Guild Wars Wiki talk:Formatting/Skills

From Guild Wars Wiki
Jump to: navigation, search
Shortcut:
GWWT:SKILLS

NPC AI.[edit]

I propose that a section noting how NPCs use each skill be added to each skill page. At the very least, each page should list obvious shortcomings or obvious advantagesin the way that NPCs use them (e.g. spamming 5e skills instead of casting ZB). — Raine Valen User Raine R.gif 18:21, 12 Apr 2011 (UTC)

I'd just put it under Notes, since a lot of skills have quite normal AI usage.
We could, however, have a little template -- like we do with Anomaly and Bug. | 72 User 72 Truly Random.jpg | 23:14, 12 April 2011 (UTC)
I agree with 72: most skills follow a generic AI (not one particular to a skill). I would support creating an AI note template, because that would also make it easier to organize the skills, how we call attention to them, and ensure that the relevant generic or skill-specific notes are covered in hero behavior (or other articles).
For example, I don't think interrupt or Cry of Frustration should have a specific note that the AI can interrupt quickly (but chooses what to interrupt indiscriminately); I think we should document the general behavior. I think it's only noteworthy when the AI behaves unusually.  — Tennessee Ernie Ford (TEF) 23:34, 12 April 2011 (UTC)
There are a lot of skills with unique functionalities (like Diversion); there are many more with functions that mimic the functionality of other skills, butthat heroes use differently (like Poison Arrow: it mimics Apply in application, but heroes don't spread poison with it like they do with Apply). Even for things like interrupts, heroes have specific usage patterns beyond "fast reflexes" (for example, heroes attempt to Tease or Cry multiple targets). Even of there is a general article, there are a *lot* of anomalies that should be documented. Of there's an AI note template, I think it should be on every page, even on those with expected usage. — Raine Valen User Raine R.gif 1:15, 13 Apr 2011 (UTC)
Yeah I totally agree with this. How about we make a template with a structure like this User:Yoshida_Keiji/Sandbox for each skill page and then we also create a list to put them all together like skill listings. How about we start a voting to receive consensus? User Yoshida Keiji Signature.jpg Yoshida Keiji talk 07:26, 3 January 2012 (UTC)
You can see that this conversation didn't generate that much discussion 7 months ago, so I'm not sure that we're ready to discuss consensus (and, in any case, it's not a matter of votes). As it happens, I disagree with Raine's contention: I think before we can call something anomalous, we have to be clear about what we should expect to see. I also disagree with the assumption that it's reasonable to expect that heroes use skills like Apply Poison and Poison Arrow the same way: one is an attack, the other is a preparation; humans are smart enough to understand that optimum usage involves spreading the condition (and returning to primary target to destroy it), but I don't expect the computer to figure that out most of the time. I do expect the AI to use other condition-causing attacks similarly to PA and other condition-causing preparations similarly to AP.
Regarding next steps: I do agree there would be use for a specific template that related to AI behavior, but I'd like to see more examples of how it would actually appear in articles before we try to decide whether to start creating one and making use of it. I'm keen that we do a better job of documenting the AI, but I'm not convinced that we agree on what that might look like. – Tennessee Ernie Ford (TEF) 09:21, 3 January 2012 (UTC)


(Reset indent) 5 months later and a bit more experienced...

As the community may already be awared of skill usage by AI is currently inconsistent through the entire wiki since it is spread across different main pages and their respective talk pages, most commonly: skill, hero behavior and NPC (as they all say different things and none are confirmed). GWW:AI PROJECT will take the rides to maintain high standards: Replicable and Verificable. AI skill behavior notes that are NOT documented will be removed from main pages and transfered to their talk pages when worthy. Vague notes such us:

  • Heroes fail to use this skill.
  • Heroes don't use this skill as I want.
  • Heroes seem to need micro-management.

Are of no use, as they say nothing...these (imo) can be removed mercilessly GWW:BOLD.

The main issue is once again: Replication and Verification. What is needed so that the information can meet high quality values? In-game talking players would need a build to be "pinged" for the receiving player be able to test it. Well...we can't ping through Wiki, so... how do we duplicate that?

