You can't just ignore the fact that the average gamer doesn't have an FPGA. For such a game you would have to ship the FPGA with the game, making it prohibitively expensive.
Well, you could ignore it, if you would be developing a new gaming console with an integrated FPGA unit.
It's just never worked out to justify itself by the stakeholders who would approve such a thing. The standard gaming hardware configuration has always tended towards a little bit of general purpose computing capability plus a very focused specialization in graphics. And the standard incentive structure of game production tends towards a rapid deployment of assets authored in general-purpose content creation software, not a game driven by new algorithms. These things drive away from the experimental places where FPGAs could shine - unique ways of creating audio and video signals, customized inner loop optimizations, etc. The first thing everyone will be asking is "OK, but how do I make it portable?"
Sure, you’re completely right, but as a hobby project it would be pretty neat!