I've tried to get into Factorio multiple times because I really love 2D isometric games, but the learning curve is incredibly steep. In my opinion, franchises like the original Command & Conquer were much more accessible for non-hardcore gamers like myself. Factorio feels less like a game and more like managing a complex software architecture.

The systematic and scientific approach is absolutely warranted if you want to produce legendary quality at scale. I brute forced my way to legendary mech armor with manually steered recycler loops and wasting millions of iron and copper in the process. Since playing on a Linux developer laptop with 0GB VRam and an onboard GPU (at 60FPS!) I couldn't have scaled it much larger anyways but the entire game, it's a fever dream of optimization and I love it.

The difference between Washing and Upcycling is enormous. That does fit my observations, it's nice to see it quantified.

Another aspect that is nice to see quantified is how big the effect of better quality modules is. You likely underestimate that initially since it's not linear.

The Space Casino mentioned in the post is not possible anymore in Factorio 2.1. Which is experimental at the moment, but I'd assume will be released soon. And experimental Factorio tends to be as stable or more as the stable versions of other games.

I would elect the people that does this kind of analysis and management for President.

You can make the descriptive / non-math text in LaTeX look better in math mode by using \text{} e.g. $T_\text{products}=...$!