I agree with scififan68 on this one.
Looking at the trace scififan68 provided shows the issue is more than likely not on the coax distribution side of their network (Field issue). The hop that spikes to almost 2 seconds appears to be referring tot he Tengig port on the CMTS chassis.
17 2427 ms 2180 ms 1516 ms tenge18-10-nycmnys-c401a.nyc.rr.com ----- 2 seconds.
18 272 ms 228 ms 195 ms cpe-24-XXX-XXX-XX.nyc.res.rr.com
The hop after that is your PC/CPE and since the there is no further layer 3 devices between the CMTS and you then it is more than likely an issue with that port based on the response times. It could be utilization at peak hours or a faulty connection but if it increases, then you would need to call them and explain to them where on your trace you are seeing the spikes. If you notice latency, run a trace just to verify since the trace done by scififan68 is him coming into the local network there and you want to prove that this is showing up on outbound traces as well.