Eavesdropping is an important threat in the context of mobile ad-hoc networks due to the use of open air as the transmission medium. As a consequence, some works trying to prevent this threat have proposed the use of anonymous routing protocols. In general, these works do not include measurements to assess how easy it is for an attacker to deduce traffic patterns from observed traffic. In this paper, we propose several metrics to characterize the degree of traffic exposure. We apply these techniques to ANODR, one of the most popular anonymous routing protocols for MANETs. The results show that, independently of the traffic pattern, it is relatively easy to deduce what the sources are, especially those with the highest transmission rates. In contrast, we show that inferences relative to recipients are minimal, meaning that recipient anonymity is easier to safeguard.