What does it take to launch a rocket in 6 hours?

[mrvn]

I've played Factorio for a long time now and I'm certainly no speed runner. But there is one Achievement left I still want to get: Launching a rocket in under 8 hours.

So I went into helmod and entered all the science packs one needs to launch a rocket plus a few cheap extra sciences, like long inserters, that are just a pain to work without. I've added the rocket silo and 1 rocket too and then all the recipes to produce the inputs down to raw materials.

Next I set the time interval to produce all that to 6h. Why 6h? Well, because that's one of the defaults in helmod. I want to be under 8h so 1h is way to low and 12h is over the limit. I was hoping to calculate everything for 4h, leaving 4h for building and slippage. But 1h and 6h are the time intervals helmod has that are close.

Last I added all the factories required, 10000 belts, 1000 undergrounds, 500 splitter and 600 fast inserter at a guess for what I need for a belt system and steam engines and boilers to produce all the power for the factory. This increased the required factories so I had to increase the counts a bit till it didn't grow no more.

Overall I got this overview of what is needed:

Overview.png (176.48 KiB) Viewed 9488 times

At this scale the factory takes 6h to produce all the required science packs and material for the rocket. That would leave 2h to build the factory. If only I could already build it. I need the research before I can build the advanced stuff. And if I don't start the production of something within the first 2h then I will have to produce it faster. Or if production of something, iron gear wheels, backs up and stops the assembler then that also increases the time. At the given scale every machine needs to run exactly 6 hours non-stop. Only 2h slippage is allowed. A second problem is Oil. I need some oil processing and plastic to research the advanced oil processing the numbers are based on. So I probably need to add a bunch of oil refineries for the initial plastic production since their output is so much less.

So this will never work with exactly those numbers. But I hope they are a close enough guide to what I need to build. I will build everything a bit bigger, maybe double so it only needs to run 3 hours and I have 5 hours to build it. So instead of 2 (1.55) belts of iron I will go for 4 belts and so on.

Anyway, in case anyone else want's to try this or is simply interested in the details here are the numbers for all the recipes required:

mining.png (98.76 KiB) Viewed 9488 times

oil.png (65.98 KiB) Viewed 9488 times

factory.png (1.41 MiB) Viewed 9488 times

[shopt]

I'm not familiar with helmod so may be misunderstanding the output, but it looks like you aren't using productivity modules anywhere? P3s on the silo more than pay for themselves, even for a single rocket. It's such a no-brainer that every single speed runner does it. Throwing P2s in yellow and purple science (which should definitely use AM3s so you can fit 4 modules in them), as well as labs once you get to the 4 pack techs should more than pay for themselves even just for a single rocket launch. Plenty of places where P1s would probably also be a net positive. And that is just a very conservative list.

I know this is just a spreadsheet exercise, but given you are targeting 6 hours, there is not a single recipe where P1s don't pay for themselves in 6 hours. P2s also break even in under 6 hours for every recipe on your list except solid fuel and rocket fuel, though that assumes 100% machine utilization which doesn't always seem to be the case in your list. P3s also break even in under 6 hours for more recipes than the ones they don't.

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

[Kyralessa]

At this scale the factory takes 6h to produce all the required science packs and material for the rocket. That would leave 2h to build the factory.

No, it still leaves 8 hours to build the factory. It's not as if you do 2 hours of initial building and then stand still while waiting for everything to be manufactured. You and the factory will be working in parallel.

[mrvn]

At this scale the factory takes 6h to produce all the required science packs and material for the rocket. That would leave 2h to build the factory.

No, it still leaves 8 hours to build the factory. It's not as if you do 2 hours of initial building and then stand still while waiting for everything to be manufactured. You and the factory will be working in parallel.

The above figures need 6h of yellow science production to finish all the research on time. Any minute past 2h where yellow science isn't produced means it needs more assemblers. The above figures also need 6h to produce all the red science. So unlocking yellow science will never be possible in the first 2h. So everything needs to be scaled a bit.

Doubling everything so it aims at producing everything in 3h actually worked out quite well. I had to tripple some rows late stage because they lagged behind the rest and increase copper and iron plate production massively to match that. But I managed to launch the rocket with 36 minutes left on the clock and without stressing the gameplay.

launch.png (2.99 MiB) Viewed 9369 times

final-base.png (2.84 MiB) Viewed 9369 times

PS: I did add PM3 modules to the rocket silo. That really is a no brainer. I just forgot to do it in helmod. Gave a nice buffer at the end because the rocket gets a lot cheaper.

