Factorio Speedrun 9 hours 57 minutes

[trevbillion]

I went as fast as I could to build factory and get missile defence. In-game clock shows a hair under 10 hours. This video is a tour of the winning factory plus victory screen!

https://www.youtube.com/watch?v=Ac3s9iT6QYE

How fast can this game be beat? Anyone else interested in speedrunning it?

Thanks for the great game Factorio developers!

P.S. speedrunning marathon is going on now for anyone interested in it. Check it out! http://gamesdonequick.com/

[hoho]

I'm quite certain that speed of completion GREATLY depends on the specifics of map. Time difference between good and bad resource/enemy spawns can be huge.

Though this game is certainly one of the more complex games to plan out the fastest and best time possible and it would always be possible to shave off a few minutes here and there

[DaveMcW]

I'm waiting for functional replays before I show my speedrun time.

I'm more shocked at your 6 advanced circuit assemblers than your 4 electronic circuit assemblers. The optimal advanced:electronic circuit ratio is more than 2:1.

[Drury]

It's incredible that there's already enough content in the game that it takes 9 hours to speedrun.

[hoho]

It's incredible that there's already enough content in the game that it takes 9 hours to speedrun.

Indeed. Can't wait for the "proper end-game content" they'll be adding Soon(tm)

[DasMonzta]

It's incredible that there's already enough content in the game that it takes 9 hours to speedrun.

Just set the ressources demands for all products higher and it will take more time. That is not only a matter of "content".

[RawCode]

there is no much content at this time, but much grinding due lack of advanced devices.

energy generation is great sample of grindy nature of current gameplay.

[Sander Buruma]

I'm very impressed with your speed run and how tidy everything is.

You say a speed run under 3 hours... I'm not sure that's possible..

[Artman40]

Under 3 hours is probably possible. What have to do though is the following:

1. Find a good map seed.
2. Lay out a plan ahead.
2.1. Order of research.
2.2. Dealing with the enemies.
2.3. Schematics for optimizing the factory.
2.4. And so fourth.
3. Not just stand around, meaning you should always be moving and crafting things.

The pre-research part is also crucial as appropriate openings will save you a lot of time.
Obviously, after that, the run needs to focus on two things:
1. Finding a good research path to rocket defense, skipping as much unneeded research as possible.
2. Getting module 3 factories going, along with everything it needs. Less demanding ingredients should be less of a problem.

[Sander Buruma]

I can not fathom how <3 hours might be possible on a default settings map because the resources just aren't very abundant and oil in particular is very scarce.

[Artman40]

This is where speed modules come in. They work great for oil wells as the resource is infinite but the resource wells are not numerous. Need for additional power should be compensated for speed of which the player builds new things.

[Sander Buruma]

I just did it in 7 hours, 7 minutes and 9 seconds.
Some screenies:



What the first screenie doesn't show are the 4 assemblers I used to produce the modules because I disassembled them. They stood right where the rocket defense now proudly stands. What it isn't showing are about 70 miners for copper 35 for iron and 20 for coal.

This was all default settings except it was peaceful mode and biter base richness was set to max (which didn't seem to have any effect whatsoever.) I planned everything out in a sheet and tried to keep the production rates at 1 module / min, which meant 1080, 190 and 30 green red and blue Circuits per minute. (As indicated by my spreadsheet.) In the final hour or so basic circuit production & consumption was at an average of

It was a bit challenging to find all the oil but by shrinking my research part of the base I could recover the needed oil. Every important assembler was given speed 1 2 and 3s and the processing unit assemblers were given productivity 3s and + speed 3 beacons because I wouldn't bother to move them to more open grounds. After accumulating enough modules I recovered all the level 3 modules from assemblers+beacons (some of which were consuming 1.3MW of power) I built the rocket defense. It all ran off of 40MW of power.
Major faults in my approach were trying to attack biters with guns instead of laser creep, not using productivity 3s earlier to produce purple science and not moving my major factories to more open space so I didn't have to use a ton of modules to keep up with production. Another thing I forgot was to put speed 1 mods in my miners earlier on. Next time I'll try to put my processing unit assemblers straight next to my processing unit assemblers because it saves on logistics. Most of my assembler 3s had speed 1s (or better).

