How to evaluate pollution?

[SilverShadow]

Using several assumptions I can bring raw resources and energy to one equivalent and then calculate all costs in it. Which makes it very handy to compare all sorts of investments. However, I don't know how to evaluate pollution. Currently I just ignore it in my calculations, but I am not happy with that.

Any suggestions?

[bobucles]

Pollution generates biters. Biters cost ammo, repair packs and losses in damage.

https://wiki.factorio.com/Pollution

Pollution is absorbed by trees and terrain. The initial cost of ammo is pretty high, but higher tech gradually brings that cost down to near nothing.

[SilverShadow]

Pollution generates biters. Biters cost ammo, repair packs and losses in damage.

https://wiki.factorio.com/Pollution

Pollution is absorbed by trees and terrain. The initial cost of ammo is pretty high, but higher tech gradually brings that cost down to near nothing.

Yes, I know all of that. The question is how to put all that knowledge into the formula.
I want to make something like this viewtopic.php?t=5705 to efficiency modules and solar panels

[BlueTemplar]

This is going to be very hard, as what the biters cost you depends on a lot of factors !

I guess it's still not extremely hard to calculate the cost spent on killing them (rather than the cost of what they have destroyed, or the cost of distracting you),
but you will still have to do big simplifications about assumptions of trees and terrain...

[SilverShadow]

Ok, lets start with simpler question. Lets say we have one electric mining drill, one assembling machine 2 and one efficiency module

Electric mining drill has
Energy consumption - 90 kW
Pollution - 9

Assembling machine 2 has
Energy consumption - 150 kW
Pollution - 2.4

Where to put efficiency module?

[Edit] Note that pollution that building produce is relative to its energy consumption, so efficiency module directly reduce pollution in addition to indirectly lowering pollution from boiler

[theolderbeholder]

I think there was something about pollution and its spread in one old FFF. Or maybe it was the wiki? Can´t remember. In either case, you have to consider energy is "free" after setting up a stable nuclear reactor or going fully solar, as neither cause direct pollution like boilers.
Which does not really help with your question, actually I just made it worse

[eradicator]

@SilverShadow:
Modules reduce energy consumption (==pollution) on a percentage basis, so you put it into the machine with the highest absolute pollution. The absolute energy consumption does not matter for this (until you start calculating energy cost anyway). But you have to consider that you can't reduce consumption/pollution below 20%.

As for a generic "pollution cost" that's impossible to calculate on a per-recipe basis as it depends on the actual factory layout. If your machines are one chunk away from a biter nest then any amount of pollution will attract a constant stream of biters, while if you have already cleared a 50 chunk radius around the machine the pollution per-recipe has no cost at all.

[SilverShadow]

It seems that it is actually possible to evaluate pollution and it is quite directly linked to resources. Miners and furnaces are the key elements of every base and the main contributors to pollution. Basically, you are paying for all of the resources with pollution.
Pollution "cost" of the resource differs with the setup and I've done some math to get to know what it would cost at different stages of the game:

Burner mining drill, Stone furnace
50.78 pollution per plate

Electric mining drill, Stone furnace, Steam power
25.62

Electric mining drill, Steel furnace, Steam power, Mining productivity 3 (research) +6%
23.90

Electric mining drill, Steel furnace, Steam power, Mining productivity 7 (research) +14%,
2*Efficiency module in miners
13.13

Electric mining drill, Electric furnace, Free power, Mining productivity 15 (research) +30%,
3*Efficiency module in miners
4.21

Electric mining drill, Electric furnace, Free power, Mining productivity 15 (research) +30%,
3*Efficiency module in miners, 2*Productivity module 3 in furnaces, 2*Speed Module 3 in 8 beacons
4.95

What I find worth notice:
• Placing efficiency modules in miners is a good way to reduce pollution
• Switching to steel furnaces or solar isn't helping with pollution that much
• In many cases placing productivity modules in assemblers does reduce overall pollution. That would increase pollution made by assembler itself and boiler, but reduce pollution made by miners and furnaces
• In some cases placing productivity modules in assemblers does reduce overall pollution more than placing efficiency modules