There can be extended discussions like my first ever intervention in wiki: Talk:Glyph_of_Lesser_Energy#Hero_usage, forgive my rustyness as that was my day 1. People can bring in several comments but still NOT HELP. What we need are:

  • Builds.
  • Iterations (number of test runs).
  • Normal mode / Hard mode.
  • Location of test ground.

These factors should provide enough information as a "build ping through chat", like this Guild Wars Wiki:Projects/AI skills usage/Testing. Plus every each player's heroes set-up will bring different results, which can be exposed like this: User:Yoshida Keiji/My default heroes.

In a few words, I will be removing skill main pages notes that "do not tell anything", or "does not provide enough information" so that others can analise as well. Which is what I have called as Talk:Hero_behavior/Unexpected_behavior#Documentation_Fail. Will be awaiting for opinions before proceeding. User Yoshida Keiji Signature.jpg Yoshida Keiji talk 01:48, 24 May 2012 (UTC)


AI secrets by developer Joe Kimmes[edit]

See also: Talk:Hero_behavior/Unexpected_behavior#A_note_on_Hero_AI_vs_NPC_AI

Since Joe Kimmes shared information to help documenting issues, we know that AI skill usage is not generic but has difference whether activated in Normal Mode or Hero/Hard Mode. This means that all skill pages with "old" reports need to be double checked. If notes starts saying "the AI uses it like this", we will have to assume its the Normal Mode half of an observation. While if a note starts saying "Heroes use it like this", we will have to assume its the Hard Mode half of an observation. We will need to make a distinction since this might cause despaired comments among players.

Also, we need to update the syntax, because players are adding their observations like Anomaly or maybe even as Bugs, e.g.: Wandering Eye.

Plus, main page notes shouldn't be added without discussion to be consistent with Hero behavior/Unexpected behavior where we ask users:

Info-Logo.png Before posting here:

Please confirm potential issues on the talk page, so that we can provide focused statements on this page. When identifying problems, please include the location where the problem occurred, the name of NPC using the skill, and the difficulty (hard mode or normal). See also: this post by developer Joe Kimmes.

...and this in the talk page:

Info-Logo.png AI complexity and how to report issues: "Most AI problems with skill use are because of problems with the individual skill AI. Further complicating matters, a skill can ...have more than one AI ... Since Heroes need to use skills very effectively, many skills have had their AI enhanced - ...[and] ... skills have chapter-based AI restrictions - Heroes and Hard Mode characters use the best AI, and characters in normal mode use decreasing tiers based on chapter....

"Most of the time, when I update the AI, I'm only updating the Hero/Hard Mode version - otherwise, normal mode could be affected negatively. So, when you're reporting AI issues, make sure you're specific about whether it's affecting Heroes or just normal mode characters."Joe Kimmes (Original Post)

We are once again facing Talk:Hero_behavior/Unexpected_behavior_Archive_1#Documentation_Fail. And as it is written above on May 24th... Notes on skill pages without discussion should be removed mercilessly. I said this 2 months and had cero opposition, which is not same as consensus but..."The lack of responses is primarily because the people who've most contributed to the guidelines and who were most active in applying the guidelines are also mostly inactive now, User:Ab.er.rant"

So I will go through skill pages and... remove everything related to AI skill usage that has no document in its talk page in a matter of days... User Yoshida Keiji Signature.jpg Yoshida Keiji talk 16:13, 22 July 2012 (UTC)

Next steps: template and clean-up[edit]

I am currently designing this Template:AI behavior to incorporate to skill pages in an attempt to resolve the problems mentioned above:

