Weapons and Elite specs need a full rework for weaponmaster to work

[JusticeRetroHunter.7684]

2 hours ago, draxynnic.3719 said:

I think there is an argument by which opening up the weapons could cause the collapse to happen faster rather than slower, even if there are more options on paper.

Yes, this is largely what my past 2 comments above was saying. What the options do, matters.

To repeat what I said previously: You can have 1000 options that all do the same thing which in effect is the same as just having 1 option where all options are superficial. In the same token, a game with 10 options that all do the same thing, will again be equivalent to having 1 option because all options are superficial.

The act of opening up the weapons, as an operation, is not what does that effect...it is what the options do, that makes that effect happen. That's why the following statement makes sense: that when there are 10 options that all do different things, is less diversity, than 1000 options that all do different things. The space simply grows, since it's biased in that axis (to simply grow larger).

And that's why the two operations (homogenizing skill effects, vs opening options) are qualitatively different kinds operations.

Quote

The reason is that by breaking up buildcrafting into more, smaller decisions, the complexity of each decision is sometimes reduced. Let's say that hypothetically, you have two elite specialisations. One has good mechanics and traits for a particular role, but the weapon is less suited. The other has a really good weapon, but the mechanics and traits are a poorer fit for what you have in mind. (For simplicity, we'll ignore the core weapons, maybe there's either no core weapon that does the job, or the core weapons that do the job are already on the build but there's a space yet to be filled.) This can create a situation where a player has to weigh up a good weapon on a mediocre specialisation versus a good elite specialisation with a mediocre weapon, which might be a complex decision. Now, let's consider this scenario when the weapons are decoupled from the elite specs. In theory, there are now four choices - but instead of a complex choice between two alternatives, you just couple the good weapon with the good elite specialisation and ignore the other three.

Like mentioned above, and something I explain in another way in this thread or this thread from a while back, is that the same problem (the homogenization problem) is happening in this scenario as well in the following manner: The fact that you can say one specialization is "good" and another is "mediocre" comes from the notion that you could compare them, easily enough to make an equivalence statement about them, Specifically with numbers. you can imagine for example, that you have 4 skills (options), each defined by a number that you can easily categorize.

Skill A = 1000 Damage

Skill B = 800 Damage

Skill C = 500 Damage

Skill D = 300 Damage

The obvious best skill combination here is A and B together, which together do 1800 damage, which means for all n skills, the skill with the highest number, will always be coupled with the skill with the next highest number. If you had 1000 options, each with 1 damage margin of separation between them from 1-1000, then the two options that do 1000 damage, and 999 damage, will be the best combination to play.

But the problem here is the fact that you could make an equivalence statement about the skills in the first place...and this is the problem of decidability. Because you could evaluate it to give a decidable answer...you in effect, min-max the combinations and it leads to a homogenization problem.

Consider now that you didn't have a decidable problem...but a more undecidable problem... where the mechanics were not easily comparable...for example

Skill A 800 Damage 100 radius 3s Stability

Skill B 100 damage 600 radius 1s Immobilization

Skill C 500 damage 180 radius, 4s fury, 6s chill

These mechanics : Stability, immobilization, radius...these are not easily comparable mechanics, not enough to make sound equivalence statements between those mechanics. Easy example...is that 180 radius being equivalent to 500 damage or 3s stability (to make comparison) is not a provable equivalence. We even recognize from experience in the game, that things with 100 radius suck in comparison to things with 600 radius, but we don't know how much or how little they compare...like asking the question : "At what point(number) does radius become good?" There's no provable answer to that.

So for all options N with an undecidable problem space, it becomes impossible to make sound equivalence statements about different non-trivial behaviors, and therefor both the act of trying to balance them becomes impossible. In the same token, it is this same property that allows the options to have qualitative differences between said options, to be diverse...that's what it means to be different in the first place.