[bobucles]

That's some intriguing math, silvershadow! I think it also helps explain why early game is so much harder when using this pollution chart:

Pollution Type
200 Small biter
1000 Medium biter
4000 Big biter
20000 Behemoth biter
200 Small spitter
600 Medium spitter
1500 Big spitter
10000 Behemoth spitter

For the cost of ~200 pollution you can generate one stack of yellow ammo, and you lose a lot of that shooting a small biter. Meanwhile the larger biters take vastly more pollution to spawn yet the overall pollution you need to generate extra firepower is only decreasing. It leaves late game waves feeling a bit underwhelming and more like backdrop noise rather than a true threat.

At the same time you need to tech up to the more efficient weapons so that you aren't burning 6+ red ammo stacks on each big biter. That's a losing game.

[4xel]

If you got a low pollution strategy, biters basically are not an issue, you evolve faster than them and they seldom ever attack you, and it's good practice for this strategy to secure a very large territory around your factory. Effectively, what I mean by low pollution strategy is when no pollution is absorbed by biter nests.

If you opt for a high pollution strategy, you can reason at the margin : any excess pollution will generate biters, proportional to pollution you emit. Some pollution will be absorbed by trees and grass, but it does not matter ; if you intend to pollute more by an amount of X, that's X pollution worth of biters more to kill.

The table

Pollution Type
200 Small biter
1000 Medium biter
4000 Big biter
20000 Behemoth biter
200 Small spitter
600 Medium spitter
1500 Big spitter
10000 Behemoth spitter

given in https://wiki.factorio.com/Pollution#Native_life, along with proportions depending on your evolution factor (https://wiki.factorio.com/Enemies#Spawn ... ion_factor) allows you to translate pollution into biters and spitters.

Using the stats of your ennemies (https://wiki.factorio.com/Enemies) and the stats of you turrets (https://wiki.factorio.com/Turret) or landmines and the damage formula (https://wiki.factorio.com/Damage) allows you to translate biters and spitters into ammunition and energy cost, depending on your defense strategy, which can directly be translated into ressources upkeep and infrastructure cost.

Area of effect and mixed defense are harder to evaluate, it should be possible to link it to the proportions of said mixed defense, but ranges difference and rock paper scissor factor changes the result, usually in favor of mixed defense (laser turret may be able to target spitter, gun turret will tend to focus on heavilly armored bitters). Repair costs may also be tricky to evaluate.

Circuitery can be used to track any upkeep cost of each outpost or wall segment, from the energy drain of your turrets down to the repair packs used by bots, and can be compared against our predictions, to help refine it.


It should be noted that pollution has a double cost : the biter it directly generates, and the raise of the evolution factor. For the early stage of the game, I don't think the computations above may replace trial and error or even help at all, a more polluting strategy also enabling a faster progression in a non linear way. But if you are wondering whether you can spam production3 and speed3 beacons or whether the pollution cost is to high, your factory and defense infrastructure should be already optimized for the endgame and you can assume an evolution factor of 1 then.

On the benefits of expansion. Any chunk that you put between your factories and the biters will absorb pollution over time, thus reducing the pollution absorbed by nest and the number of biters spawned. Bonus is said chunk has trees. The cost of expansion might be very hard to evaluate (maybe not so much if assuming artillery rather than complex non automated fight involving actual actions and effort from the player), but evaluating the benefits is straightforward assuming one can compute the cost of pollution.

TL;DR :
Factory --> pollution --> biters --> damage points needed --> amunitions --> cost, and there's a formula for each steps, albeit not a simple one for steps 2 and 4, and one very dependant on the specifics of your factory for steps 4 and 5.

[SilverShadow]

Factory --> pollution --> biters --> damage points needed --> amunitions --> cost, and there's a formula for each steps, albeit not a simple one for steps 2 and 4, and one very dependant on the specifics of your factory for steps 4 and 5.

