Also TS3 wasn’t the first game to introduce fires, firefighters, burglars, cops, etc. And just a lot of things in general.

One is entirely exclusive of the other. You either want an intelligent neural simulation, which can’t be developed, until physicists can accurately image human neural activity during cognition, or a scripted interactive cut-scene, which will never convince a human.

Games such as Crysis don’t look real, because it’s illegal to realistically simulate violence. Governments will not allow you to publish a simulation, which is so realistic, that children become emotionally bound to their virtual characters, or so intelligent that ownership/treatment of the virtual entity must be legally evaluated.

Again, what you’re suggesting is a new intelligence algorithm for entertainment, and there is no interest. You can’t convince the human without an intelligent algorithm, or a statistical database with such depth, that it predicts every possible response from the human. That’s impossible.

“What does a competitor game need?”

There’s no objective answer, and it should obviously include as much of what consumers are demanding, as possible, or they won’t pay for it. Try looking at all of the compaints in The Sims 4 forums, and on twitter #stopEAabuse #EAListen, especially any regarding base content for The Sims 4 vs previous versions.

“How do we make a character in a videogame feel alive?”

This is where independent developers fail entirely, especially those who begin with the intention of telling the members of a competitor’s community that version one of their product will be published with FeatureBuffet™. If you want characters to animate and react in a human-like manner, then you’ll need to study human emotion and culture, capture and/or dynamically calculate motion for animation of models, and develop an engine which can efficiently run processes on distributed/multi-threaded architecture, without delay, and while synchronizing multi-actor interactions – The Sims are animated independently, and each sequence must fall within a fixed offset, range, and time slot. Dynamic animation is very consumptive and code-heavy.

Your work should begin with human behavioral studies, and requests for feedback regarding personal interests which are of most relevance to the features of your simulation, but you should also consider candidates who are educated in Drama/Stage Performance, Psychology, and Philosophy. It would be a mistake to blindly create a “stereotype simulator,” then spend infinitely on modeling and marketing, while consumers find nothing relatable within your content. If your only interest is accuracy of motion, then there are plenty of devices and APIs which can be used to capture and simulate human motion and/or newtonian physics, but players will want their characters to have personaities which relate to their own experiences. The key element is intellect, and you’ll eventually discover a cultural paradox, which mandates a choice between content depth and inclusivity.

A living character should accelerate within the range of norms for average physique (use motion capture), react to forces (hair, clothing, limbs, think ragdoll physics + nvidia hair), exhibit emotional and physical tendencies, store and react to memories, react to objects and events in a sane manner. Obviously, you should also consider the current state of entertainment industries when deciding on model resolutions and texture quality, since common experiences will affect the consumers’ perception of your work. PC gamers who are accustomed to the ‘ultra experience’ will never appreciate anything which looks and feels like it was written for android tablets, and I’m sure most simmers will all agree that The Sims 4 didn’t meet anyone’s expectations for model quality.