[bhouston]

It does feel like they bet on Unity's High Definition Render Pipeline and it locked them into a specific way of development that was hard to escape from once it proved problematic.

City simulation games (Sim City, Factorio, etc.) are sort of a unique beast in that they have a ton of small scale detail that is animated and and dynamic.

The choice of engine here matters a lot, because engines are often highly optimized for specific assumptions and the assumptions of standard games (mostly static worlds with just a few dynamic entities - a platformer, a first-personal shooter) do not hold.

The studio taking this over should ensure they have some really good low level 3D devs guys on the team and a flexible engine.

I think that a home built engine could work in these cases, but only if you have the right guys for the job.

[bob1029]

Unity can definitely support scenarios involving 100k+ scene elements. The problem is that you cannot place the art team any meaningful distance away from the development team when using something like ECS. Game objects + components allow for a high degree of decoupling between the art & technology teams, assuming the total # of scene elements is well bounded (<10k).

A custom engine is probably not a bad idea for a city builder. I feel like ECS is a good middle ground because you would need some pattern like this anyways and it would serve as a good reference if you decided to go fully custom.

I would challenge the notion that we couldn't develop an effective city builder within the constraints of a reasonable # of game objects. Players may be compelled to accept an experience that has a significantly smaller world size if the richness within each unit of world space is much higher.