Context
I was talking with a friend about this simple problem and we couldn't figure it out easily. I spent a little while with paper and pen and googled a bit and checked previous questions but have not been successful finding a solution. I think I am missing some language/terms.
Problem
Let $D_i$ be the $i$th outcome of a 20-sided die event. That is, $D_i\in[1,20]$. What is the probability that $D_1<D_2<D_3$ for three sequential die events?
Scratch Work
My pen & paper work thus far (I feel like I am abusing notation):
$P(D_1)=1$. $P(D_2>D_1)=\frac{20-D_1}{20}$. $P(D_3>D_2)=\frac{20-D_2}{20}$.
$P(D_1<D_2<D_3)=P(D_1)P(D_2>D_1)P(D_3>D_2)=(1)(\frac{20-D_1}{20})(\frac{20-D_2}{20})=1-\frac{D_1+D_2}{20}+\frac{D_1D_2}{400}$.
$D_1+D_2\in[2,40]$; $D_1D_2\stackrel{?}{\in}[1,400]$.
Closing
I know that I am doing something gravely wrong and I want to know what it is. All help appreciated, especially generalizations.
Hint: Partition over the middle value: $$\begin{align}\mathsf P(D_1<D_2<D_3)&=\sum_{t=2}^{19}\mathsf P(D_1<t)\mathsf P(D_2=t)\mathsf P(D_3>t)\\[1ex]&=\sum_{t=2}^{19}\dfrac{(t-1)(20-t)}{20^3}\end{align}$$