For example for $n=2$ the new phase $y_0$ is the average of $y_1$ and $y_2$ but this dowsn't seem to be true in general.
And what about simplifying $\sum_{i=1}^{n} A_i sin(x - y_i)$ ?
Analytically it's easy to see why we get one sine. But is there an elementary reasoning for this?
$$ \sum_j^n A_j\sin(x-y_j)={\bf Im}\sum_j^n A_je^{ix-iy_j}={\bf Im}\left(\sum_j^n A_je^{-iy_j}\right)e^{ix}={\bf Im}\,re^{ix-i\vartheta}=r\sin(x-\vartheta), $$ where $$ \sum_j^n A_je^{-iy_j}=re^{i\vartheta} $$