For the first one I think it can be balanced out for expenditure point of view, more "floor tile" should not linearly translate like 10x10 tiles is just 100/9 times more expenses than a 3x3. And 10x10 production throughput isn't just 100/9 times as 3x3, also the adding of a single unit will be like buying production line machine cost way more than current level, than this will be a economy of scale problem. How to balance the longer-term average expense with production quantity and prices. Players will have to find the right throughput and layout not just using every tile possible, too much it's a burden of expenditures, too few you will not be able to sell enough goods, and become just a shadow compare to AI's much higher base production (essentially means the right scale for human players equivalent to current scale have to be at the far end in very large size not the smaller 3x3 size, in smaller scale players can not get nearly enough income to expend fast enough)megapolis wrote: 1. Even now at a certain point we reach the moment when we don't have a place where to invest the money. Billions are just rotting on our bank accounts. There's just no need to optimize anything.
2. 3x3 or 10x10 - does not matter. Within a week we will have a perfect factory design for every product on the forum. Just copy it and save in layout plans library and the game is over.
3. There's a lot of games that are focused on factory planning. Big Pharma, for example. Or Factorio that I can only describe as a masterpiece.
As to the second issue, that's not possible, you underestimate the possible permutation of such a combination. A 9 grids with 3 major different units have at least 3^9 permutation (obviously it's much more, since there isn't just 3 types of unit in a factory, and in fact if you consider empty tile as one, this is at least 4^9), and if limited to like only 1 to 3 of input types is permitted, at least 1 output is needed and every manufacturing unit is equal with no difference, as well as the rotation of 8 direction is also equal, and finally most of them are not functional and usable, this number will be reduced to around merely thousands, but even so there's no true optimize layout yet, sometimes two layout with different throughput has different advantage, or sometimes you need to mix with other products, hence even in this limited form, the optimization answer is still interesting enough for human to play as a mini-game. But for production chain like up to 10, and 100 tiles, the scale will be at least 10^100 per product (completely discarding other more permutation like how many different types of goods can be produced intermediately and products that can be shared in the same factory with similar intermediary products), even let's say 1 out of 1 trillion trillion (1/10^24) of them are usable and functional connections, it's still at the level of 10^76 possible combination. This is at the scale equivalent to the total number of atoms in the universe. And I can guarantee a lot lot of them will be roughly the same "effective" as each others. I doubt we will see the most optimize layout anytime soon. (This is also why I asked before how to make AI be able to play such a game, this is essentially impossible to solve dynamically and certainly impossible using libraries of default layouts)
As to comparison with other games, I think it's really comparing apple with orange, the appeal of this will be there is an established open world procedurally generated economic environment for players to base their layout for. Imagine if you can sell your products, and purchasing components in factorio not just have to build every one of them through production chain, and the price is not some fixed data generated from random function, but from the larger world outside that are basically many competition factorio AIs building their own "factories" in real time and supplying them to you or from you. That's what I was talking about worthy of a new game concept.