You can imagine a homogenization procedure of the above, being the following:

Skill A 250 Damage 240 radius 1s Stability 1s Immobilization 1s fury, 1s chill

Skill B 250 Damage 240 radius 1s Stability 1s Immobilization 1s fury, 1s chill

Skill C 250 Damage 240 radius 1s Stability 1s Immobilization 1s fury, 1s chill

Where we have essentially taken all of these diverse mechanics, and distributed them to each skill. NOW it becomes easy to make the skills comparable, to make equivalence statements. In this case, All options do the same thing, and therefor each option, for all options N are effectively the same choice.

In each scenario you can see that, what the skills do, and how different they are from one another (and by proxy how not-balanced they are) is what determines the diversity of the game. It is through the undecidability of trying to make comparisons between things that can't be compared to one another.

If you actually want to do an application and not just listen to what I said here, you just have to ask the following question to yourself, and do the following exercise: Take just two, non-trivial skills from the game, and attempt to perfectly balance them. If they are non-trivial, You will find that this procedure is impossible to do, and impossible to prove that the two things are perfectly balanced. It can not be done without homogenizing their mechanics, and the numbers that define those mechanics.

So in conclusion, when the mechanics of the game are non-trivial (not easily comparable to one another) then said options are truly diverse, and in any situation where the options are indeed diverse, adding more options, compounds the possibility space for more diverse options.

Quote

From what I've seen on the forum, there are quite a few people who have already done that analysis, collapsing what had been a complex choice into a simple one, even if they weren't thinking of it in those terms.

Guild wars 2's problem/diversity space is MASSIVE...not doable by a human being (and if it was an actual analysis of the full game state which includes human player decisions, it would be undecidable, and not doable even by a super computer.) I'm sure you've heard of super computers like Alpha Zero...that is the kind of computational power it takes to fully evaluate the combination/complexity state of a game like chess, which is way a simpler game than guild wars 2. I'm sure that whatever the analysis is, it's more than likely effective theories of the game (which I've done stuff like that myself), but not a description of the games complexity or diversity space. For example, just because one figures out optimal DPS on a Golem or Understand Roles in SPVP, doesn't make that the objective solution state for player choices. In other words, not everybody cares to do maximum DPS in all scenarios...and these effective theories change from game mode to game mode, from situation to situation.

Still, i wouldn't mind seeing links to any kind of analysis if you have any.

Below, is a good example of 5 interesting WvW builds that came out of this beta...As players continue to create new solutions in the possibility space, others will treat their solutions as problems to solve, developing their own solutions. That is the human driven force of metagaming at it's heart taking place.

 

 

Edited July 4, 2023 by JusticeRetroHunter.7684

[JusticeRetroHunter.7684]

1 hour ago, Leo G.4501 said:

@JusticeRetroHunter.7684 I think you're not reading the room.

Firstly, I'll say you're not in university or having a study peer reviewed. It's mostly just a game forum where people share opinions or numbers of their in-game feats (or facilitating obtaining said numbers). You don't have to cite your sources unless we're talking about "x Anet dev tweeted this" or "the decrease in x game mechanic is proved by that". If you have to use unrelated jargon to make your point, you're not communicating well. Revise your method...just a friendly pointer there.

I'm not here to cater to whatever message you want me to convey, nor do I care about how you want me to convey it. Takes a lot of nerve, telling someone else how they should live, and has nothing to do with the topic is just some personal issue you seem to have with information getting conveyed to you. So go check yourself... Those are how those concepts get conveyed and that is how I will convey them here. Just cause you don't get it, that's not my fault. Go do your diligence, or believe whatever you want to believe...it's a free world, free country, nobody is holding anything to your head. I provide links (to reputable sources) so that, people who want to do their diligence and learn something, can go and learn the it for themselves, and not have to rely just on what I said.

Edited July 4, 2023 by JusticeRetroHunter.7684

[Leo G.4501]

1 hour ago, JusticeRetroHunter.7684 said:

Below, is a good example of 5 interesting WvW builds that came out of this beta...

What, no interesting Warrior builds?

