Markov's inequality bounds the probability that a nonnegative random
variable
reaches or exceeds a positive threshold. For every real
number
,
|
(1)
|
Here
denotes the expectation value of
. The bound also holds when
, although it then gives no restriction
on the probability. For the indicator
function
of the event
, the pointwise inequality
gives
. Dividing by
proves the bound for arbitrary probability
distributions (Durrett 2019, pp. 29-30).
When
has a probability density function
, the theorem also has an integral proof. Write
|
(2)
| |||
|
(3)
|
Since the probability density function and
are nonnegative,
|
(4)
| |||
|
(5)
| |||
|
(6)
| |||
|
(7)
| |||
|
(8)
|