Imo <4 hours is probably possible on default resource settings, but much less than that surely not...? In order to achieve that I need to make a more comprehensive and better planned out construction plan so I get everything up and running much faster. The processing unit and and module 2&3 part of the chain is more flexible because it doesn't take a lot of production time.

[Artman40]

Nice! How did you tackle the beginning of the game until the first research?

By the way, did you use assemblers to produce most of the buildings stating from the mid-and late game?

Since now you can select the random seed, should a good seed be used?

[Sander Buruma]

I crafted most main buildings by hand but not the ingredients from which they were made. Several assemblers were dedicated to maintaining a stock of 500 gears and 200 / 100 red / blue belts ready for use. The first modules produced were speed 1s because they allow you to build mass AM3s. 10 oil wells were made with some super fast assemblers worth about a total of 9.0 crafting speed because those just took too darn long by hand.

I don't recall how long it took me in this run but in my currently running attempt I got to 0.6 R/G packs per minute feeding 20 labs, 50 min after game start. I started by putting the starter burner/furnace on iron and getting some nearby lumber (40 units worth) and 15 stone then making two furnaces and a boiler. Manually mining the next bits of iron while keeping my burner/furnace fueled as well as the extra furnaces stocked with iron and fuel of their own. After smelting about 50 iron packed my extra furnaces and went to the copper patch nearby and smelted about 15 copper from manually mined ore. Then went to copper and set up a boiler/steam engine with a nearby electric coal miner and about 10 patches worth of belt to feed onto. Used some of the coal to go back to iron and go to the iron patch with electric poles. Once there I made another electric miner with two furnaces (periodically putting surplus ore into the 2nd furnace by hand) and used the new iron to make an electric copper miner with two stone furnaces. After that I went back for some extra coal, extended the belt to the boiler added 2 boilers, 1 steam engine and 3 electric inserters to make it automatic. Then I ran back to the copper patch for some copper and then the iron patch to extend it further with 2 more electric miners, while ensuring that the new miners were vertically aligned so I would not have to replace the miners once I'd go reshuffle the furnaces to make them automatic.

By about now I'm making 10 red science packs and a lab and get automation researched. Then I ramped up iron to about 6 miners while maintaining copper to just 1... and so... on... so I could quickly build an array of 3 AM1s and 10 extra boilers and 8 extra steam engines... Also added another copper miner and made that automatic eventually... By this point I've made 3 assemblers to make gear/circuit production semi-automatic and its taking off from there. Unfortunately the replay desyncs after tick 8909. That would be much simpler than typing a report.

You should have a good seed, or you should restart your map until you have a decent iron/copper field (100K iron/copper in each). and at least 10 oil spots in your starting area. Its otherwise hopeless to get a good speed run on default settings in my opinion.

ps. the seed I used for the 7h speedrun is 142857. (If someone here is a number nut, they know what that number represents.)

[Artman40]

Very interesting.
How many beacons with speed modules were oil wells surrounded by?

[Sander Buruma]

well none, there were about 15 spots within about a 7x7 sector grid. enough for the plastics & batteries that were needed. Counting at 28K adv circuits and 5K batteries, I used about 100K petrol and probably about 120K worth of of oil, seeing as every drop of it was being continually converted to petrol.
There were 3 beacons and they were spent boosting AM3s with Speed 3. The highest priority AM3s for processing units were at an average productivity of about 30% and needed 2 beacons to stay at a speed of about 1.5. I tried putting productivity into the module assemblers but it won't let you. I'm a little disappointed.

[MtNak]

Wow! That was amazing Sander Buruma

[Artman40]

Is it best to do all the logistics with conveyor belts?

[Sander Buruma]

it is especially for the base resources. It costs a lot of resources to research bots, and yet more to make them economical to use, and yet more to build enough of them even when you're at that point. You're just delaying things with logistics bots. By the time you're ready to use them you've spent so much time that it's become a slow-run.

[Artman40]

Has anyone tried to speedrun campaigns?