53 minutes ago, JusticeRetroHunter.7684 said:

I'm not here to cater to whatever message you want me to convey, nor do I care about how you want me to convey it. Takes a lot of nerve, telling someone else how they should live, and has nothing to do with the topic is just some personal issue you seem to have with information getting conveyed to you. So go check yourself... Those are how those concepts get conveyed and that is how I will convey them here. Just cause you don't get it, that's not my fault. Go do your diligence, or believe whatever you want to believe...it's a free world, free country, nobody is holding anything to your head. I provide links (to reputable sources) so that, people who want to do their diligence and learn something, can go and learn the it for themselves, and not have to rely just on what I said.

And learn to take constructive criticism. I know I may seem lowly compared to you with your verbose explanations but there is something to be said about brevity.

I'm just giving you the layman's retort to your posts: you're not really describing anything that isn't already inherently understood by most players.

You're welcome. I'll see myself out.

[Mell.4873]

Didn't we just get a full rework, now you want another.....

[JusticeRetroHunter.7684]

1 hour ago, Leo G.4501 said:

What, no interesting Warrior builds?

You won't know until you explore the landscape (the possibility space). Others will do the same. That's the point.

[draxynnic.3719]

10 hours ago, JusticeRetroHunter.7684 said:

Like mentioned above, and something I explain in another way in this thread or this thread from a while back, is that the same problem (the homogenization problem) is happening in this scenario as well in the following manner: The fact that you can say one specialization is "good" and another is "mediocre" comes from the notion that you could compare them, easily enough to make an equivalence statement about them, Specifically with numbers.

I used "one specialisation good, the other mediocre" mostly as a simplification to demonstrate the point that increasing the number of choices by decoupling them can actually result in decomplexifying the choice. Even then, you'll note that I specified "for a particular role". You gave your example of three skills, but if what you need for a particular encounter is Stability, obviously you'd take the stability skill for that particular encounter. Of course, adjusting your build based on what you expect to face is pretty healthy gameplay, so let's consider the examples I gave to be more general than "specific encounter that requires X".

The thing is that you're assuming that in order to make that judgement, people have to quantise everything. Of course, if you COULD quantise everything, then it would be a simple choice either way - whichever build accrued the most points would be the one you'd take unless you specifically needed something else. However, people are generally pretty good at coming to a gut feeling of "X is better than Y", taking into account not just damage but survivability, boon access, and a variety of other considerations, without having to quantise it and assign different points to different capabilities and sum them up. One example I saw before a previous post, for instance, was someone talking about Revenant, and how previously they had a difficult choice between playing herald and not having access to greatsword, and playing vindicator in order to get greatsword at the expense of playing an elite specialisation they didn't think was as good. Currently, that's a complex problem, but after the expansion, it will become a solved problem: they'll just play greatsword herald and get the best of both worlds. If enough people agree with that assessment, then the player base as a whole would have collapsed a complex decision between two options into a simple decision with three wrong choices and one right one. 

Of course, that "if" is a big question. If you also get a bunch of people deciding that, say, vindicator with shield is the best thing since sliced bread, then you go back to having two choices. If you get disagreement between whether herald works better with greatsword or with the weapons it previously had access to, then you have three choices. If there are people who disagree with the poster I was referring to and say that in their opinion greatsword vindicator is better than any herald (unless you specifically want quickness...) then that opens it up to the four choices that were there on paper.

How many cases there are where weaponmaster genuinely opens up range of viable options, versus how many cases where there are where it collapses them, will determine whether it leads to more or less homogenisation in the medium term. (Ultimately, more frequent balance changes and introduction of new options are likely to keep it from homogenising all the way since frequent enough changes won't give people the opportunity to determine what is Objectively Better except for maybe at the very top levels of play.) However, I think there are cases where it will, at least, result in some reduction of variety, which is part of why I brought in the second example. I'd give good odds, for instance, that every condition guardian is going to be running axe after Weapon Master, while every revenant that brings a ranged weapon at all will bring shortbow. Now, these are relatively inoffensive examples: if a condi guardian, for instance, could still be running core, firebrand, willbender, maybe even dragonhunter, then these are all still different enough that it might not matter that much if they're all running axe/torch. But it does mean that condi firebrand and condi willbender are just a little bit more similar if they're both using the same weapons rather than condi willbender using a sword instead.

