To be honest I don't know what was the motivation of TC39 for adding class syntax to ECMAScript. Perhaps for easier transition of object oriented people (Java/C++) into JavaScript? Any way, as you mentioned, delegation pattern (I prefer this name over prototypical inheritance) gives more flexibility. And I think we should use it, always.
Firstly, it has better performance against object-oriented approach. If you do not want to copy common set of properties to large number of objects, you just use Object.create(commonProperties) to create new object based on the object containing common set of properties. Within class approach it's not possible, you must create set of common properties for each object instance, which might have greater memory consumption impact.
What's more, class inheritance by design leads to tight coupling. Imagine you want to model tesla car in your application. Probably first thought is to define ElectricCar which extends Car class. Then we need some gas-based car, let's say ford. So you define GasCar which also extends Car class. But what happen, when we want another car, let's say steem, which have both electric car and gas-based car capabilities? Which class should we instantiate, both of them? Object composition is more flexible: we might just call const steem = Object.extend({}, electricCarCapabilities, gasCarCapabilities) where objects have functions and properties describing specific set of capabilities.
JavaScript is so powerful tool, it does not need object-oriented paradigm to be part of the specification. I think functional programming is more fun. Strongly encourage you to make some shift in way of thinking and check it yourself. I did so.
Nice article by the way!