The last two decades of network science have discovered stunning similarities in the
topological characteristics of real life networks (many biological, social, transportation
and organizational networks) on a strong empirical basis. However our knowledge about
the operational paths used in these networks is very limited, which prohibits the
proper understanding of the principles of their functioning. Today, the most widely
adopted hypothesis about the structure of the operational paths is the shortest path
assumption. Here we present a striking result that the paths in various networks are
significantly stretched compared to their shortest counterparts. Stretch distributions
are also found to be extremely similar. This phenomenon is empirically confirmed on
four networks from diverse areas of life. We also identify the high-level path selection
rules nature seems to use when picking its paths.