[Burnfall.9573]

A Support Role Profession using a long range weapon; :Longbow and outperforming a Spiker Range Profession Role-Ranger Profession damage

Whose idea was it to give a Guardian Profession; a support role Profession..access to play..match and surpass Ranger Profession to its range, damage, role and its identity??

-Nothing more to say-

Edited July 5, 2023 by Burnfall.9573

[JusticeRetroHunter.7684]

49 minutes ago, draxynnic.3719 said:

I used "one specialisation good, the other mediocre" mostly as a simplification to demonstrate the point that increasing the number of choices by decoupling them can actually result in decomplexifying the choice. Even then, you'll note that I specified "for a particular role". You gave your example of three skills, but if what you need for a particular encounter is Stability, obviously you'd take the stability skill for that particular encounter. Of course, adjusting your build based on what you expect to face is pretty healthy gameplay, so let's consider the examples I gave to be more general than "specific encounter that requires X".

The thing is that you're assuming that in order to make that judgement, people have to quantise everything. Of course, if you COULD quantise everything, then it would be a simple choice either way - whichever build accrued the most points would be the one you'd take unless you specifically needed something else. However, people are generally pretty good at coming to a gut feeling of "X is better than Y", taking into account not just damage but survivability, boon access, and a variety of other considerations, without having to quantise it and assign different points to different capabilities and sum them up. One example I saw before a previous post, for instance, was someone talking about Revenant, and how previously they had a difficult choice between playing herald and not having access to greatsword, and playing vindicator in order to get greatsword at the expense of playing an elite specialisation they didn't think was as good. Currently, that's a complex problem, but after the expansion, it will become a solved problem: they'll just play greatsword herald and get the best of both worlds. If enough people agree with that assessment, then the player base as a whole would have collapsed a complex decision between two options into a simple decision with three wrong choices and one right one. 

Of course, that "if" is a big question. If you also get a bunch of people deciding that, say, vindicator with shield is the best thing since sliced bread, then you go back to having two choices. If you get disagreement between whether herald works better with greatsword or with the weapons it previously had access to, then you have three choices. If there are people who disagree with the poster I was referring to and say that in their opinion greatsword vindicator is better than any herald (unless you specifically want quickness...) then that opens it up to the four choices that were there on paper.

How many cases there are where weaponmaster genuinely opens up range of viable options, versus how many cases where there are where it collapses them, will determine whether it leads to more or less homogenisation in the medium term. (Ultimately, more frequent balance changes and introduction of new options are likely to keep it from homogenising all the way since frequent enough changes won't give people the opportunity to determine what is Objectively Better except for maybe at the very top levels of play.) However, I think there are cases where it will, at least, result in some reduction of variety, which is part of why I brought in the second example. I'd give good odds, for instance, that every condition guardian is going to be running axe after Weapon Master, while every revenant that brings a ranged weapon at all will bring shortbow. Now, these are relatively inoffensive examples: if a condi guardian, for instance, could still be running core, firebrand, willbender, maybe even dragonhunter, then these are all still different enough that it might not matter that much if they're all running axe/torch. But it does mean that condi firebrand and condi willbender are just a little bit more similar if they're both using the same weapons rather than condi willbender using a sword instead.

My point was that, when things are comparable, it means the skills are trivial (easy enough to compare) and that’s an issue (a homogenization problem) in how some (a lot) of skills are designed.

let me use another example: introducing especs is equivalent to putting more options in the game too. If you recieved 2 especs that did the same stuff, and one of them is clearly better than another…your issue is not going to be focused on the fact that presenting new especs is flawed…your issue would be asking why the especs are the same and why one of them is just an upgrade.

it is no different of a situation here. Options being presented isn’t a problem, it is what the options do. 

Edited July 5, 2023 by JusticeRetroHunter.7684