That's fun because I had calculated damage per resource for firearm magazine, piercing rounds magazine and laser against each biter previously. I think I should share that too. It's the best to post it under General discussion thread?

[4xel]

Factory --> pollution --> biters --> damage points needed --> amunitions --> cost, and there's a formula for each steps, albeit not a simple one for steps 2 and 4, and one very dependant on the specifics of your factory for steps 4 and 5.

That's fun because I had calculated damage per resource for firearm magazine, piercing rounds magazine and laser against each biter previously. I think I should share that too. It's the best to post it under General discussion thread?

By all mean go ahead. I had a previous thread like that in mind but could not find it again (was it yours). IIRC this thread was about the part "damage points needed --> amunitions --> cost" for laser turrets and most gun turret configurations, and converting the cost in electricity.

I'm not sure of the fairness of converting the cost to electricity though. Iron and copper ores also cost the land ou need to clear to build new mines, not just electricity. In any case, the "cost" part of the equation will be very tricky and will depend on what you want to optimize. Laser turrets along solar panels are upfront cost, while gun turrets are upkeep cost, hard to compare. Unless you plan on building an exponentially or geometrically ever growing factory, I'd say reducing upkeep costs is more important though.

[BlueTemplar]

Yeah, for my part, in my first drafts I'm instead neglecting energy costs...
(And assuming that 1 iron = 1 copper = 1 ₣)
- but it's also because I haven't even tried Laser Turrets yet !

factorio_graph.png (89.14 KiB) Viewed 9287 times

(P.S.: Alien stats taken from *vanilla*-based wiki - bobenemies might change those...)

[bobucles]

Evaluating the cost of gun or laser ammo vs. biters is fairly straight forward. Evaluating the cost of oil for flame turrets is much more difficult. Flame turrets deal many forms of damage, cause a devastating fire debuff and have a large AoE effect that makes them even more efficient. I had done some math on it a while ago but lost the numbers. This was my takeaway on ammo efficiency:

- The efficiency of ammo is awful. Simply awful. It is in fact SO bad that creating ammo factories on death worlds and filling turrets with 200 stacks may very well be considered griefing. It's THAT bad.

- The efficiency of grenades are bad, but not anywhere near as bad as ammo and coal is much easier to have in spare supply. If you can wipe out a cluster of small biters it's done a good job.

- Shotgun ammo has decent efficiency. It's nothing stellar but it won't wipe out your resources either.

- Laser turret efficiency is VERY high, at least an order of magnitude better than ammo.

- Flame thrower oil is absurdly good and may in fact be even better than lasers. 90% of the damage is hidden behind the fire debuff so the secret is to tap biters with tiny sprays and they simply die. Flame turrets use so little ammo that it hardly matters.

- Land mines are pretty good and are absolutely superior to grenades.

- Cluster grenades are like grenades except more expensive and scatter their damage everywhere. Pretty bad.

- Combat drones are incredibly ammo efficient over their duration but for some reason players don't like them. Perhaps the range is too low or the tech feels too far out of the way. Basic drones definitely suffer due to their 5 base damage when the current level of biters have 4 base armor. Drones need to be used in a critical mass or they die before being useful.

- Repair packs > just about everything. It's far cheaper to repair a wall against a behemoth than to use bullets, at least. You can even keep them as a pet since the repairs are cheap enough.

[4xel]

- The efficiency of ammo is awful. Simply awful. It is in fact SO bad that creating ammo factories on death worlds and filling turrets with 200 stacks may very well be considered griefing. It's THAT bad.

- Laser turret efficiency is VERY high, at least an order of magnitude better than ammo.

And setting up enough laser turrets and power generation for them to do the job of a gun turret cost at least one order of magnitude more than setting up a gun turret. A gun turret is basically free, and by the time you can drop your first laser turret, gun turrets have more DPS than late game laser Turret, even against heavily armored turret, meaning you need less of them, and they help minimizing damages in case of big waves.

