Yes, ping times are affected by which hops they flow through. We always wanted a more 'real world' result for pings so wanted those to go through many switches getting to our network. We just needed a reliable destination. It didn't matter where or what, so long as the agents algorithm was able to build up averages from which it would recognize latency changes.
We've tested using closer (to the agent) destinations but it's never 'real world' like. We keep testing this to see how we can improve it.
Thanks very much for bringing this up. We're looking into the other things too.