Red teaming is campaign-based testing that runs for an extended period of time, e.g., multiple weeks or months of emulating the same attacker.
With enough time and resources, all security defences will fall. Red teaming is not meant to score the skill of defenders – but the cost to adversaries. Blue teaming is meant to increase the cost to adversaries. Purple teaming integrates the defensive tactics and controls from blue teams with the threats and vulnerabilities found by red teams into a single narrative that maximizes both while minimising the inconvenience for operations adding these defensive tactics and controls.
I agree with Miessler I don’t much care for the word “team” being assigned to the colours, as I too think in most cases they’re actually mindsets, or functions, rather than dedicated groups of people. We add and change: