New Achievement Ideas

[mxlppxl]

I'm working my way through the achievements slowly...some I might never get...but the game never gets old.

I would like to see some other achievements such as:

1 million copper bars
1 billion copper bars
1 trillion copper bars...etc

Could anyone realistically get 1 trillion coal/iron/copper etc? Who knows...but I'd certainly give it a go.

Mxlppxl

[Deadly-Bagel]

Well Copper Bars don't exist so no.

Assuming you mean Copper Plates, have you got the 20 million Electronic Circuits achievement yet? They are by far the highest consumer of copper (something like 99% of your copper will become circuits) and that achievement requires 30 million Copper Plates. That's the sort of scale we're talking here. To get to one billion, you would essentially need to make 33 times that achievement, then a thousand times that to get a trillion.

How about looking at the data and getting a handle on the game before throwing random numbers out?

[ssilk]

I'm not sure, if that is really an enhancement of the game. IMHO this could be done as mod.

[Vexos]

I think what you really need is some goal, that seems to be already adressed by the 0.15, with the new science. I still agree with you because there is a huge gap between other achievements and the circuit one, witch is much more time consuming than the others, unless you make a game focused on that. I really liked the challenge achievements like "there is no spoon" and "lazy bastard" and more of that type whould be nice and improve a lot the replayability (as if it was lacking...).

[mxlppxl]

Copper plates...yea...got my wires crossed with WoW.

I am working on the 20M achievement now but I estimate with playing 2-3 hours a day, it'll take me at least a month.

Just threw those out there as ideas...

"How about looking at the data and getting a handle on the game before throwing random numbers out?"

appeciate the encouragement there...I'll probably not post again.

[Ragnaman]

Personally i did not enjoy the 30million green circuit achievement.

The throughput achievements however "300 items per hour" were fun, as the challenge was to scale up your factory (which in turn surfaced flaws in your design from ground up), not pointlessly copy paste green circuit factory across the map where copper/iron was found.

Launching 2, 5, 10 rockets per hour from the same rocket silo (not sure if 10 rockets from the same silo is even possible) would prove to be the ultimate challenge for any factorio player.


I hoped my idea was understood:
Achievements that require more thinking than wasted time are fun,
Achievements that require less thinking than wasted time are not fun.

[BenSeidel]

I don't know how the speed improvements will go with 0.15, but looking at some numbers out of one of my larger factories:
Iron & Copper plate: max production 117k plate/ minute COMBINED - it's a smart furnace setup.
Green circuit: 61k/min

This runs at 30 UPS on my system, so you have to halve these numbers to get real time production, or double the game time values below.

So, the 20 million achievement *should* take about 328 minutes or 5 & 1/2 hours to achieve. (11 hours real time).
1 billion copper plate, if I get everything to just smelt plate: 8547 minutes or 142 & 1/2 hours or just under 6 days.

Hmmm.... I started this to show that 1 billion would just be stupidly insane, but it appears doable. 1 Trillion however....

[thereaverofdarkness]

Trillions are definitely possible as long as the paging file is large enough to handle the number of units you'd need to get that amount of production done, and your computer can handle the CPU/GPU/memory load.


Perhaps the one million achievement says: "Good work, newbie! Keep this up and you'll soon be a veteran player!"
Then the one billion achievement says: "Impressive! Your dedication to this game is surpassed only by your vast amounts of free time!"
And the one trillion achievement says: "New goal: go outside and discover the world beyond Factorio!"

[Deadly-Bagel]

Trillions are definitely possible

Assuming you don't constantly need to search for ores (which would involve constantly reloading the game as achievements only count in vanilla) and being generous at producing 100,000 plates per minute it would take over 19 years to reach one trillion plates. That's assuming you are able to reload the save each time the average drops to nearly 100,000. I would say that is outside the realm of possibility. Technically you could do it but realistically you could not.

[thereaverofdarkness]

Trillions are definitely possible

which would involve constantly reloading the game

The game world is limitless as far as I know. One could simply continuously set up new trains to get supply from farther and farther away, and feed it all to one enormous factory.

[orzelek]

Trillions are definitely possible

which would involve constantly reloading the game

The game world is limitless as far as I know. One could simply continuously set up new trains to get supply from farther and farther away, and feed it all to one enormous factory.

There are some practical limits. I belive it was changed at some point that map actually ends around 2 mil tile mark. I'm not sure if it's 2 millions to left and right or total. Going that far on map even with something very fast would take a lot of time.

After recent changes with ore amounts increasing further away or with RSO which does same thing you should start getting more and more ore further out you go. RSO uses exponent there so after reaching some distance (should be quite far - never tried that or calculated how far ) ore piles could go into high numbers.

[Deadly-Bagel]

Trillions are definitely possible

which would involve constantly reloading the game

The game world is limitless as far as I know. One could simply continuously set up new trains to get supply from farther and farther away, and feed it all to one enormous factory.

One could not do this for 19 years even if the map were truly infinite.

And keep in mind that as you are producing 100,000 per minute, that means every ore patch has to be tapped and connected in under one minute per 100,000 ore in the deposit. So if the deposit has 500,000 ore, you have 5 minutes to get it to your furnaces. It is much more feasible to set up but not connect several mining outposts, then connect them all at once and let the game run for however long it takes for them to run out. To have it running like this for an hour you would need to provide 6 million ore in that time. Can't say how many outposts that would be, maybe 1-2 million ore per outpost on highest ore settings? You will also want to put speed modules in the miners closer to the centre to better balance supply as the outer miners will run out and the supply will drop. You would need some wicked balancing to get this outputting properly... And then apply that to at least 6 outposts, just so you can run for an hour.

