Uniform and Exponential Problems with Boundary Checks

Why Boundary Checks Matter

Uniform and exponential models describe different kinds of uncertainty. A uniform model spreads probability evenly over a bounded interval; an exponential model describes a nonnegative waiting time with a constant rate. Problems that use either distribution can look straightforward until an event extends beyond the model’s support, a rate is measured in the wrong time unit, or an endpoint is treated as if it had positive probability.

A reliable calculation begins by identifying the distribution’s support and writing the event in the units of the model. Only then should you apply a density or distribution function. At the end, check that the answer is between zero and one and that complementary events fit together as expected. These checks are especially useful when a problem combines a uniform quantity and an exponential waiting time.

Definition: A continuous random variable $X$ is uniform on $[a,b]$, where $a<b$, if its density is $f_X(x)=1/(b-a)$ for $a\le x\le b$ and zero otherwise. Its support is $[a,b]$. A waiting time $T$ has an exponential distribution with rate $\lambda>0$ if its density is $f_T(t)=\lambda e^{-\lambda t}$ for $t\ge 0$ and zero otherwise. Its support is $[0,\infty)$, and $\lambda$ has units of inverse time.

For $X\sim\operatorname{Uniform}(a,b)$, the probability of an event is the length of the event’s intersection with $[a,b]$, divided by $b-a$. For $T\sim\operatorname{Exponential}(\lambda)$, the distribution and survival functions are $P(T\le t)=1-e^{-\lambda t}$ and $P(T>t)=e^{-\lambda t}$ for $t\ge 0$. If $t<0$, then $P(T\le t)=0$ and $P(T>t)=1$, because negative waiting times are outside the support.

Boundary rule: For any continuous random variable $X$, $P(X=c)=0$ for every fixed number $c$. Consequently, including or excluding a single endpoint does not change an interval probability. This rule does not mean that the endpoint is outside the support; it means that one point has zero probability under a continuous distribution.

Start With the Support

The support gives the values a model can take. For a uniform variable, an event outside $[a,b]$ contributes no probability, and an event only partly inside the interval must be clipped to the support before its length is calculated. For an exponential waiting time, negative thresholds must be handled using the nonnegative support rather than substituted carelessly into a formula derived for $t\ge 0$.

The endpoint convention can be handled simply: for continuous models, a probability over $[r,s]$, $(r,s]$, $[r,s)$, or $(r,s)$ is the same whenever the endpoints are fixed. Nevertheless, writing the event precisely helps reveal whether it lies within the support and whether it is empty. In particular, a continuous variable can have support including $c$ while still satisfying $P(X=c)=0$.

Worked Example: Clipping a Uniform Event to Its Support

A hypothetical inspection time $X$, measured in minutes, is modeled as uniform on $[10,40]$. Find $P(5\le X<25)$, $P(X\le 10)$, and $P(X\ge 30)$. Use the support check to verify that the three calculations are consistent with the model.

Method and model. Here $X\sim\operatorname{Uniform}(10,40)$, so the full support has length $40-10=30$ minutes and the density on that support is $1/30$ per minute. For each event, intersect its range with $[10,40]$ before dividing by 30.

For $5\le X<25$, the portion of the event that lies in the support is $[10,25)$. Its length is $25-10=15$, so

P(5\le X<25)=\frac{25-10}{40-10}=\frac{15}{30}=0.5000.

The event $X\le10$ contains only the support endpoint $X=10$, since values below 10 are impossible. A single point has probability zero, giving $P(X\le10)=0.0000$. For $X\ge30$, the part inside the support is $[30,40]$, of length $40-30=10$, so

P(X\ge30)=\frac{40-30}{40-10}=\frac{10}{30}=\frac{1}{3}=0.3333.

As a check, the events $X<25$ and $X\ge25$ partition the support, and the second probability is $15/30=0.5000$, the complement of the first. The probability $P(X\le10)=0$ is also consistent with the fact that only one endpoint is included. The common mistake is to use the untrimmed length $25-5=20$ for the first event, obtaining $20/30=0.6667$. That incorrectly assigns probability to values from 5 to 10, which are outside the model’s support.

