The issue he has is that if the other bloke (and this is supposition on my part) had engaged in a long campaign of offensive stuff, then that doesn't fulfil the provocation defence for this incident in and of itself, because the reaction isn't immediate, looking at the footage.
Now he could obviously evince evidence of what the man may have said that suggests very bad shit was about to go down, and hey, that might cover the first punch. But, after that, the bloke is on the ground, the possible bad shit hasn't occurred, and isn't going to at that point, so the follow-up punches are pretty hard to describe as provoked by the strict legal defence.