Switch to multiplicative stacking for modules

[obstinate]

Hi. I think you should switch to multiplicative stacking for modules. Two l1 efficiency modules are too good. If you used multiplicative stacking, they would give ~50% reduction, rather than 60%. Conversely, efficiency modules mixed with other module types should be better than they are. You should be able to mix efficiency and speed l1s to get a 20% speed boost at a 5% energy cost. Multiplicative stacking would fix both of these problems, and free you from the awkward artificial floor on energy cost.

Yours in automation,
Obstinate.

[ssilk]

Moved from suggestions to balancing.

I mean it can make sense: Stacking of modules is simple, but I think sometimes a good mix of modules is better.
On the other hand: Why then not just disable stacking of modules? Allow only stacking of maximal 2 modules of the same type for an assembly: Game-play-wise the same effect.

It's an idea that needs to be calculated through and be tested.

[Marconos]

Hi. I think you should switch to multiplicative stacking for modules. Two l1 efficiency modules are too good. If you used multiplicative stacking, they would give ~50% reduction, rather than 60%. Conversely, efficiency modules mixed with other module types should be better than they are. You should be able to mix efficiency and speed l1s to get a 20% speed boost at a 5% energy cost. Multiplicative stacking would fix both of these problems, and free you from the awkward artificial floor on energy cost.

Yours in automation,
Obstinate.

What is the largest size factory you have built? I know if the factories that I have setup the modules are a long ways from being unbalanced. All three module types have their uses depending on the resource types they are being used on and the type of power you are looking to use.

I'm trying to understand what mathematically makes you think they are too good?

[obstinate]

While I could go into a lot of nit-picky examples, my best example would be electric miners. If you only care about the energy floor, it's easy to hit for a converted resource cost of less than sixty units. That is a very low price to pay. It basically erases any incentive to user higher tier modules on mines.

There's also an elegance/correspondence to real world processes argument. Normally, you get diminishing returns from applying optimizations. With additive bonuses, you get better than linear improvements. Consider an electric smelter with zero, one, and two efficiency modules.

1. 0 -> 180kw
2. 1 -> 126kW (30% decrease in energy used vs. 0)
3. 2 -> 72kW (57% decrease in energy used vs. 1)

This is a very uncommon scenario in real world processes. Normally the decrease from applying the same optimization twice will be less, not more. I am not saying the game has to be realistic, but it is a side benefit to the balancing step.

[Marconos]

While I could go into a lot of nit-picky examples, my best example would be electric miners. If you only care about the energy floor, it's easy to hit for a converted resource cost of less than sixty units. That is a very low price to pay. It basically erases any incentive to user higher tier modules on mines.

There's also an elegance/correspondence to real world processes argument. Normally, you get diminishing returns from applying optimizations. With additive bonuses, you get better than linear improvements. Consider an electric smelter with zero, one, and two efficiency modules.

1. 0 -> 180kw
2. 1 -> 126kW (30% decrease in energy used vs. 0)
3. 2 -> 72kW (57% decrease in energy used vs. 1)

This is a very uncommon scenario in real world processes. Normally the decrease from applying the same optimization twice will be less, not more. I am not saying the game has to be realistic, but it is a side benefit to the balancing step.

But only looking at power is naïve. Finding resource patches that have the quantities you need can actually be much harder than power supply, again depending on how you play. In those case productivity are much better than efficiency by giving you a nice 30% more from the same patch. That increases power and decreases speed, exact inverse of the one you are saying is OP.

Power is only one part of the game, space is an issue, number of machines is an issue, length of belts/distances between bots can become an issue, scarcity of resources can be an issue. In all of these cases what is OP in once case is irrelevant in the other. To most advanced factorio players the efficiency module is the least used of all the modules because it does affect an aspect of the game that makes a difference to them. If they were so important then why are people running assemblers at 5 - 6 MW -vs- the 42 KW you can get with efficiency. Here it's definitely not about power, it's about resource consumption and production speed. It really disproves the entire point about them being too cheap as most people don't even use them as they don't give a real benefit.

Still not seeing the OP portion here.

[TuckJohn]

I do think that two of one moduals of a single type in a single assembler should apply more of an effect then if you put the moduals in two seperate assemblers.

[leadraven]

September 2015... Bump
Vote for multiplicative stacking.

[bobucles]

It's an idea that needs to be calculated through and be tested.

Was three years enough time? Doesn't look like it.

[Ranakastrasz]

Yep. The question is if we have enough API now to have a mod that does this manually. If so, someone make a mod, and we can test it.

[leadraven]

I forgot to mention one very important thing : multiplicative stacking will be absolutely broken with beacons. Beacons should be removed.

[Ranakastrasz]

I forgot to mention one very important thing : multiplicative stacking will be absolutely broken with beacons. Beacons should be removed.

Eh, sort of. I would point out that there are some serious downsides to beacons even with multiplicative stacking.

Note that power cost will be effected. Take the supercharged oil well. 12 beacons with 2 speed module 3 and 2 in the oil well.

The overall effect is 24 instances of +25% speed and +35% power, and 2 extra 50% speed for 70% power.

Currently, it goes from a yield of 2 for 90kW, to a yield of 16 for 972kW. The slight loss of efficiency is preserved, no matter how many you add, approaching an extra 40% power consumption, capping at 60.75kj/oil, an increase of 35% exactly
However, if it is scaled instead... Well, it gets kinda ridiculous

It produces 17.2 at 8 modules (3 beacons) for 1575kW, around a 50% increase in power consumption, albiet with less infrastructure.
At the maximum of 26 modules though, it outputs 953 oil, for 349gW of power. That is a LOT of power. More specifically, around 390 steam engines, or 72% of a 2x2 reactor. Admittedly the yield is also ridiculous, but, well. At no modules, its 1 oil per 45kW, and at all of the modules ever, its 366kW per oil, or a markup of 713%.
Oh, and don't forget the pollution, which is actually a factor now.

Whether this is balanced or not, I think it isn't, but exponents get big, very fast. And as long as power consumption goes up faster than speed, well....

Anyway, this ignores efficiency modules, and so on. Beacons probably would need tweaked, but I expect outrage if they were ever removed.
Maybe add extra diminishing returns. Good stuff diminishes, bad stuff doesn't, so you get +25% speed but still spend 70% power each time. But negative power is efficiency, so behavior would be inconsistent.

Efficiency modules are impressive If you replace 6 speeds with efficiency in beacons, it drops from 950 to 250 oil, and power cost drops from 349gW to 10.2gW. So loss of around 75% production speed for a 97% reduction of energy consumption. Better? Worse? Dunno. But still want to experiment with it.

TL;DR Exponential increases are insane, and beacons don't let you dodge the power cost.

[bobucles]

it outputs 953 oil, for 349gW of power.

Did you say Megawatts or Gigawatts?

[Koub]

Just redid the math, It's 349.6 MW for 952.9 output