It's really difficult to program all those animations -- and put in the hundreds of hours of play-testing in order to make sure there aren't that many bugs (because there will always be bugs -- it's impossible to get rid of them all) by the time you launch the game.
That's why action games are becoming rarer. They're getting too expensive (and time-consuming) to make due to all the new technology in the marketplace nowadays.
If we scale down the graphics back to PS2 levels, then we'd see a lot more releases. Unfortunately, most gamers these days are graphics-whores and don't want anything less than AAA budgets.
So, two things. First: most of what you said is wildly inaccurate, and this will be a lengthy post about why almost all of it is wrong. I know you probably won't read it because you and your friend have lately been too scared to actually talk to me about anything, but this is for everyone else's information. Second: the remaining elements of truth are statements of why games in general are hard, and not why another devil may cry would be hard.
The entire purpose of long established in-house engines (like the MT framework) is to dramatically reduce the number of hours that need to go into pre-production, engine development, and bug fixing. The amount of gameplay programming and testing that remains is no more than any other game would require; making another devil may cry on the MT framework, unreal, unity, or any other established engine would actually be much easier than making any other game from scratch. That's kind of just common sense.
It's very rare for anyone to actually program an animation, the existing engines companies like capcom use have implemented loading a model binary and skeletal animation long ago. Furthermore, converting maya binaries and skeletal animation are pretty much solved problems, so the need to continually update your animation methods isn't really there. When someone invents something better than a simple scene graph or matrix stack, that's actually worth the overhead to tear out all existing animation code and rewrite it, maybe then we can talk about the "huge costs" a company like capcom would face here.
Oh, also motion capture. Pretty much every AAA studio has their own in-house setup for this, and even smaller studios have places they can go for access to this technology. Ever wonder why modern games aren't so big on non-realistic enemies any more? It's because rigging and animating a realistic looking dude with 4 arms is a hell of a lot harder than copy-pasting your tried and true 2 armed rig, and animating it with motion capture. Hell, even motion capturing a dog is easy compared to something that would require hand animating, like a centaur. There are exceptions to this; Riot are basically psychopaths and do pretty much all of their animations by hand, which sort of blows my mind. It's not even that they don't have access to the technology to animate their mostly-human cast of characters, it's just that they don't want to. As you could probably imagine, their artist
rogrammer ratio is even more skewed than normal.
Would reducing graphics quality to ps2 levels really help? Well first of all, why on earth would you do that? Most of the cool effects people associate with modern games are not hundreds of hours of manpower to implement. What usually happens is someone finds a cool trick, or neat application for a long-known method, and they just slap it in there. Even tiny indie projects made with just one or two people can still find the time to implement simple things like phong shading, motion blur, or 2D shadows (which is actually considerably harder than 3D shadows, if you can believe that). I don't think bayonetta 2 would have released a day earlier than it did if someone had decided to exclude reflection, noise, or projected texturing, (which is essentially how they achieved their cool water effects).
TL;DR - Games are hard, but the things you think would be time-savers actually aren't. The only real barriers that stand in the way of another devil may cry game are the same barriers that any other game would face. In fact, you could convincingly argue that devil may cry would be easier than most games, as it's an existing formula, modified and applied on an existing engine where that formula has been shown to work in the past.
