Not necessarily. There's like 4000 people who voted on Diablo 3 and it averaged around a 3. That's totally unfair. Diablo 3 is a very good game.
Not really, and certainly not when it was released. Compared to D2 it was unspectacular, uninspired, limited and built around the auction house. Covering up the database breaches early on, the massive exploiting and the server issues did not help. If you force people to play online, remove LAN features, and then fail to provide adequate server capacities and performance, then the user feedback will reflect this.
I had waited for D3 for ten years, and I played D2 for several years, as well. My interest in D3 died in the third week. I never played a game that annoyed and frustrated me more than D3 did with its Inferno mode. Some of the worst design I have seen in the genre. Now, I would not give it a 3/10 either, because if you just play through the story once or twice, you get 20-40 hours of good entertainment from it, and I do not regret the purchase, but gameplay-wise it is a far worse title than D2 and definitely a league below the excellent Torchlight 2, which had a fraction of the budget. D3 is an "okay" game. Cutscenes and voice acting were great, and I liked the painterly style, too, but loot, character and monster/encounter design, as well as handling of act travelling and poor, barely random map generation were all huge disappointments to me, as a genre fan.
But I agree that user ratings are pretty useless. Actually, I think the professional metascore is also meaningless after a game has been out for a few weeks and development continued. Static scores meant something 20, 25 years ago. Today, they are often counter-productive since they do not reward continued support.