How to fix module balance

[Hedning1390]

The real truth is that the problem can't be fixed without digging all the way to the bottom, dredging up the heart of the problem, fixing it at the source, and then re-building it up again from there. And the longer we avoid this uneasy truth, the harder it will be.

Productivity does so much more than just soaking up power. It rebalances the factory so that late game your factory isn't 90% boring furnaces and 10% interesting chains. Power isn't an issue. Balance of production is. I think it is cool that early game you focus more on digging and smelting, and later you focus more on the high tech stuff.

[thereaverofdarkness]

I think allowing prod modules on more products doesn't make much sense.
-Things like kovarex is balanced around getting only 1 extra. The 40 required for a cycle would turn from a negative (takes long time to get going) to a positive (high base number to multiply).
-On things like barrels you can get items by putting them in a loop which I disagree with.

Of course there should be some things that don't allow productivity, such as Kovarex Enrichment or barrels.

-On structures such as inserters it really doesn't matter so why change it?

It wouldn't be a problem to allow it, so why deny it? If a restriction has no justification, it should be removed. You don't need to find a justification for taking advantage of the removal of the restriction. Better to change it before you need it changed.

Efficiency needs more of a buff. I think eff3 should have something like 5000% to make sure a single module, even in a beacon, will put the machine on it's lowest consumption level.

Then why would you ever use more than one efficiency module?


Here's an idea:

Productivity speed reduction also reduces speed bonus from speed modules by half as much. Energy cost increase also increases energy cost bonus on efficiency by half as much.

So if you use 5 productivity 3:
-75% speed, -37.5% speed module speed bonus, +100% energy, +50% energy cost reduction on efficiency modules.

Add in 5 speed 3 and you get:
+156.25% speed, +375% energy, +187.5% energy cost reduction on efficiency modules.

At this point, our total speed is +81.25% and our total energy cost is +475%. With our adjustments, an efficiency module 3 will reduce energy cost by 253.125%. With 2 of them, our energy cost would be -31.25%. Our productivity is +50% and our rate of output is +171.875%.

Comparison with 10 speed modules:
We get +500% speed and +750% energy cost. Efficiency module energy cost reduction increased by +375%. Adding one efficiency 3 module would reduce energy cost by 356.25%. With two of them, final energy cost is +37.5%, total speed is +500%, and total rate of output is +500%.

5 prod/5 spd/2 eff
Productivity: 150%
Output rate: 272%
Energy cost: 69%

10 spd/2 eff
Productivity: 100%
Output rate: 600%
Energy cost: 138%

[Hedning1390]

Then why would you ever use more than one efficiency module?

You wouldn't. That's the point. The bonus from efficiency is not worth sacrificing several slots over, even at the maximum bonus. It's better to just build more power. To buff it further you need to either give it new abilities or reduce the price. The price is the number of slots you need, so to buff it you need to reduce the number of slots it needs. The lowest number of slots it could need is one beacon slot. If that's the price to get your machine from +500% (or whatever) down to -80% then it could be worth it.

I would caution you against making power important in order to make efficiency modules more valuable. Factorio is about logistics and power have infinite throughput through cables, hence there is no factorio challenge in massively expanding your power. It can easily turn into a tedious chore.

[thereaverofdarkness]

You wouldn't. That's the point. The bonus from efficiency is not worth sacrificing several slots over, even at the maximum bonus. It's better to just build more power.

I keep saying this, and you guys don't seem to be hearing it. Power generation in Factorio is overpowered. The reason there is a low ceiling on effective value of an efficiency module has nothing to do with how much power it saves you and everything to do with how cheap it is to get more power.

[Zavian]

But from a game design perspective it's not about whether its overpowered, nor is it about whether the balance is realistic, it about whether it's fun. A 1k science per minute base needs over 96,000 solar panels. It is already bad enough laying out enough blueprints for 100k solar panels. Nerfing solar panels so that megabase players need to build 10 times as many solar panels would not be an improvement. Neither steam nor nuclear are viable alternatives, for many megabase players since they cannot compete with solar's ups efficiency. (And if you are trying to build the biggest base you can, then ups is a significant consideration for every design decisions).

[thereaverofdarkness]

And why should we ignore balance for most players in order to balance UPS for megabases? Why don't we instead focus on the UPS cost of other power generation methods and suggest ways to optimize it? Clearly megabases need a power source with no moving parts, but it doesn't need to have unlimited use. There could be a new one added which uses a finite resource to generate a large amount of power without moving parts. But this aside merely serves to indicate why it's not important to use UPS in megabases as an excuse for overpowered solar panels. I'll consider further discussion of megabases and UPS to be off topic unless it relates to modules or some other part of the topic.

[Hedning1390]

Disregarding ups the argument still holds. There is no new challenge in centupling your power. It's literally just stamping down more blueprints, unlike other production which provides new logistics challenges at different levels. All nerfing power would do is create a chore while start breaking down the engine, because construction bots has a rather small limit on number of simultaneous construction orders.

[Zavian]

The main users of modules are megabases. If you aren't playing on after launching a rocket, then typically you just need a few modules for the silo. (Some players use modules in labs or other places, but others don't).

To fix the ups problems will require a fluid mechanics overhaul, and possible a similar optimisation of heatpipes. It might be coming for 0.17. (It's listed as a low priority item for 0.17). At that time nuclear might be more attractive for megbases than it currently is.

But regardless of whether a fluid mechanics overhaul is coming in 0.17, improving gameplay should trump realism when talking about balance changes. (Otherwise the first thing to fix is lugging multiple locomotives and a dozen rail wagons in our backpacks).

(Also what Hedning1390 said. but then I've already tried to make similar points in this thread).

[Frightning]

Regarding power situation, a better solution instead of nerfing power generation would be to increase energy usage of most buildings. Now at face value that might seem like 6 one way, half a dozen the other, but it's not because now the energy savings of Eff modules compares MUCH more favorably with the fixed cost of producing that same amount of power with whatever method you prefer to use (ditto for running costs with Steam/nuclear, too).

[thereaverofdarkness]

Disregarding ups the argument still holds. There is no new challenge in centupling your power. It's literally just stamping down more blueprints, unlike other production which provides new logistics challenges at different levels. All nerfing power would do is create a chore while start breaking down the engine, because construction bots has a rather small limit on number of simultaneous construction orders.

The problem with solar power is that it's one of the most potent forms of power generation. Given that it is free power forever, it should be the weakest form. If solar panel output were greatly reduced, and solar panel cost reduced somewhat to compensate, it would have a drawback: the space it takes up. Steam power has a drawback: the fuel it takes to run. Nuclear power has a drawback: the logistics balance as well as the cost of nuclear fuel.

Regarding power situation, a better solution instead of nerfing power generation would be to increase energy usage of most buildings. Now at face value that might seem like 6 one way, half a dozen the other, but it's not because now the energy savings of Eff modules compares MUCH more favorably with the fixed cost of producing that same amount of power with whatever method you prefer to use (ditto for running costs with Steam/nuclear, too).

I wouldn't propose simply reducing power output OR increasing power consumption, and besides they are essentially the same because they still affect the usefulness of efficiency modules the same way.


Each source needs to have its own solution offered:

Solar panels have a good material cost but take up too little space. Cut their output to 1/5th and cut their material cost lower as well. Now they are limited by space, or the time you care to put into clearing land and laying them down. As for clearing land, that's too easy because bots need to be nerfed. The whole thing is tedious in large part due to how easy it is. When something is more difficult, your mind responds differently to it and is less likely to think of it as tedious, even though it actually takes more time and energy for the same result. This is because you're busier while doing it, and also when we perform a cost analysis of a task, we often ignore time spent moving from one location to the next, time spent planning, and other similar "free moves". When the task is primarily comprised of time spent on cost-moves rather than free-moves, we tend to estimate its cost to value ratio more accurately and are less likely to get a ways into it and think "man this is boring".

Coal and oil need to be more limited. Coal shouldn't be uncommon, but I think maybe it should have a lower energy value. The player shouldn't wind up with loads of coal they don't know what to do with when they're running a steam-heavy power setup. But I find at default coal setting, I can't seem to run out of the stuff. And then oil further exacerbates the issue because solid fuel is so cheap and potent. Perhaps the largest problem is the tremendous number of oil patches per field. Since they never run out, you can eventually wind up with a never-ending fuel supply capable of running power sources indefinitely. If my math is correct, counting processing in the refinery, and only using the natural light oil output, it costs roughly 3.5 MJ to get 1 solid fuel (25 MJ) if everything has -80% power consumption. Not counting inserters or logistic bots, that's ~107 kW of steam engine power per pumpjack, with all the heavy oil and petroleum gas left over for other uses. The problem I see is less how much it is and more how many oil patches you get. I'd fix oil by making individual oil fields smaller, and maybe consider reducing the minimum yield further as well.

Nuclear needs to be more limited. It's not so much the Kovarex enrichment process which is the problem here, as nuclear fuel cells need 19 times as much U-238. The real problem is how fantastically abundant uranium ore is. It needs to be far less common. Most games you shouldn't have any uranium in your starting area, but I usually find 1-3 patches inside it and 3-5 more within radar scan range of the origin point. I've run counts a few times, and I consistently find uranium second most commonly out of any type of ore patch, only falling short of stone. Oddly, iron is the least common. There was a FFF recently which suggested they fixed the problem by making iron override other ores instead of previously being overrode by them, but in my experience iron is just plain less common without having anything spawn on top of it. So fix ore patch abundance into something reasonable, and make uranium both small and uncommon. But it could also help to make uranium fuel run out faster. I'm not sure if that needs tweaking because I haven't yet used a lot of nuclear power, but I get the feeling that if you aren't spending your uranium on much else, you'll get way more than you can ever use on fuel cells. That should not be the case.


With everything finite, you'll find that in order to keep your steam and nuclear power running long-term, you must continually find new and more distant sources of coal, oil, or uranium. As richness increases with distance from the origin, you'll gradually have each source lasting longer and longer. But if you don't like that, you can circumvent the problem of limited supply by building lots of solar power. If you try to run a power-hungry base on solar power and you find it difficult, then that is as it should be. If you make the choice to control your power consumption at the cost of your output rate, then you should have no difficulty running your base on solar. And before you say 1/5th output is too low for solar panels, that's about what I'm using with Realistic Power mod, and I do fine with solar only when I put efficiency 1 modules on everything.

Also, this ties into my suggestion to reduce the energy cost increase of productivity modules. If they appropriately reduced your net output speed, then that speed loss is already the correct trade-off for the improved resource efficiency. They should then have a small increase in energy cost, and a player should be able to easily counter it by mixing in some efficiency modules. That low rate of output setup with high resource efficiency is perfect for players who don't like to build a new base very often. It offers a solution to their problem which comes with an opportunity cost.

[Zavian]

But if you don't like that, you can circumvent the problem of limited supply by building lots of solar power. If you try to run a power-hungry base on solar power and you find it difficult, then that is as it should be. If you make the choice to control your power consumption at the cost of your output rate, then you should have no difficulty running your base on solar. And before you say 1/5th output is too low for solar panels, that's about what I'm using with Realistic Power mod, and I do fine with solar only when I put efficiency 1 modules on everything.

Space itself is cheap. It's not difficult to place 100,000 solar panels, Assuming you have decent sized landmasses, it's not hard to find the space for them for any base that needs that much power. Bots do the clearing and building, but as we have already said it's tedious to align and place 100 solar blueprints. That means it is not interesting or engaging from a gameplay perspective. Having to place 5 times as many blueprints is 5 times as bad. You practically admit that making solar tedious to use is you intent,so that solar will be less attractive to players, and they will choose another option for power. Yet solar will always win when it comes to ups efficiency. (Even if fluid mechanics were free, the inserters, reactors, boilers, turbines etc all require individual processing, so neither nuclear or steam can ever equal solar's ups' efficiency).

Now onto the choice of adding efficiency modules. Are you also suggesting that all assemblers get an extra module slot, limited to only accept efficiency modules so we can add 4 prod modules, and 1 efficiency module? If you are suggesting we should use efficiency in beacons then they need an extra dedicated slot as well. Because the energy savings aren't going to be worth losing that extra 10% productivity, and needing to add more assemblers and miners. Beacons would be worse, replacing a speed module with an efficiency module would almost halve the effective speed of assemblers lvl 3s. (From speed 5.5 down to 3) That alone almost doubles the number assemblers and inserters you need for a given output. Effectively almost doubling the number of active assemblers and inserters could almost half ups for some bases. Anyone who cares about ups would say, "I'm not adding efficiency modules at that cost", I'll just build the solar.

If you really want to nerf solar then just double the cost of panels. The people who you might be able to convince to use something else instead of solar will care about that.