[mrvn]

I'm not familiar with helmod so may be misunderstanding the output, but it looks like you aren't using productivity modules anywhere? P3s on the silo more than pay for themselves, even for a single rocket. It's such a no-brainer that every single speed runner does it. Throwing P2s in yellow and purple science (which should definitely use AM3s so you can fit 4 modules in them), as well as labs once you get to the 4 pack techs should more than pay for themselves even just for a single rocket launch. Plenty of places where P1s would probably also be a net positive. And that is just a very conservative list.

I know this is just a spreadsheet exercise, but given you are targeting 6 hours, there is not a single recipe where P1s don't pay for themselves in 6 hours. P2s also break even in under 6 hours for every recipe on your list except solid fuel and rocket fuel, though that assumes 100% machine utilization which doesn't always seem to be the case in your list. P3s also break even in under 6 hours for more recipes than the ones they don't.

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

I did throw in PM3s in the silo but nothing else. I know it's worth it but having to produce all those modules is somewhat disruptive to the game play. You should really throw all resources into producing modules as soon as you research them and put them in every assembler. Updating assemblers to AM3s with even more modules is worth it too. But again, too lazy to disrupt the flow to make that switch. It's not required for an sub 8h run and much harder to plan out the required resources and assemblers.

I noticed that I wasted a lot of resources producing stuff on the belts. I have a lot for blue, yellow science left over at the end, quite expensive. Purple science got starved towards the end and it would have helped a lot to stop the blue/yellow science production from racing ahead. Similar stopping SM1s, rocket fuel and rocket control units when the rocket was actually finished would have saved a lot.

Planning this out and moving the producers closer to the consumers should reduce the wasted resources a lot. A bus design really isn't the most efficient when it comes to items wasted on the belt.

[mrvn]

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

You are right, just launching the empty rocket is enough to finish the game. Did that change? I thought you needed the satelite. That's probably wasted the last 1-2h.

[shopt]

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

You are right, just launching the empty rocket is enough to finish the game. Did that change? I thought you needed the satelite. That's probably wasted the last 1-2h.

https://www.factorio.com/blog/post/fff-275, last section (Freeplay victory condition)

[SoShootMe]

I did throw in PM3s in the silo but nothing else. I know it's worth it but having to produce all those modules is somewhat disruptive to the game play.

What do you find "somewhat disruptive"? You need PM1s anyway for productivity science packs; PM3s for the silo and the PM2s they require can be made with a couple of hand/chest fed assembling machines.

You are right, just launching the empty rocket is enough to finish the game. Did that change? I thought you needed the satelite. That's probably wasted the last 1-2h.

Perhaps less than you think: the vast majority of resources to launch a rocket with a satellite go into producing the rocket silo, filling it with PM3s (as you wrote elsewhere, a no-brainer) and building rocket parts.

[mmmPI]

The "there is no spoon" achievement require you to play with standard settings. It is disabled when using peaceful mode or lower than default enemy base settings. It's not just launchnig a rocket under 8 hours.

[mrvn]

The "there is no spoon" achievement require you to play with standard settings. It is disabled when using peaceful mode or lower than default enemy base settings. It's not just launchnig a rocket under 8 hours.

Funnily enough though you can select a large starting area and turn of pollution, reduce the spreading, the group sizes and the evolution.

The no pollution setting is the biggest change you can make. You won't trigger an attack because it stinks. It's basically peaceful mode. With even oil right in the starting area on my map I didn't even see a single alien, didn't have to explore at all.

[mrvn]

I did throw in PM3s in the silo but nothing else. I know it's worth it but having to produce all those modules is somewhat disruptive to the game play.

What do you find "somewhat disruptive"? You need PM1s anyway for productivity science packs; PM3s for the silo and the PM2s they require can be made with a couple of hand/chest fed assembling machines.

It changes the resource demands when you focus on producing all the modules quickly. It then changes the resource ratios when you place them in assemblers.

In helmod I would have to do 3 spreadsheets: pre modules, the modules production and post modules to get an even halfway accurate count of required resources. As I said, I was lazy.

You are right, just launching the empty rocket is enough to finish the game. Did that change? I thought you needed the satelite. That's probably wasted the last 1-2h.

