[bo1024]

Hang on - the one time correlation does strongly suggest causation is when trials include randomized interventions, which i pretty much the case here (once in a while, move my mouse and see if network traffic spikes).