Speaker
Description
Hydro simulations are compared with recent experimental data by CMS, on a Principle Component Analysis. The trends for the scaled leading and sub-leading components for n=2,3 agree with data. In contrast, for n=0 there is a qualitative disagreement: the leading component increases with $p_T$ while it is constant in data. Using a toy model where the principal components can be computed analytically, it is shown that the n=0 leading and sub-leading components are extremely dependent on fluctuations in $N$ and $\bar{p_T}$ as well as their covariance and provide a new and clean way to study fluctuations and rule out initial condition models.