range

Find P(0<X<6).

Hint

Only 2 is strictly inside both bounds.

Worked solution

0.4.

11 · Include the endpoints

Find P(0≤X≤6).

Hint

Include 0,2,6.

Worked solution

0.2+0.4+0.3=0.9.

12 · Complement

Find P(X≠2).

Hint

Complement the mass at 2.

Worked solution

1−0.4=0.6.

07 / Reject an invalid model before calculating

Both tests protect the meaning of probability.

13 · Negative weights

Could choosing a normalising constant turn weights 1,−2,4 into probabilities?

Hint

Their total is 3, so the only normalising multiplier is 1/3.

Worked solution

No. Normalisation gives 1/3,−2/3,4/3, which are invalid. Nonnegative weights are required.

14 · Formula without a domain

Why is P(X=x)=x/10 incomplete as a distribution specification?

Hint

Which values of x are allowed?

Worked solution

The supported values must be given. For example, on {1,2,3,4} the masses sum to 1; on {1,2,3} they sum to 0.6. The same expression on different sets can have different validity.

08 / List, normalise, validate, select

Keep support and probability separate.

Make a table if a formula is hard to read. Sum all entries to find a constant, then check their signs. For an event, choose the supported values satisfying the condition and add their masses. Never extend a PMF formula beyond its stated domain.

Section 1 of 8 · Check both requirements