December 19, 2013

Question From A Newbie: "What is stacking penalty?"

"Stacking penalty" is a model for diminishing returns on identical ship/module effects. Or, in other words, the more effects of one kind you pile on, the less powerful each successive effect will be. If you're to remember one thing from this post, remember this:


If you look at the description of many modules, you will see this message or something similar:
Penalty: Using more than one type of this module or similar modules that affect the same attribute on the ship will be penalized.
It sounds ominous, but all it means is "the more of these you use, the less effective they will be". The above rule of thumb is simply boiling this down to the most popular cutoff point.

But... why?


Back in the day, things didn't use to be this way. All modules applied their full effect, all the time. This meant that some ships which had very high numbers of mid or low slots, could fit multiplier modules in all of them, resulting in insane total modification. We're talking "insane" as in "battleships moving faster than frigates by filling their low slots with nanos" levels of insane. It was so crazy that it started breaking the laws of physics.

So it was changed, and the Machariel shed a manly tear.


How big are the penalties?


It's not a linear progression, and the penalty grows the more modules there are. The actual equation and visual representation of how effective your modules are is below:


Note that 100% simply means 100% of the module's effect. For example, a Gyrostabilizer II increases projectile damage by 10%. The second gyro would increase the damage by a further 86.91% * 10% = 8.7%. And so on...

Summing up these modifiers, this is how much you get out of fitting a number of identical modules:

Update: Turns out I'm a big fat wrong liar, and there is no such thing as additive stacking. The graphs have been replaced with graphs showing the real multiplicative effect.


Math is fun, isn't it? Note how fitting the fourth module is only of tiny benefit, and all further modules are nearly useless... hence the rule of thumb.

In case you really love math and would like to play around with the stacking penalty calculation, you can use this function for the total multiplier (M) of n modules each giving a m bonus:


Does this affect guns? If so, it's stupid.


It would indeed be stupid if it affected guns, but fortunately, it does not. "Absolute value" modules such as guns, shield extenders, armor plates, etc. are exempt from stacking penalty.

Are there other exceptions?


Yes. Here are some notable ones:

  • Damage control modules. The boost they give to shield and armor resistances does not count as part of other resistance-modifying effects.
  • Fitting modules. Co-processors, reactor control units, capacitor rechargers, cargohold expanders, and other such modules do not suffer stacking penalty.
  • Trimark rigs and CDFE rigs. The amount of extra armor or shield hit points you receive from fitting multiple of these rigs does not diminish with each successive rig. This makes them extraordinarily powerful and popular on ships that need a lot of buffer tank (such as battleships or capital ships).
  • Implants. Bonuses from implants do not count towards the stacking penalty of a certain attribute. Snake implants are an exception, to avoid excessive speed problems.
There are a few others, but I'm not going to attempt to go hunt down all of them. With very few exceptions (such as falloff "ambit" rigs), all stacking-penalized modules are indicated as such in their description. A quick search should let you know if a module is stacking penalized, even if it does not have the warning in its description.

What if I'm using modules of different strengths? Is it possible that the strongest one gets penalized for weaker ones "counting before it"?



No, not at all. Effects are "counted" in order of their strength. 

Suppose I had an Energized Adaptive Nano Membrane II fit (20% resistance bonus to all armor resistances) along with an Armor Explosive Hardener II (55% bonus to explosive armor resistance). Since the hardener is the highest modifier, it gets counted first, and thus receives the full 100% of its effect, or 55%, while the EANM's explosive component gets counted second, receiving only 20% * 86.91% = 17.38% of its effect Since the other resistance bonuses the EANM provides do not conflict with anything else, they receive their full effects of 20% each.

My head hurts. Can I take advantage of all this somehow, or is this just hot air?


You can! By just knowing and applying all this, you already have an advantage over many other players. Some don't know better until it's too late (as one thing wrong with that fit). 

Now go forth and conquer!

13 comments :

  1. YES! YEAHHHHH! This be a DAMN PATRIOT, CEO of Dropbears Anonymous, and that's my dread pilot. That lossmail was the last thing I expected when I clicked on that link but it made my day, especially since it literally happened like, yesterday. That's a moros we stole and then sacrificed to BOB at the sun. http://www.reddit.com/r/Eve/comments/1t8snl/brave_collective_moros_killed_with_acreative_fit/ for the full story xD

    ReplyDelete
  2. I feel the need to point out that in some cases the second mod gives more then the first due to the bonus not being applied to the base but to the modified.

    ReplyDelete
    Replies
    1. Uh... what? That's just not true. The second mod can be extremely useful, such as how a second web improves holding someone down by a lot, but its effect is always weaker than that of the first mod.

      Delete
    2. What he means is:

      Say I have 10km falloff. I add one t2 falloff rig, which is a 20% bonus, bringing me to 12km falloff. So, the first rig gave me 2km falloff

      I add a second falloff rig. This is stacking penalized, so the bonus is (20% * 86.91%) = 17.4%. 1.174 * 12km = 14.088km falloff, so the second rig actually gave more falloff than the first one in absolute terms

      Now, most modules don't give a 20% stat bonus so usually this is not the case, but the way multiplicative stacking works does greatly mitigate the stacking penalty for the second module in most circumstances.

      Delete
    3. http://i.imgur.com/St0omNW.jpg

      Looks like you're right, but only partially. So long as the module's bonus is either below -15% or above 15%. So... Your T2 falloff rig is a good example. A gyro is not.

      Delete
    4. He did say "In some cases".

      It's also worth calling out anyway that the multiplicative stacking effect does counteract the stacking penalty for the second module in a large fashion. In the case of a gyro's damage bonus, the second gyro doesn't provide quite as much as the first, but it provides a lot more than 86.91% of the first.

      Delete
    5. A gyro is still an example of this if you look at dps
      The damage and RoF bonus of the first gyro will increase your dps from 100 to 123.
      A second gyro will increase your dps from 123 to 147, an increase by 24.

      Even though the combined damage bonus is slightly less than 1.2 and the combined RoF bonus less than 19%, you still get a bigger total dps increase from the second gyro when you multiply them together.

      100*(1+0.1)/(1-0.105) = 122.9
      122.9*(1+0.1*0.8691)/(1-0.105*0.8691) = 147

      Delete
    6. I was trying to stay away from incorporating the gyro's RoF component, since this post already had a lot of math going over people's heads. You're right, though.

      Delete
  3. This comment has been removed by the author.

    ReplyDelete
  4. May I ask where you got these formulas from? Experimenting on Tranquility/Singularity, Petition, ...?

    ReplyDelete
    Replies
    1. I'm guessing the original formula came from the data dump at some point in the past or something. As of right now I can't find any official source that just has it posted, but it is a known value. An easy way to confirm it ingame would be to

      -Put a tracking enhancer on a ship with guns. Note the optimal or falloff.
      -Add a second tracking enhancer, and note the new optimal/falloff
      -Manually do the math to make sure it makes sense given the numbers

      Delete
  5. This comment has been removed by the author.

    ReplyDelete
  6. A bit late, but should note that the reactive armor hardener module receives stacking penalties from the damage control module, but not from other armor resistance modules.

    ReplyDelete