I see various mentions of differences between the two modes but I can't find anything definitive.
For instance I see mention that the Video Mode is not pixel accurate and Video Pattern Mode is, but I don't see an explanation of why. They are both reading the same video signal after all.
I saw somewhere that Video Mode uses built in patterns whereas in Video Pattern Mode the patterns have to be specified explicitly. If I'm not trying to do anything clever with the patterns why incur the extra work of setting up the pattern?