Factorio Forums

Because it kind of reminds me of a sea anemone.


Copy blueprint

This lets a ship go full speed 500,000 km (probably more) toward the shattered planet before turning around, and uses no rockets whatsoever. Due to the asteroid collector layout, it also picks up nearly all of the promethium chunks from the front of the ship so almost nothing goes to waste. It can actually handle speeds of around 650-700 km/s, but I expect the devs will take away engine stacking at some point, so I'm not designing around that. I prefer loading biter eggs onto the ship then making the science as I gather the promethium, since that results in the highest possible SPM and really isn't all that hard. Hence why I optimized for a full-throttle design.

Required upgrades are: Laser Weapon Damage 17, Physical Projectile Damage 24, Railgun Damage 3, and Railgun Shooting Speed 25 (yes levels above 5 do actually speed it up, albeit asymptotically). Note that you'll need angled/sideways railguns on the top corners/sides of the ship, and some companion gun/laser turrets to deal with any asteroids with sideways motion. Gun turrets require red ammo.

Anyway, I figured I'd share this because it turned out to be pretty awesome; I didn't think a no-rocket design was feasible. And because we need more cursed designs for space platforms.

_________________

A bit of background: I have been optimizing promethium ships for late endgame. Originally I set out to produce 28,800 legendary promethium science per minute, but quickly discovered that the scale of such a thing was infeasible. However, 57,600 rare promethium science looked doable on paper. So I made a really freaking huge ship (4500 tiles wide) designed around maximal chunk collection to test my theory out.

What I discovered is that Factorio (in its current state) is nowhere near being up to the task. According to the debug menu, 95% of the game update time is being spent on asteroid collector and turret updates. Sadly, it looks like neither is multithreaded right now, so they each saturate a single CPU core and that's pretty much all you can squeeze out of the game.


Hopefully the devs can do something about this, because currently promethium science is far and away the biggest bottleneck for megabases right now. I thought it was going to be agricultural science, but it turns out you can crank out a lot of legendary agricultural science (32K/minute in my case) without too much of a UPS hit. So I got kind of excited, thinking that the cargo landing pad would be what ultimately dictates that upper limit of physical SPM. But, sadly, it was the promethium science.

Awesome design! Looks very efficient.

Too bad the promethium science doesn't scale well. It would be great to have an infinite research that scale like research productivity and that adds +10% promethium science productivity per level. Or being able to use the centrifuge/crusher to refine promethium into itself (1:1 ratio) to upgrade its quality.

FasterJump wrote: Tue May 13, 2025 9:36 amIt would be great to have an infinite research that scale like research productivity and that adds +10% promethium science productivity per level.

Funny, I was actually thinking the same thing yesterday. But then I realized they'd have to cap it at 300% like everything else, and promethium science needs WAY more than a 4x speed increase.

I think the design idea for promethium science (and how it encourages you to produce it in realtime as you gather the chunks) is really cool. The issue is there's no good way to gather promethium at scale for megabases, due to how horribly CPU-intensive a trip toward the shattered planet is. And I don't expect they can magically squeeze a 100x performance improvement out of the existing code.

I think the solution would need to be something like Kovarex enrichment, where you can multiply existing promethium by adding additional inputs. That way it could be produced much more UPS efficiently in a cryo plant or something. And maybe add a restriction that it has to be done on a moving space platform, so you can't just park the platform over Nauvis. But at least you'd be able to do it around the inner planets where the game isn't keeping track of literally hundreds of thousands of asteroid chunks of various sizes.

Actually, I changed my mind. Promethium doesn't need a buff. Why? Because as infinite research gets researched, we need to do promethium research less and less often.

To optimize material output:
After you research level 23 research productivity, you would then research 13 levels of mining productivity, and to do that you would need 10 less promethium pack than total red packs for all these mining levels.
After you research level 24 research productivity, you would then research 16 levels of mining productivity, and to do that you would need 12 less promethium pack than total red packs for all these mining levels.
After you research level 25 research productivity, you would then research 17 levels of mining productivity, and to do that you would need 17 less promethium pack than total red packs for all these mining levels.
After you research level 26 research productivity, you would then research 23 levels of mining productivity, and to do that you would need 21 less promethium pack than total red packs for all these mining levels.

I calculated these in excel, so I might be off by a level or 2. Still, the trend is clear: when research level approach infinity, the percentage of time you spend doing research productivity rather than mining productivity approaches 0. In practice, that means that as time goes on, promethium ships becomes less and less needed.This explains why many players find their promethium reserves pile up as they spend most of their time doing more and more mining productivity research. That is assuming the player wants to optimize research gained per depleted ore.

That's sad because infinite mining productivity research is the only one that scale linearly, meaning at high research levels, all science pack lines slows down (except the 4 basic ones). If the devs wanted players to spend 1 hours doing research productivity for every x hours of mining productivity, the cost of both research should have been both linear or both exponential.

However, there could be a fix for all science packs to continue to be used often: Mining productivity, the only linear research, should use all science packs!