Perhaps less than you think: the vast majority of resources to launch a rocket with a satellite go into producing the rocket silo, filling it with PM3s (as you wrote elsewhere, a no-brainer) and building rocket parts.

The satelite is comparatively cheap. But you have to research it first https://wiki.factorio.com/Space_science_pack_(research). Those 2000 science packs of every color represent a lot of time. At the end I had the rocket sitting there just waiting for the research to complete. All the resources for the satelite where already in chests next to the assembler, the assembler connected to the rocket silo, the silo set to auto-launch. All I had to do to finish was set the recipe on the assembler (see save game).

So 2000 science packs out of the 3292 purple and 2748 yellow weren't needed, so the majority of them. .oO(Wait, the rocket silo also needs 1000 yellow science, did I drop a 1000 yellow science packs in the spreadsheet?)

I could have just hit launch without satelite.

[SoShootMe]

The satelite is comparatively cheap. But you have to research it first

Ha, I somehow forgot it was a separate research . 2000x 5 science packs, plus a satellite nearly quadruples total launch cost (including the rocket silo and 4x PM3) in terms of resources. Moving on...

[mrvn]

I'm not familiar with helmod so may be misunderstanding the output, but it looks like you aren't using productivity modules anywhere? P3s on the silo more than pay for themselves, even for a single rocket. It's such a no-brainer that every single speed runner does it. Throwing P2s in yellow and purple science (which should definitely use AM3s so you can fit 4 modules in them), as well as labs once you get to the 4 pack techs should more than pay for themselves even just for a single rocket launch. Plenty of places where P1s would probably also be a net positive. And that is just a very conservative list.

I know this is just a spreadsheet exercise, but given you are targeting 6 hours, there is not a single recipe where P1s don't pay for themselves in 6 hours. P2s also break even in under 6 hours for every recipe on your list except solid fuel and rocket fuel, though that assumes 100% machine utilization which doesn't always seem to be the case in your list. P3s also break even in under 6 hours for more recipes than the ones they don't.

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

Let me get back to this as I did some math in helmod. I started out with the basic science setup ratios producing 0.75 science per second and the factory for building the rocket silo in 6h, all with AM2s and no modules at all. That's the top third of the picture. Included an extra line for putting PM3s in just the rocket silo as that's a "no brainer". It doesn't save that much factory but a whopping one third of copper and one fourth of iron.

The middle section is putting PM3s into everything but sticking with AM2s. It doesn't save that much on science and only the yellow science pack get significantly smaller to build. The rocket now only costs 1/3 copper and 1/2 iron of the totally no modules cost. Less drastic compared to modules in the rocket silo.

The bottom section I upgraded everything to AM3s and maxed out the PM3s, a whole 642 of them. Some size reduction for the non trivial science packs and some resource reduction. The rocket silo gets even cheaper down to 1/4th copper and 1/3rd iron. But you need 642 PM3s for that. And producing those over 6h would take 653 AM3s, which themself need 2418 PM3s, and cost 357k iron and 494k copper. That's as much as producing a rocket + satelite costs on it's own without any modules (except iron and copper switched around). Not to mention that if it takes 6h to produce the modules making it 12h to build the rocket instead of 6h. You would have to double everything to keep it in a 6h time window as without modules. So you have to pay 2 times what a rocket costs to save 75% of the rockets cost.

Summary: No, PM3s everywhere do not pay for themself. Not in a 6h window.

are-modules-worth-it.png (238.78 KiB) Viewed 9227 times

Last I also calculated the cost of PM3s just in the rocket silo and AM2s with PM1s everywhere else: It costs 12k iron (7.5k for PM1s) and 24k copper (16k for PM1s) to produce 4PM3s and 501 PM1s and it brings the cost of the rocket down to 24k iron and 42k copper. So that is a net plus just from the rocket but also makes the science a bit cheaper.

So I can confirm that PM1s are really worth it. You also get them early so that you can produce them in spades and put them everywhere.

Note: I have not factored in PMs in the science labs as I was mainly focused on the rocket and cost of a science pack, not their total amount. PMs save 7%, 11% and 20% in science labs. My original spreadsheet had 4 science labs, or 8 PM3s. You need 4 PM3s for the rocket silo anyway and you can remove them from the science lab once you have researched the rocket silo. So the first 4 PM3s are basically free. I actually ended up needing a bunch more science labs, so consider at least 16 labs. 32 PM3s are probably a bit expensive but 32 PM2s likely pay for themself, you need 20 PM2s to produce 4 PM3s anyway. So only 12 extra PM2s to build to save ~10% of blue science (need some blue for PM2s), 11% of purple/yellow science and 11% of 5/8 red and green (about 6.9%).

