Wow. And, "is regressing" ?? It's done, for a long time now. And it's all the time.
But on topic, I'm not sure I understand your comment as it seems to be self conflicting. At first you say,
"Divinity has a much more even performance profile, meaning that the devs of that game have actually done their work without any over bias to any vendor."
Then you say,
"A poorly optimised/one-sided game is not taken as a proof to bash AMD."
Just wondering what Divinity is to you. A game with a much more even performance profile that the devs knew what they were doing? Or a poorly optimised/one-sided game? I don't see how it can be both. Unless I'm completely missing the context or you misspoke...?
And is your observation of obduction's performance solely based on the variances between Nvidia and AMD scores? How do you know that Obduction was coded poorly and Divinity was coded in a well balanced manner? How do you know, off the cuff, that Obduction doesn't just favor one architecture over another, and Divinity isn't as architecture dependent?
Unless you thoroughly researched this?