From an expert investigation - maximum energy at a plug-in protector was 35 joules, and almost all cases under 1 joule, with a power line surge including those caused by a 100,000A lightning strike to the primary wire at an adjacent utility pole. The maximum energy wasn't even from the worst surge.
When the voltage at service panel busbars reaches about 6,000V there is arc-over to the enclosure. The voltage of the established arc is hundreds of volts. Since the enclosure is connected to the earthing system that dumps most of the surge energy to earth. Since the "ground" and neutral are also connected to the enclosure, the exposure beyond the panel if far less than imagined.
A strong surge will drive the voltage on the busbars to 6,000V and arc-over. With a weaker surge, a plug-in protector on a short branch circuit may keep the service panel voltage below arc-over. The 35 joule energy in the investigation was with one of those weaker surges.
I don't suggest that people use plug-in protectors. I suggest they make decisions based on science. Discussion centers on plug-in protectors because of the misinformation posted by westom (and other misinformation that has been posted).