This is a pretty good video that examines why people might have liked Skyrim. He's pretty correct, but his praise of the stealth system is a bit odd. It's neat that it tracks so many things for detection, but alert states need to be longer and enemies need to search more.
There two distinct kinds of gameplay systems and those aren't "broken" and "not broken".
They are those that are fundamentally soundly designed and those that are fundamentally broken/misdesigned.
The former can still be fixed easily if they are broken. The later can merely have their badness covered up and pretend they are good.
As it happened, most of the stuff in Skyrim, while broken, belongs in the first category, making Skyrim a solid foundation for modding that yields to having it's various shortcomings easily fixed by modding.
Stealth, for example is solid at core but broken (I'd guess purposefully, for the benefit of consoletard audiences) to make AI oblivious and demented. But since the core is actually good, just tweaking a few variables fixes that.
Skyrim in general seems to have a much more ambitious (unfinished) game inside wanting to get out, and, unlike Oblivion, the existence of this game can be actually seen in game data, rather than being purely based on devs' claims.
Tweak a few variables - bam, good stealth.
Tweak a few more - actually sensible AI.
Unearth some cut stuff - dynamic civil war campaign emerges.
Up the casting costs - the College of Winterhold actually expects you to be a capable caster and demonstrate it before letting you in.
And so on.