Note2: miners can have PM1s too. But that really only makes the ore field last longer. Unless you don't have enough ore for the rocket and can't go out and use a second ore field due to aliens or something just don't bother there.

Conclusion: Put PM1s into everything and upgrade the science labs to PM2s. Then when the rocket silo is researched shut down all science and build 4 PM3s from the PM2s. Any remaining ones can replace some PM1s.

PS: I always hated that the ratios between assemblers get screwy with modules. It just isn't nice to build.

[mmmPI]

Funnily enough though you can select a large starting area and turn of pollution, reduce the spreading, the group sizes and the evolution.

The no pollution setting is the biggest change you can make. You won't trigger an attack because it stinks. It's basically peaceful mode. With even oil right in the starting area on my map I didn't even see a single alien, didn't have to explore at all.

good to know

[shopt]

I'm not familiar with helmod so may be misunderstanding the output, but it looks like you aren't using productivity modules anywhere? P3s on the silo more than pay for themselves, even for a single rocket. It's such a no-brainer that every single speed runner does it. Throwing P2s in yellow and purple science (which should definitely use AM3s so you can fit 4 modules in them), as well as labs once you get to the 4 pack techs should more than pay for themselves even just for a single rocket launch. Plenty of places where P1s would probably also be a net positive. And that is just a very conservative list.

I know this is just a spreadsheet exercise, but given you are targeting 6 hours, there is not a single recipe where P1s don't pay for themselves in 6 hours. P2s also break even in under 6 hours for every recipe on your list except solid fuel and rocket fuel, though that assumes 100% machine utilization which doesn't always seem to be the case in your list. P3s also break even in under 6 hours for more recipes than the ones they don't.

Also I'm pretty sure you can drop the satellite if you are just after the achievement.

Let me get back to this as I did some math in helmod. I started out with the basic science setup ratios producing 0.75 science per second and the factory for building the rocket silo in 6h, all with AM2s and no modules at all. That's the top third of the picture. Included an extra line for putting PM3s in just the rocket silo as that's a "no brainer". It doesn't save that much factory but a whopping one third of copper and one fourth of iron.

The middle section is putting PM3s into everything but sticking with AM2s. It doesn't save that much on science and only the yellow science pack get significantly smaller to build. The rocket now only costs 1/3 copper and 1/2 iron of the totally no modules cost. Less drastic compared to modules in the rocket silo.

The bottom section I upgraded everything to AM3s and maxed out the PM3s, a whole 642 of them. Some size reduction for the non trivial science packs and some resource reduction. The rocket silo gets even cheaper down to 1/4th copper and 1/3rd iron. But you need 642 PM3s for that. And producing those over 6h would take 653 AM3s, which themself need 2418 PM3s, and cost 357k iron and 494k copper. That's as much as producing a rocket + satelite costs on it's own without any modules (except iron and copper switched around). Not to mention that if it takes 6h to produce the modules making it 12h to build the rocket instead of 6h. You would have to double everything to keep it in a 6h time window as without modules. So you have to pay 2 times what a rocket costs to save 75% of the rockets cost.

Summary: No, PM3s everywhere do not pay for themself. Not in a 6h window.

are-modules-worth-it.png

That's interesting, but also not that surprising. A naive "PM3 everything" approach will put a lot of modules in some very slow recipes, whereas a minority of the modules do most of the heavy lifting. So for 6 hours while a majority of recipes are worth putting PM3s in, that's not the same as a majority of machines, nor all the machines.

These are the recipes where you would get the most benefit:
* Rocket Part
* Yellow Science
* Purple Science
* Blue Circuit
* Labs once you are up to 4 and 5 pack techs
* Green Circuit
* RCU
* Iron gear wheel

These recipes would come out slightly ahead or roughly break even:
* Sulfuric acid (Largely comes down to low machine utilization)
* Sulfur
* Plastic
* Iron Stick
* Copper cable
* Flying robot frame
* LDS
* Red Circuit
* Advanced oil processing (sorta depends how much you value oil compared to ore, I assumed 10:1)
* Electric Engine Unit

Recipes where the AM3s cost more than they save over 6 hours:
* Green Science
* Battery
* Engine
* Light oil cracking
* Heavy oil cracking
* Red science
* Smelting (iron, copper, steel, stone)
* Lubricant
* Solid fuel
* Rocket fuel

I think even if you just left out solid fuel, rocket fuel, and smelting you would see very different results. They take lots of modules and have a very long payoff time.

