Lowering the polycount will not bring any notable optimization.

why? less polygons mean less surfaces being rendered, and from what I've experienced it seems that the scenes with high poly models are when I notice the most chug

Nowadays renderable polygons do not impact performance too much. I once tried putting a mountain into Serious Engine, which had around 4 million polygons. The editor had a hard time processing it for edit, but the game rendered it without any issues. That's not where all the problems come from. Now, if you enable collision on this model... Yeah, that's where any PC is gonna die.

Thing is, today improving performance isn't as easy as "just remove X and everything will be fine". FPS gets lower and lower by many different pieces, which all require improvement and optimization to work better together. Hundred of enemies (AI logic processing), far trees draw distance (lots of geometry to render - EVEN considering that each tree is just 1 polygon if it's farther than several meters from camera), beautiful numerous particle effects (brings down the CPU) - now imagine that there are dozens and dozens of such factors, interacting with each other and making life worse.

So, what am I saying. Can someone make an optimization mod, which also includes polycount lessening on problematic models? Sure, of course. This will be quite an effort, to optimize so much stuff, and it won't be too good without access to engine's source code. But it can be done.

Can someone make a mod which just reduces polycounts? Ugh, yes, but it won't help at all. Like, I dunno, maybe it will add 1-2 fps in certain scenes, but that's it.

Please, don't underestimate the complexity of game engines.

I have no interest in undermining or underestimating the complexity of game engines. I actually graduated with a degree in computer science and a concentration in game design. Please do not assume i'm here to rag on the game or be a nonce.

You present interesting points. What would you say needs to be the number one focus then when talking about optimizing the game?

If you mean just Serious Sam 4, I'd say I don't know, even though Im pretty familiar with SE. Croteam did their best to optimize the content part, even if there are some overlooked parts like polycount of St. George statue. Problems, most likely, lie in various code parts, fixing which will probably require so much work that Croteam couldn't afford to keep doing it - remember, there is a business side of things, they need to pay salaries etc, cant keep working too long on stuff which won't bring cash anymore.
For us, modders, that means we can just clean up some content things which Croteam guys overlooked, and that's it.

However, if you mean optimizing games in general - my whole point is that there is no such thing as number one priority. Have to see the exact case of performance issue, then be familiar with exact case's engine, in order to check what can be done for improvement.

https://steamhost.cn/steamcommunity_com/sharedfiles/filedetails/?id=3153499291 problem solved :)