The DiffServ Code Point (DSCP) field in the IP header allows to specify adesired per-hop behavior as packets traverse routers. Setting the DSCPfield opportunistically, without prior contractual agreement, has recentlybecome accepted practice for Internet end hosts. Measurement studies findthat there is reason to hope for a DSCP setting to have an effect ontraffic, and at least configuring this value is not heavily detrimental:systematic drops of packets due to non-zero DSCP values are rare, and thevalue is often left intact along an end-to-end path. What these studies donot discuss is whether per-hop behaviors truly are honored: what happens topackets in terms of the delay they experience? In this paper, we make anattempt to find a first answer to this by mining a dataset of our ownrecent large-scale measurement study. Using a deep neural network, weobtain the importance of the factors which help us understand the delayimpact of the DSCP.