[thereaverofdarkness]

Are you also suggesting that all assemblers get an extra module slot, limited to only accept efficiency modules so we can add 4 prod modules, and 1 efficiency module? If you are suggesting we should use efficiency in beacons then they need an extra dedicated slot as well. Because the energy savings aren't going to be worth losing that extra 10% productivity,

Have you read nothing in this thread? The entire point is to balance the modules. If the modules are balanced, there will be a reasonable trade-off between productivity and efficiency. I'm discussing ways to fix power production because as I have explained earlier, efficiency currently cannot be a good module when power production is too easy.

[Hedning1390]

Solar panels have a good material cost but take up too little space. Cut their output to 1/5th and cut their material cost lower as well.

I think you vastly underestimate either the power need of late game bases or how much power they actually produce. They already do take up a massive amount of space.

Here's my solar farm. It produces 8.7 GW. I need 4 times this for my finished base. With your suggestion I would need 20 times this.
At the top is a 2.4GW nuclear power plant. You can barely see it in the picture because it is so small in comparison, but it's not small. It's large enough to fill a screen at max zoom out.
https://i.imgur.com/Yk2USNV.png

You can get a size reference from the radars.
When I'm expanding it my other construction stalls, even my personal bots, because the solar expansion is eating up all construction orders.
What's the difficulty here? Why is making me build 5x as much not just a tedium? Like farming xp in a bad rpg.

Space itself is cheap. It's not difficult to place 100,000 solar panels, Assuming you have decent sized landmasses, it's not hard to find the space for them for any base that needs that much power.

Wrong. Have you actually tried it? There's both cliffs and water in the way.
Here's a base where the current solar farm was tediously laid down by using cliff explosives. It took quite a bit of exploration to find the cliff and water free landmass in the top right corner (where solar expansion has started) which is large enough. This is pure panels with no accumulators, but it still needs help from nuclear during the day. It's going to be rectified, but it took until this point to get it going. https://i.imgur.com/CAWgZPe.png

[bobucles]

Of course there should be some things that don't allow productivity, such as Kovarex Enrichment or barrels.

Enrichment would be perfectly fine if the 40 good Uranium wasn't duplicated. That's all it takes to fix the recipe for prod modules.

Nuclear power has a drawback: the logistics balance as well as the cost of nuclear fuel.

Those aren't drawbacks. Providing a stack of fuel every hour is the easiest thing in the world. You can even do it on foot. Everything about nuclear fuel can be done on foot without a single ounce of logistics.

The drawbacks of nuclear power are learning the heat system and the high tech barrier. The former adds a considerably difficulty curve for new players and the latter is a hard gating that pushes nuke power to the mid-late game.

Clearing cliffs is tedious

We need cliff explosive artillery.

Buffing efficiency modules

I doubt you'll ever find a realistic number that works. The reason productivity works is because every single thing with module slots is an item producer. All item producers have the exact same goal, produce as much as possible. Whatever is best for one producer is the best for every other producer. There are no module guns, radars, steam engines or anything else that might have different module priorities.

Of course the current module scaling is pretty atrocious. 30/40/50% is not good scaling. It doesn't even double, which 4/6/10 prod and 20/30/50 speed manages to do. Eff modules absolutely positively need a buff. Maybe 30/50/100%? It's still not good scaling, but it's not as terrible as before.

Nerfing energy generation

Can't say I'm a fan. Big bases need big power and players will choose whatever works best. There aren't many ways to make a system that both has moving parts, is fun to put together and is friendly on UPS.

[thereaverofdarkness]

I think you vastly underestimate either the power need of late game bases or how much power they actually produce. They already do take up a massive amount of space.

Here's my solar farm. It produces 8.7 GW. I need 4 times this for my finished base. With your suggestion I would need 20 times this.

No, you're locked into this idea that there's a certain amount of energy it costs to run your factory, because in vanilla that's how it is. I'm advocating for making the energy costs flexible.

Those aren't drawbacks. Providing a stack of fuel every hour is the easiest thing in the world. You can even do it on foot. Everything about nuclear fuel can be done on foot without a single ounce of logistics.

The drawbacks of nuclear power are learning the heat system and the high tech barrier. The former adds a considerably difficulty curve for new players and the latter is a hard gating that pushes nuke power to the mid-late game.

I agree with this. It seems mostly alright to me, but I don't have a solution for the problems.

I doubt you'll ever find a realistic number that works.