The other thing is the overall module count (speed plus productivity) can be significantly reduced by using speed beacons in the right places.

[mrvn]

Green science is 10 assemblers or 40 PM3s, Battery, oil stuff red science is all less. Smelting was steel furnaces so no modules support there. Electric furnaces come too late to make much of a difference, most of the time they wouldn't be available so I never tried those.

So your exclusion list is just a fraction of the 642 PM3s, lets call is 500 PM3s left after removing all those. You're not even close to saving what they cost nor have the time to produce them all.

Did you notice the 2418 PM3s needed in the assemblers to make the PM3s at even the cost of a rocket (with satelite and all the techs for it)? Without PM3s in the PM3 factory the cost is much more than the overall cost. Putting PM3s into the "get the most benefit" list might be paying of. But every module you remove lowers the savings so I would say it's a 50:50 chance even with just that list. Labs are definitely the most likely candidate for paying of the modules because of the amount of resources they process for the last 3000+ science packs each of every color.

[SoShootMe]

That's interesting, but also not that surprising. A naive "PM3 everything" approach will put a lot of modules in some very slow recipes, whereas a minority of the modules do most of the heavy lifting. So for 6 hours while a majority of recipes are worth putting PM3s in, that's not the same as a majority of machines, nor all the machines.

But, starting from nothing, you don't have PM3s available for most of those six hours. As has already been mentioned, PM3s in the rocket silo more than pay their resource cost back with one rocket. With that metric, PM3s in production science pack assemblers are worthwhile if they give a net saving for the 1300 needed to research SM3 and the rocket silo, and so on. There are clearly many possibilities here (combinations of machines with modules, speed modules and/or beacons) and other factors (eg productivity gain in making modules), but I'm not aware of any case other than the silo where level 3 modules yield a net resource reduction in the context of reaching the first rocket launch... level 2 modules in some of your "most benefit" list are good though.

[shopt]

I think what's happening in this discussion is that there is actually 2 discussions going on.

One is the theoretical spreadsheet exercise of 6 hours of production to get the science and components for a rocket launch, which is mostly what I'm talking about.

Then there is the discussion of actual practical gameplay which seems to have started after I raised questions about the original helmod pics. In this discussion you don't build a factory then let it run for 6 hours.

I'll admit for that I'm not considering the cost of the factory to build the modules (though using the worst case factory cost with zero beacons or speed modules as a counterpoint doesn't really illustrate much), nor that you need to spend a lot of science before you get PM3, and therefore can't use PM3s to save resources that you spend before you get PM3. Those are all valid points, and my list is meant to address a theoretical 6 hours, not an actual game to rocket launch. But still, I'll admit my list is based on applying PM3 backward in time, which can't be done. I got that wrong.

My points was that PM3 in everything is not optimal, even in theoretical spreadsheet land, over 6 hours. It seemed like my post was misread as claiming that PM3 everywhere would break even, rather than the narrower point I was actually making.

As for whether PM3 does or does not pay for itself for the first rocket launch, in and of itself you can't say whether it will or wont. It depends on how long those modules get to run for, with speed beacons increasing the effective run time of each module. If I'm aiming for launching a rocket as cheaply as possible, I could beeline to PM3 and then build a very small factory where PM3s will break even in many recipes for the first rocket, due to the long run time. Or I could build a speedrunner esque factory where the time from PM3 tech to rocket launch is so short that it comes out at you say; PM3s only have enough time to break even with rocket parts. Over 6 hours PM3 will break even on lots of things, over the practical time PM3s are usable in a vanilla rocket launch where speed is being aimed for, I agree they wont.

I think the one thing we can all agree on is that PM3 in the rocket silo is a no brainer, and PM3 everywhere doesn't make sense if the goal is to launch one rocket. Hopefully soon I get some time to run the calculations I think would produce a better result, which I suspect is a mix of all three module tiers. Moving to the other side of the world turns out to take up a lot of time.

[mrvn]

if you don't care about time you build 1 assembler and no beacons. And since we already established that PM3s for the silo pay for themself you use them in the assembler till you need them in the silo. Can't get any cheaper.