If my proposed method to test multi-threading/cpu usage in games is not right, then what would you need it for in games at all. In games all you need multi-threading for is more objects in the scene, better lighting, more physcis-enable objects etc., so more going on screen.
The foliage tool test case is a good example. You want grass as far away from the player as possible and not cull it to early to avoid naked polygons being visible. It is the classic example for performance/fps critical elements in a level and you do not want to have cpu idling if fps drops because of it. There are many more examples.
Essentially, what I want to bring across is, that when you ship a game done in UE4, and you specifiy the maximum requirements, then you can write:
maximum cpu: dual core
Any money spend on more cores is wasted.