Gun turret is basically no upfront cost. Laser turret is basically no upkeep cost (assuming their huge electricity drain is covered by the upfront cost of solar panels).

For places with a lot of combats, laser turret will quickely pay for themselves and for the underlying power grid.

For weponizing your inner towns just in case, or when you play the low pollution game, laser turrets are horrendously wastefull, all you need is a cheap gun turret which seldom ever fires.

- The efficiency of grenades are bad, but not anywhere near as bad as ammo and coal is much easier to have in spare supply. If you can wipe out a cluster of small biters it's done a good job.

Grenade is a very efficient way to clear nests with low tech. True, they are not as efficient as combat drones, but low tech being less efficient than high tech should not be surprising. The ability to use low tech earlier makes up for that.

The low tech argument also applies to gun turrets too. In any serious marathon death world like setting, you have automatized gun turrets wall before you getting anywhere near the much beloved solar panels and laser turrets.

The low tech argument applies to shotgun too.

- Combat drones are incredibly ammo efficient over their duration but for some reason players don't like them. Perhaps the range is too low or the tech feels too far out of the way. Basic drones definitely suffer due to their 5 base damage when the current level of biters have 4 base armor. Drones need to be used in a critical mass or they die before being useful.

The few youtubers I follow actually use them a lot, and they look efficient and convenient. What do you mean by "amo efficiency"? do you include the cost of the expandable robot itself?

They are very endtier though if I'm not mistaken, not necessarilly a surprise they are not muched used in multiplayer. Also, there are a lot of other methods to clean nest which work just as well even late game, are lower tech and are barely less convenient or fun, so I guess people might pick them and not bother changing when finally acquiring tech. Like turret jump (assisted with construction bots) or artillery.

[bobucles]

What do you mean by "ammo efficiency"?

One magazine goes in, and 45 seconds of pewpew comes out. Combat bots make very good use of the ammo you use to make them. I don't use them nearly enough, and I don't see them used online half as much as they should be. If you want a quick practice with them try the First Steps 3 mission. It gives you a small stack of tier 1 drones and a few nests to try them on. It makes the final section a breeze.

They are very endtier though if I'm not mistaken

The basic capsule is fight+ military 2, so it's available pretty damn early. The basic recipe is 1 red ammo, 3 gears and 2 green circuits. It's not expensive at all for what you get and it even benefits from bullet research. It's the later versions that end up scaling extremely high in terms of cost and assembly time. Chances are players ignore the first tier, look at the end tier and say "too expensive" and skip them all. I can't really blame them because what good is a tech you can't take to the very end?

[4xel]

I'll definitely give them a try then.

I think reaching the critical mass is also a big issue. I've played some few RTS so now that you say it, I totally get what you mean by that. But trying a new weapon in factorio, my first thought is to use very few ammunition. Except combat bots are not ammunitions, they are troops following Lanchester's square law (https://en.wikipedia.org/wiki/Lancheste ... square_law).

People trying bots in very small quantities and seeing them die easily sounds like a very compelling explanation as to why they may be underrated.

[BlueTemplar]

You cannot use bots in "large quantities", as when you first research them you can only have 1 at once.

And the research is pretty damn expensive, at 150 MilSci, and with a lot of pre-requisites (a lot of which require oil to be useful).

So by that point, on DeathWorld, you'll be likely facing Blue Biters (with their 8 armor),
and the mk1 robots (with their 5+upgrades "regular" ammo damage) are almost completely useless against them
(you'll probably even be wasting resources by converting an APmag to the robot !)
- you'll be better off rushing (about as expensive?) capsule and tank research !

(Unless you can't get to oil for blue science, but if you can't, mk1 robots probably will help less than basic grenades.)

[Koub]

Hasn't this gone a little offtopic ? There's a topic not far from here about combat robot count. Shall I split and merge the posts into that thread ?

[bobucles]

Combat bots are weapons, and all weapons play into evaluating the cost of pollution. There's just a little bit more about them to understand than at first glance.