In games it's a matter of level of detail though (either for the CPU/memory or simply workload for the animators). A body skeleton typically consists of around 30 bones. If two hands with 5 fingers get 3 bones extra that doubles the load for relatively little effect. For a single player character it's still worth it. For NPCs it typically does not make the cut.
In animation each bone typically has 6 degrees of freedom so it is no problem to copy any hand movement. Doing the same with servos in robots at a natural speed and in a coordinated manner is a lot harder.

Does controlling a mechanical hand count? Or communicating electronically (however primitive)? And that's what he was up to in 2002.

What the article fails to mention is the underlying reason for this: resale. If a gamer finishes the game it is done, a coaster in a pretty box. If the game always has something left to do, whether in the form of downloadable content, achievements, replayability or open endedness, it will retain some value and not end up traded in for a new game quite so soon. The game resale market may seem pretty small (mostly because stores take a huge second profit margin on them), but add to that the number of copies lended to a friend or rented for the weekend. In the end significantly more people will buy their own box if it provides limitless enjoyment.
In my opinion adding more value to a game is the most customer friendly way to do it. Far better than strong arming stores to not take trade ins or locking installations to hardware, creditcard and so on.

If by 'the same kind of violence' you mean 'a different kind of violence'.
The NY times article refers to the ESRB rating. I'm pretty sure the article with the alternatives went by those. In your example the alternative, Overlord II, is rated Teen while its counterpart, Left 4 Dead 2, is rated Mature.
There are standards for these ratings. Now you may disagree with the standards, but dismemberment, animated blood and gore fall in the M category. Morality choices, like playing on the side of evil in Overlord, are not totally excluded from the standard, but usually have less impact.

I'm sorry if I ignored a crucial difference in comparing the two. I'm not from the US. In my country (the Netherlands) there is not such a big gap in performance between gym class and sports outside of school hours.
HCM can apparently cause heart failure even under light stress, even just emotional stress. But then I'm not a medical professional either so perhaps I am again assuming too much about the motivation of the grand parent poster.

Full article here
The catch is that the "otherwise healthy" claim is arbitrary here. The most common cause, according to that article, is HCM. The problem is that this is hard to test for (many false positives) and can occur in athletes, even on the professional level.
"Seemingly healthy" would have been a better choice of words.

Roger Penrose makes a hard to refute argument for pretty much that statement. As it's based on completeness theory and turing machines it's pretty mathematical.
The Chinese Room argument of John Searle makes a similar claim, although it's less mathematical.
I would say these proofs are support too strong a conclusion by comparing the brain to a single algorithm. They merely raise the question what the magic ingredient is. Massive parallelism is my favourite candidate and seeing as how that happens to be the direction that computation is progressing anyway I confidently voted AI.

A car drives from A to B. At its top speed this takes one hour. The driver decides to get out after 40 minutes without pulling over. How much time does it take him to arrive at his destination?