# Probability of a sum of uniform random variables [closed]

Let $X_1,X_2,X_3,X_4$ and $X_5$ be uniformly distributed random variables over $(0,1)$.

What is the probability that $X_1+X_2+X_3+X_4+X_5 < 1$? I would like to know the shortest way to do this.

## closed as off-topic by dustin, user99914, Dario, Did, Claude LeiboviciFeb 6 '15 at 7:52

This question appears to be off-topic. The users who voted to close gave this specific reason:

• "This question is missing context or other details: Please improve the question by providing additional context, which ideally includes your thoughts on the problem and any attempts you have made to solve it. This information helps others identify where you have difficulties and helps them write answers appropriate to your experience level." – dustin, Dario, Did, Claude Leibovici
If this question can be reworded to fit the rules in the help center, please edit the question.

• See: math.stackexchange.com/questions/1128396/… for a geometric representation ... or check out the Irwin-Hall distribution at the link provided – wolfies Feb 6 '15 at 7:18
• The wiki page for Irwin Hall distribution provides the probability distribution function for sum of up to 5 variables. However it doesn't provide the method for derivation. – Achal Premi Feb 10 '15 at 9:30

Hint: Can you describe the region $\{(x_1,x_2,x_3,x_4,x_5) \in [0,1]^5 : x_1+x_2+x_3+x_4+x_5<1\}$ as a geometric object? What is this object's volume? Try with $5$ replaced by $2$ or $3$ first if you don't see it.

Hint 2: it looks like you don't yet have intuition about the relationship of expectations and integrals (that okay! you have to start somewhere). Consider that if a random variable $X$ taking values in $\mathbb{R}^d$ has density $f$, then $$P( X \in A ) = \int_A f(x)dx$$ for nice sets $A$. Note that the set $\{X_1+\cdots + X_5 < 1\}$ is the set where the random variable $X:=(X_1,\ldots,X_5)$ is in the set $A:=\{(x_1,x_2,x_3,x_4,x_5) \in [0,1]^5 : x_1+x_2+x_3+x_4+x_5<1\}$. Take for granted that this set is nice enough. Assuming the $X_i$ are independent, we know the density of $X$ is $f(x) = 1_{x_1,\ldots,x_5 \in [0,1]}$. Check that integrating this over $\mathbb{R}^n$ gives $1$. In particular $$P(X_1+\cdots+X_5 < 1) = P(X \in A) = \int_{A}1_{x_1,\ldots,x_5 \in [0,1]} 1dx = \int_{A} 1 dx.$$ Now, we know that the integral of $1$ over a region is the volume of that region. Thus to calculate the integral, we can try to geometrically recognize what kind of shape $A$ is and just calculate it's volume.

Lets looks at the shape in a few different dimensions, replacing $5$ here with $n$.

$n=1$: $[0,1)$ thats an interval with length $1$.

$n=2$: $\{(x,y) \in [0,1]^2: x+y<1\}$ is a triangle with area $1/2$.

$n=3$: $\{(x,y,z) \in [0,1]^2: x+y+z<1\}$ is a tetrahedron with volume $1/(2\cdot 3)$.

In general the shape is a simplex with volume $1/n!$. To calculate this we can just do the integral $$\int_0^1\int_0^{1-x_1}\int_0^{1-x_1-x_2}\cdots \int_0^{1-x_1-x_2\ldots-x_{n-1}} 1 dx_n\cdots dx_3 dx_2 dx_1 = \frac{1}{n!}.$$

• @Eupraxis1981 It's the standard textbook approach for multivaritate uniform distributions in most textbooks e.g. Pitman's Probability. – Fizz Feb 6 '15 at 10:00
• @RespawnedFluff Yes, and....? I was simply commenting on the poster's speed and for not resorting to convolutions. – user76844 Feb 6 '15 at 12:11
• I was able to derive for x_1 + x_2 + x_3 < 1. By following: let X_3 = u. So P(x_1+x_2 < 1-u)= 0.5*(1-u)^2 P(x_1 + x_2 + x_3 < 1) = Integration u=0 to u=1 [(1-u)^2 du]. Is the same approach required to calculate P(x_1 + x_2 + x_3 + x_4 < 1) ? – Achal Premi Feb 10 '15 at 9:37