Exponential Events: Check the Time Unit

In an exponential model, the numerical value of the rate depends on the unit used to measure time. If $\lambda=0.4$ per hour, the exponent $\lambda t$ is dimensionless only when $t$ is expressed in hours. A threshold given in minutes must first be converted to hours, or the rate must be converted to a per-minute rate.

For $0\le r<s$, subtracting two tail probabilities gives the probability of falling between the thresholds:

P(r<T\le s)=P(T>r)-P(T>s)=e^{-\lambda r}-e^{-\lambda s}.

The formula assumes that $r$ and $s$ are nonnegative and are expressed in the time unit corresponding to $\lambda$. If a threshold is negative, use the support directly: no waiting time is negative. Probability calculations should also respect bounds: every probability lies between zero and one, and the probabilities of a partition into disjoint events sum to one.

Worked Example: Converting Minutes Before Using an Exponential Rate

A hypothetical equipment alert occurs after a waiting time $T\sim\operatorname{Exponential}(0.4\text{ per hour})$. Find the probabilities that the alert occurs within 45 minutes, between 45 minutes and 2.5 hours, and after 2.5 hours. Also find the probability it occurs exactly 45 minutes after monitoring begins.

Method and model. Convert every threshold to hours before using the rate $0.4$ per hour. In particular, 45 minutes is $45/60=0.75$ hours. For nonnegative $t$, use $P(T\le t)=1-e^{-0.4t}$ and $P(T>t)=e^{-0.4t}$.

Within 45 minutes,

P(T\le0.75)=1-e^{-0.4(0.75)} =1-e^{-0.3} =1-0.7408182 =0.2592.

Between 45 minutes and 2.5 hours,

\begin{aligned} P(0.75<T\le2.5) &=e^{-0.4(0.75)}-e^{-0.4(2.5)}\\ &=e^{-0.3}-e^{-1}\\ &=0.7408182-0.3678794\\ &=0.3729. \end{aligned}

After 2.5 hours,

P(T>2.5)=e^{-0.4(2.5)}=e^{-1}=0.3679.

Because $T$ is continuous, $P(T=0.75)=0.0000$. For a boundary check, the three time ranges before, between, and after the thresholds are disjoint and cover all possible waiting times. Before rounding, their probabilities add to $0.2591818+0.3729388+0.3678794=1.0000000$. The common mistake is to substitute 45 directly into $e^{-0.4t}$, giving an exponent based on 45 hours rather than 45 minutes. Converting $45$ minutes to $0.75$ hours prevents that unit error.

Combining Independent Uniform and Exponential Variables

When a problem compares a uniform quantity $U$ with an exponential waiting time $T$, state whether the variables are independent before calculating. If they are independent, the conditional probability of $T>U$, given $U=u$, is the exponential survival probability $e^{-\lambda u}$. Averaging that conditional probability over the possible values of $U$ gives the probability of the comparison event.

Result: If $U\sim\operatorname{Uniform}(a,b)$ and $T\sim\operatorname{Exponential}(\lambda)$ are independent, with $0\le a<b$ and compatible time units, then $$ P(T>U)=\frac{1}{b-a}\int_a^b e^{-\lambda u}\,du =\frac{e^{-\lambda a}-e^{-\lambda b}}{\lambda(b-a)}. $$

The integral has a direct justification. Conditional on $U=u$, the event $T>U$ becomes $T>u$, which has probability $e^{-\lambda u}$. The density of $U$ is $1/(b-a)$ on $[a,b]$. Integrating the conditional probability against that density yields the joint event probability:

\begin{aligned} P(T>U) &=\int_a^b P(T>U\mid U=u)f_U(u)\,du\\ &=\int_a^b e^{-\lambda u}\frac{1}{b-a}\,du\\ &=\frac{1}{b-a}\left[-\frac{1}{\lambda}e^{-\lambda u}\right]_a^b\\ &=\frac{e^{-\lambda a}-e^{-\lambda b}}{\lambda(b-a)}. \end{aligned}

The assumptions matter: the interval for $U$ must be nonnegative for this waiting-time interpretation, $U$ and $T$ must be independent, and the rate and interval must use the same time unit. Since both variables are continuous and independent, $P(T=U)=0$; consequently $P(T\le U)=1-P(T>U)$.