Also keep in mind that this is bound to generate a TON of pollution and if you want to run it unattended for any period of time (not like you're going to sit there for 19 years) you'll need a hell of a network to supply outposts with spare construction robots, repair packs, walls, towers, etc.

[BenSeidel]

Trillions are definitely possible as long as the paging file is large enough to handle the number of units you'd need to get that amount of production done, and your computer can handle the CPU/GPU/memory load.

I'm assuming that you mean that it will be possible in the future, with more powerful machines & a multithreaded factorio? The way the game currently runs, a trillion of anything as Deadly-Bagel said, would take about 19 years (give or take a decade) as we have pretty much found the upper limit of the production capabilities of a base. Sure, if you only focused on ONE production and used something like compression chests to store it, you could reduce the time. It may even only be a year or two of game time for the one achievement.

However, if the game is corrected to take full advantage of any number of cores, then it would not take that long at all. I get about a 0.2% increase in GDP of my bases every minute. This is fairly consistent now with the increasing resource values the further you are away from the origin. It's lower than it should be because I am continually having to redesign areas to deal with longer flight times and to reduce UPS consumption, etc. I also don't play with biters, so that makes a massive impact.

Anyway, assuming that it is possible to get a 0.2% GDP increase, then it should only take 3338 minutes of game time to increase all the production by 1000x, thereby making the trillion achievement reachable. In Fact, each additional 1000x increase in an achievement will only take an extra 2.3 days worth of game time to achieve.

There are some assumptions here I should mention,
Firstly that the GDP growth of a base is actually exponential. I am only guessing that it is because all my production graphs show this trend.
Secondly that the exponential growth is not an artifact of the increase in technology, but is indeed part of the growth of the base.
Thirdly, that the limiting factor is the production of the base, and not the ability of the player to place the new items. From the looks of all my production curves, up to 100k/min or so, this appears to be correct. It also seems to be the case because I appear to be "standing around" at this level waiting for more furnaces/bots/factories/modules etc.

So all in all, until the limiting factor of the base construction becomes the speed of the construction orders, additional achievements would be good to have - especially once we can run on a 16 core machine.

[Deadly-Bagel]

First off Factorio can never truly be multithreaded. SOME tasks can be multithreaded but as everything interacts with everything else, a great deal of the game must be processed synchronously.

Personally I would like to see more challenge achievements than farm fests. The Lazy Bastard achievement is brilliant, I loved getting it, many of the others however seem underwhelming such as Logistic Embargo or Steam All The Way, as these don't really have a massive impact on your game. I would like to see something like Ground Embargo where you are not allowed to place belts, forcing you to use logistic bots for everything. Achievements like this encourage you to change the way you build your base and give you better understanding of game mechanics, and they feel like an accomplishment. An achievement for completing the Supply challenge wouldn't go amiss either as that is a tricky scenario to keep up with. When you complete Lazy Bastard you think "I did it!" but when you get 20mill circuits it's more like "ugh finally" which are the types of achievements to avoid for the most part.

[BenSeidel]

First off Factorio can never truly be multithreaded. SOME tasks can be multithreaded but as everything interacts with everything else, a great deal of the game must be processed synchronously.

That is truly the funniest thing I have seen on these forums. Of course I am assuming that it was supposed to be a joke.

[orzelek]

First off Factorio can never truly be multithreaded. SOME tasks can be multithreaded but as everything interacts with everything else, a great deal of the game must be processed synchronously.

That is truly the funniest thing I have seen on these forums. Of course I am assuming that it was supposed to be a joke.

I'm not sure why you would think that it's a joke.
It's a conclusion based on the fact that some things in Factorio are hard to separate and determinism requirements are really strict which is hard to enforce in any kind of multithreaded environment. Syncing data between threads can take so much work that actual benefits are not visible.

There was an attempt to multihread inserters but last dev word on it (Rsending I think) was that it causes a lot of problems in code and no real performance benefits.

[Deadly-Bagel]

At that point I realised he's not taking it seriously so I'm not bothering anymore.

[BenSeidel]

Nope, I am deadly-serious.
Factorio is extremely parallelable because they are essentially list processing with a well defined synchronisation point of the end-of-update cycle.
Orzelek's statement that "things in Factorio are hard to separate and determinism requirements are really strict which is hard to enforce in any kind of multithreaded environment" is correct when you say that each item's update action is based upon every other actions current action. But that it an incorrect way of looking at it. If you view each item being updated based upon every other item's PREVIOUS state, that is the state that the items were at in the last updated tick, then you can see that no matter what order you update the items in, because they have a known "base-state" that is unaffected by the update process, you end up with a deterministic outcome. If you think about it in the correct way you can even get it so that if two threads update the same memory at the same time, they end up setting it to the same values. This is quite easy to do if you use tick-tock update structure. It requires a doubling of memory usage for updated values, while gaining #CPUs times more memory throughput and update speeds. While it may not seem like a good trade-off, in a machine with 2 or more free memory channels the bonus becomes more clear.

Anyway, this is extremely off topic.
If you wish to discuss this further then let me know and I will start a thread in the general forum. I am always happy to discuss asynchronous techniques with people as it's my life's work. If we can get the industry to adopt multithreading as the default we will eventually get better games and programs.

[ssilk]

Some "simple" type of parallelism I explained here: viewtopic.php?f=5&t=8670&start=10

[Leris]

So the text reads "Construct more machines using robots than manually."

However I first completely misunderstood. This is not about constructing things in hand (manually) at all, but about placing on map. So build a machine once, place it, tear down, replace elsewhere -> will count as two (2)!

Furthermore this definition of "machines" seems extremely broad, including but not limited to: Power pole, Wooden chest, every single transportation belt, etc.

Maybe rephrase or recode this?

Also, how about a new achievement (like I originally asumed here): "Construct more items using factories than manually."

Should be quite simple feat, but there are always more noobs like me