Page 1 of 1
Unsolved mysteries
Posted: Mon Apr 03, 2017 7:13 am
by gustaphe
I'm taking a course in experimental design and statistics, which includes a project.
I'm thinking of writing mine about Factorio, but I'd have to find a question that lends itself to statistical analysis (I don't have time to e.g. measure the effect of burner boiler feeding on game length).
So, what unsolved Factorio mysteries are there? Is there anything you want to see formally examined?
Re: Unsolved mysteries
Posted: Mon Apr 03, 2017 9:23 am
by dog80
How is it possible to store multiple chemical plants inside a wooden chest?
Re: Unsolved mysteries
Posted: Mon Apr 03, 2017 10:35 am
by terror_gnom
Maybe some data about different train accelerations and speeds depending on the number and placing of locos and cargowagons?
Re: Unsolved mysteries
Posted: Mon Apr 03, 2017 8:23 pm
by ili
dog80 wrote:How is it possible to store multiple chemical plants inside a wooden chest?
Re: Unsolved mysteries
Posted: Tue Apr 04, 2017 5:49 am
by Frightning
Inserter throughput to and/or from belts based on corresponding throughput between containers, as a function of base inserter throughput, stack size, and belt speed (tier).
Re: Unsolved mysteries
Posted: Tue Apr 04, 2017 10:37 pm
by ssilk
The car has a bigger inventory than me, but I can put it into my pocket.
How can I make a 180 degree turn with a radius of 12 meters with a speed of over 250 km/h without dying?
Why are the enemies so stupid and take always the same path for dying?
You can built power cables that are in the way like the net of a spider. But you even don't touch it.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 1:21 am
by Mehve
You can put thousands of boxes inside a box.
Productivity modules increase both the viable output AND pollution. Where is this extra mass coming from? Or, where was it going without the module?
More seriously, a comprehensive analysis of solid fuel and rocket production, from pump to final production, with varying combinations of efficiency and production modules, as related to their net energy yield, could be interesting.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 5:38 am
by iceman_1212
I think OP is asking for questions that lend themselves to statistical analysis as opposed to deterministic questions (e.g., oil processing math).
I'll throw out a few hypotheses:
-getting electric furnaces earlier correlates with lower income / production at, say, the ~20 hour mark
-same as above except with solar energy (i.e., staying on steam longer leads to more income/production at some time mark)
-maybe something having to do with the distribution of logistic vs. construction bots and production in late-game
-in practice, bigger train size does not result in nearly as much more throughput as would be predicted by theoretical calculations (because of effects that are hard to account for e.g., acceleration/deceleration at intersections)
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 9:55 am
by Frightning
iceman_1212 wrote:I think OP is asking for questions that lend themselves to statistical analysis as opposed to deterministic questions (e.g., oil processing math).
I'll throw out a few hypotheses:
-getting electric furnaces earlier correlates with lower income / production at, say, the ~20 hour mark
-same as above except with solar energy (i.e., staying on steam longer leads to more income/production at some time mark)
-maybe something having to do with the distribution of logistic vs. construction bots and production in late-game
-in practice, bigger train size does not result in nearly as much more throughput as would be predicted by theoretical calculations (because of effects that are hard to account for e.g., acceleration/deceleration at intersections)
Having data to analyze is enough to do some basic statistics. Hence why I suggested inserter throughputs involving belts.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 11:05 am
by gustaphe
iceman_1212 wrote:I think OP is asking for questions that lend themselves to statistical analysis as opposed to deterministic questions (e.g., oil processing math).
Yes, I want to be able to conduct experiments, each of which preferably takes less than 20h.
I was thinking of pipe physics or train throughputs in roundabouts vs. crossings.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 11:51 am
by mophydeen
pipe physics would be interesting.
max capacity
difference between fluids: lube, petro, crude, water, acid, heavy, light
distances
with or without pump(s)
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 1:43 pm
by Nich
The problem I see is factorio is very deterministic. Statistics needs more randomness doesn't it?
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 2:10 pm
by gustaphe
Nich wrote:The problem I see is factorio is very deterministic. Statistics needs more randomness doesn't it?
Not neccessarilly. I'm looking for questions that can't be solved by simple calculation, but is better examined through experimentation.
But yes, I'm sure that most quick experiments in Factorio could be replaced by asking the devs exactly how the thing works (i.e. the religious method). Figured it was worth a shot to check if there are any systems which are random or chaotic enough to be examinable.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 2:30 pm
by DerivePi
I think the only thing in Factorio that would lend itself to statistical analysis is the behavior of the players. Collecting this information would be a feat in itself. Perhaps look at badges or comments on Steam. Maybe open a server that collects information on the people that play. Given a population of n players in a multiplayer game, what is the average number of belts placed by a player, standard deviation, etc...? Maybe do a review of the youtube plays available for other player behavior.
I just think anything about the game function itself is a distinct calculation (it is deterministic) and would not reveal anything special through statistics.
Re: Unsolved mysteries
Posted: Wed Apr 05, 2017 3:16 pm
by Deadly-Bagel
I convinced my bro to do a school project on fast food advertisements vs reality. The closest a product got to as advertised (usually relative to the provided drink) was the McDonalds McChicken burger - at something like 53% of the size advertised. But we did get to eat a lot of junk food ^^
If you want my opinion, try not to pick something specific to the game, you want something with real-world parallels. Pipe flow for example is abstract and while it would be nice to have more efficient designs involving fluids it will only be of interest to those who play the game, without the context the report will be meaningless unless you can say something like "This is how you can use Factorio to plan and predict X".
Re: Unsolved mysteries
Posted: Sun Apr 09, 2017 6:20 am
by dog80
How does the Lazer shoot slow moving lazer objects? What kind of technology is that. - yes i know there is a mod for that... its not so funny anymore now that i though of it.
Re: Unsolved mysteries
Posted: Mon Apr 10, 2017 5:17 pm
by Ingress
Piggybacking off of the laser comment, How does the worm "spit" go through walls and turn corners.
Re: Unsolved mysteries
Posted: Mon Apr 10, 2017 7:17 pm
by Ranakastrasz
Well, the game has a few random bits that could work.
Alien base design. Its random, but how many spawners/worms of what type show up, what kind of organization, how many any how often. There is info on this stuff, but I think it could work.
Resource spawning. Without actually reading the code and figuring out the mechanics using stats, figure out the kind of size and depth you expect to get from patches based on various game settings.
Re: Unsolved mysteries
Posted: Thu Apr 13, 2017 11:04 pm
by ssilk
The shadows show always to east. The sun goes up in west and keeps there until afternoon.
Also the moon...
The lights in the nights don't cast any shadows... but the moon does.
I have no clue, how they found that planet. The Factorio world exists probably in a geocentric universe... on another planet than earth.