Worked Example: Comparing an Independent Uniform Duration and Waiting Time

A hypothetical review duration $U$ is uniform from 1 to 5 hours. Independently, a response waiting time $T$ is exponential with rate $0.5$ per hour. Find $P(T>U)$, $P(T\le U)$, and the probability that the review duration is at least 3 hours and the response takes longer than the review.

Method and model. Use conditioning on $U$, because the event $T>U$ has an exponential survival probability once the value of $U$ is fixed. All durations are in hours and the rate is per hour. The uniform density is $1/(5-1)=1/4$.

First,

\begin{aligned} P(T>U) &=\frac{1}{4}\int_1^5 e^{-0.5u}\,du\\ &=\frac{1}{4}\left[-2e^{-0.5u}\right]_1^5\\ &=\frac{1}{2}\left(e^{-0.5}-e^{-2.5}\right)\\ &=\frac{1}{2}(0.6065307-0.0820850)\\ &=0.2622. \end{aligned}

Because equality has probability zero, $P(T\le U)=1-P(T>U)=1-0.2622228=0.7378$.

For the joint event $U\ge3$ and $T>U$, integrate only over the uniform values from 3 to 5:

\begin{aligned} P(U\ge3,\ T>U) &=\frac{1}{4}\int_3^5e^{-0.5u}\,du\\ &=\frac{1}{2}\left(e^{-1.5}-e^{-2.5}\right)\\ &=\frac{1}{2}(0.2231302-0.0820850)\\ &=0.0705. \end{aligned}

As a check, $P(U\ge3)=2/4=0.5$, so the joint probability $0.0705$ is no greater than the probability of the condition $U\ge3$. Also, $P(T>U)=0.2622$ lies between zero and one, and its complement is $0.7378$. The common mistake is to use $P(T>3)$ for the joint event: that ignores the fact that the threshold is the random duration $U$, not the fixed value 3. The integral over $3\le u\le5$ accounts for the full event.

A Practical Boundary-Check Sequence

A short sequence of checks catches many otherwise plausible errors. Identify the support first; express every threshold in compatible units; intersect the event with the support; and only then use the relevant probability formula. For a mixed event, identify the conditioning variable and confirm independence when the calculation requires it. Finally, check point probabilities, complements, and the range of the answer.

Write down the model and units.
Record the support, parameters, and units. For an exponential rate, write the time unit explicitly.

Check the event against the support.
Remove impossible portions and recognize events containing only a single point.

Calculate and verify.
Use interval lengths for uniform probabilities, survival or distribution functions for exponential probabilities, and conditional integration for independent mixed comparisons. Check the result against zero, one, and any relevant complement.

A frequent exam trap is to treat a boundary as a separate probability-bearing outcome. For continuous distributions, changing $<$ to $\le$ at a fixed endpoint does not change the probability. A different trap is to apply an exponential formula at a negative time without first checking the support. The formula $e^{-\lambda t}$ is the survival probability for $t\ge0$; for a negative threshold, the waiting time is always greater than that threshold, so its probability is one.

Key takeaway: A probability formula is only as reliable as the support and units supplied to it. Clip uniform events to their bounded interval, keep exponential thresholds nonnegative and in the rate’s time unit, and assign zero probability to a single point under either continuous model.

Check Your Understanding

For each problem, identify the support and units before calculating. Round probabilities to four decimal places.

Let $X\sim\operatorname{Uniform}(-3,9)$. Find $P(X\le0\text{ or }X>6)$ and $P(X=9)$. (Answer: $0.5000$ and $0.0000$.)
A waiting time $T$ is exponential with rate $0.2$ per hour. Find $P(2<T\le7)$ and $P(T=2)$. (Answer: $0.4237$ and $0.0000$.)
Let $X\sim\operatorname{Uniform}(4,10)$, measured in days. Find $P(0\le X<6)$. (Answer: $0.3333$.)
Independent variables $U\sim\operatorname{Uniform}(0,4)$ hours and $T\sim\operatorname{Exponential}(0.5\text{ per hour})$ describe a hypothetical duration and waiting time. Find $P(T>U)$. (Answer: $0.4323$.)
Independent variables $U\sim\operatorname{Uniform}(2,6)$ minutes and $T\sim\operatorname{Exponential}(0.25\text{ per minute})$ describe a hypothetical duration and waiting time. Find $P(T>U\mid U\ge4)$. (Answer: $0.2895$.)

What You'll Learn