I already did, multiple times. My favorite is the second comment down on page 2 of this thread. Individual numbers can be tweaked from there, but I've already suggested mechanics changes which can fix the inherent underlying problem.

[Frightning]

Why go w/ half measures, I'd say in general, make subtractive effects apply separately from their additive counterparts (with a cap of -80%), then not only is speed penalty on prod modules not countered by speed beacons, but now efficiency modules can potentially save you a LOT of energy.

So the new formula for final effect would be:
(base)*(1+positive effects)*(1-negative effects)
where negative effects is not allowed to exceed 80%=0.8.

[thereaverofdarkness]

Why go w/ half measures, I'd say in general, make subtractive effects apply separately from their additive counterparts (with a cap of -80%), then not only is speed penalty on prod modules not countered by speed beacons, but now efficiency modules can potentially save you a LOT of energy.

So the new formula for final effect would be:
(base)*(1+positive effects)*(1-negative effects)
where negative effects is not allowed to exceed 80%=0.8.

Good point. We were discussing this earlier. I also came up with the idea to have the penalty have partial effect on the bonus, so that the modules have a reduced constructive effect with each other, without eliminating it entirely. I should update the OP with this info, but I was hoping to discuss it further first, to get a better idea of what values would be good. I'm fine with simply applying bonus and penalty cumulatively with each other, but I think it's possible to achieve something even better by allowing some constructive interference between modules.

[BlakeMW]

I'm a proponent of separately multiplicative effects, it might make eff modules good, but it probably wouldn't do much to make eff3 modules good, since the eff3 module is such an minor advantage over eff2 and eff1: the maths would probably make it favorable to just shove in an eff1 module and build a bunch of solar/accu, the difference between eff3 and eff1 is too marginal to justify the more expensive module.

One idea I had, was to make eff modules even more effective at reducing pollution especially at higher tiers, and a little less effective at reducing energy usage, a scheme might look like this:

eff1: -20% energy usage, -5% pollution
eff2: -30% energy usage, -15% pollution
eff3: -40% energy usage, -30% pollution

Since energy usage and pollution are applied separately, 2x Eff3 modules would reduce total pollution to just 8%.

My reasoning is that without making power crazy expensive, or making the modules crazy cheap, eff modules will never actually be good for a power-gamer, more power and more guns is always going to be more cost-effective than being ecologically friendly. But they can at least be "flavorsome", so using them makes a real difference to your factory. If adding a few Eff3 beacons makes your pollution nearly go away some players would consider it worth doing, after all not everything has to cater to power gamers, as long as it caters to some legitimate playstyle. Note that this isn't dependent on the multiplicative thing, because pollution is already applied separately to energy usage, so even with the current scheme, if Eff3 modules had a hefty minus pollution they would do a great job of reducing the pollution output of prod3+speed3 setups. Technically this would also be accomplishable by having like -400% energy usage on eff3 modules, but I don't like number inflation like that, a -400% effect makes the system look broken.

[Hedning1390]

No, you're locked into this idea that there's a certain amount of energy it costs to run your factory, because in vanilla that's how it is. I'm advocating for making the energy costs flexible.

It's already flexible. You only suggested a buff to efficiency modules. It changes nothing. If I can place a ton more solar panels to get slightly more productivity out of my factory that's a tedium I'd have to suffer through. The question shouldn't be "is improving the efficiency of my factory too boring to be worth it?" It should be "is this fun option more efficient than this other fun option?".

[thereaverofdarkness]

One idea I had, was to make eff modules even more effective at reducing pollution especially at higher tiers, and a little less effective at reducing energy usage, a scheme might look like this:

eff1: -20% energy usage, -5% pollution
eff2: -30% energy usage, -15% pollution
eff3: -40% energy usage, -30% pollution

Since energy usage and pollution are applied separately, 2x Eff3 modules would reduce total pollution to just 8%.

I like it.

It's already flexible. You only suggested a buff to efficiency modules. It changes nothing. If I can place a ton more solar panels to get slightly more productivity out of my factory that's a tedium I'd have to suffer through. The question shouldn't be "is improving the efficiency of my factory too boring to be worth it?" It should be "is this fun option more efficient than this other fun option?".

Once again you're using vanilla to base your argument. In my suggestion, rate of output is no longer linked to material efficiency.

I edited the OP, added details to the bottom.