What to develop has been decided, but I plan to refine the plan a bit more...
Some thoughts...
The thing I like the most about T:WGM is the simulation aspects. The thing I like the least is the reliance on FMVs for H-scenes. These two elements clash, in that the simulation aspects lends itself to variation, while FMVs are totally static. Even when brute-forcing a ton of different FMVs, as T:WGM does, you still get a bunch of incongruous elements and very little interactability.
I would advise caution when it comes to Ren'Py for a few reasons:
-If you're looking at making a Ren'Py game as a vehicle for learning or improving your proficiency with Python, which is a thought I've had myself, be aware that Ren'Py uses a highly-customized version of Python 2. It's almost totally irrelevant for learning modern Python 3.
-Recently, experimental support for Python 3 has bee added, but there's a total lack of learning resources for making Ren'Py games in Python 3.
-When Ren'Py 8 launches, Python 3 is supposed to become the new default, but there's no concrete eta for the launch. At one point, I think Ren'Py Tom was hoping for 2021. I don't think that's the case anymore. When Ren'Py 8 does come out, I may try to make a game with it myself.
It's not impossible to do a simulation-heavy game in Python/Ren'PY, and there's
at least one game that does so. Most that try end up as buggy messes. C# and Unity are generally better fits for simulation and any sort of unique gameplay. The closer to a straight up VN you are though, the more of an advantage Ren'Py has.
-Do you still intend to use FMVs for h-scenes? If so, have you compared compression/format support/file sizes between Ren'Py and Unity?
-If you do still intend to use FMVs, I'd suggest looking at creating video files with alpha channels, and then compositing multiple videos together to make a scene. This is doable in Unity. Is it doable in Ren'Py?
-That said, I would really prefer to see these scenes 3D rendered in real time. Watching the same FMVs over and over without any interactability doesn't do much to get the blood pumping.
As a general thought, I'll say that T:WGM is one of my all time favorite H-games, even though the FMVs don't do much for me. My all time favorite H-game is
MLA. Both were moderately successful, but didn't become breakout hits. Both do a few things really well, but also have obvious weaknesses. The author of MLA went in another direction afterwards and well, the results have been somewhat mediocre. This isn't to say that if you go in another direction with the next game that the same thing will happen to you. Maybe you'll stumble upon something great... but the fact is that you already have something great but minus a few flaws that are holding it back. I would advise that for the next game you iterate on what you have, bolster the strengths and address the weaknesses. I think that if you do this, the next game will be much more successful than T:WGM. To the extent you veer, you're rolling the dice.
That last thing I'll advise is that, whatever direction you do go, if at all possible, work on the tech before you start working on the game. The weakest part of T:WGM is the reliance on FMVs for h-scenes. If the H-scenes were interactable, realtime rendered, and done well, I think this game would have done 10x better. If it takes you six months to come up with a great system for that while incrementally improving T:WGM in the meantime, that may pay off in the long term. If you start another project without thinking through how to improve that, and use the exact same workflow for h-scenes.... then I think you might be able to do a bit better than T:WGM (you will have some momentum), but you're putting a fundamental ceiling on how successful you can be and I don't think you'll ever get that breakout hit. If you do want a breakout, I think the most straightforward path is taking T:WGM and addressing the weaknesses.
That's just my 2c, so feel free to disregard it entirely, and no matter what you do I'll be checking out your future work