Template:AI behavior

  • NM and HM splits.
  • Unexpected behavior should be compared with common general deault behavior described at Hero behavior. Avoid comparisons between skills (which often generates more wiki inconsistency).
  • Re-directs users to document research before editing carelessly and also informs of other user's reports.
User Yoshida Keiji Signature.jpg Yoshida Keiji talk 04:28, 24 July 2012 (UTC)
I don't understand these assumptions:
  • "If notes starts saying "the AI uses it like this", we will have to assume its the Normal Mode half of an observation. "
  • "...if a note starts saying "Heroes use it like this", we will have to assume its the Hard Mode half of an observation. "
Why would this be remotely true? People use the terms "AI" and "Hero" indiscriminately, with "AI" sometimes referring to heroes and sometimes to NPCs. I don't think we should make any assumptions at all about what people meant. Mostly people don't use henchmen at all, but they, too are affected by the same AI. So we should say "NPCs use the skill..." , unless we have specific data that shows that heroes, hench, and/or allies use it differently from each other. (Possibly with the caveat that most tests are run using heroes.)
Further, we shouldn't assume that every skill behavior is wildly different in nm and HM. Different? Sure, but so different that an oddity in HM doesn't happen in nm (or vice versa)? Maybe not.
Also, people disagree about what constitutes an anomaly, a bug, or an oddity of hero behavior. I don't think we can define the difference in a useful way in the formatting guidelines. I don't think the Wandering Eye example is necessarily anomalous (or even odd) for other reasons (I would expect different behavior for an AoE hex vs a single-target hex), but it was posted as a testable hypothesis (although it was spare on details).
Finally, as mentioned on my talk page, I'm worried about removing stuff just because we don't see the discussion in an obvious place. (Many times, people discuss skills on personal talk pages, skill pages, related skill pages, or in edit summaries...and those might be missed. I'd prefer a smaller list of things to verify.) – Tennessee Ernie Ford (TEF) 05:12, 24 July 2012 (UTC)
"Further complicating matters, a skill can actually have more than one AI determining its usefulness. Since Heroes need to use skills very effectively, many skills have had their AI enhanced - however, applying this to all characters would increase the difficulty of the game, since some mobs might become much more effective. So to combat this, skills have chapter-based AI restrictions - Heroes and Hard Mode characters use the best AI, and characters in normal mode use decreasing tiers based on chapter. Most skills have 3 or fewer AIs - one for the newest version, one for Nightfall when heroes were added, and the original general version.
Most of the time, when I update the AI, I'm only updating the Hero/Hard Mode version - otherwise, normal mode could be affected negatively. So, when you're reporting AI issues, make sure you're specific about whether it's affecting Heroes or just normal mode characters. by Joe Kimmes Talk Page‎"
  • If an observation is done in Norma Mode on henchmen, other allies and foes. We will get the decreased tier analysis,
  • If an observation is done in Hard Mode or Heroes, We will get the optimal analysis.
This difference is what I am trying to split since the beginning of my comments.
User Yoshida Keiji Signature.jpg Yoshida Keiji talk 12:36, 24 July 2012 (UTC)
We cannot assume that people will use language carefully. I also disagree with the phrasing "decreased tier" or optimal analysis — it's just different to test in nm or HM, with heroes or with hench or allies. The point is merely to be specific, so that Joe has enough data to replicate the circumstances and zero in on the specific behavior. For example, in these contrived examples:
  • Skill A usage is inconsistent: it heals and removes a condition, but in nm & HM, heroes use it to remove conditions and not to heal. However, in HM, the general AI prioritizes emergency healing, so sometimes heroes will use it to heal (and sometimes not). To identify this, we need nm & HM reports.
  • Skill B is bugged: it's supposed to heal and remove a condition, but it only removes the condition. However, the AI treats it as a healing skill. Since hench and allies tend to have more limited skill sets, we might see them spam it as a heal. Heroes, with more optimized bars, might appear to be using it to remove conditions. Again, to see this, we'd have to observe different NPCs and in different modes.
The examples might be contrived (or bad), but my point is that all observations are valuable and that we don't have the same information that the dev team has, so we can easily draw incorrect or misleading conclusions. All we can do is our best to ensure consistent reporting with replicable details and to draw attention to problematic skills. – Tennessee Ernie Ford (TEF) 16:26, 24 July 2012 (UTC)


I want to start applying the template like this:

Examples[edit]

Skull Crack[edit]

Template:AI behavior

Mind Freeze[edit]

Template:AI behavior

Putrid Flesh[edit]

Template:AI behavior

Ward of Stability[edit]

Template:AI behavior


Ready to update the syntax with Template:AI behavior? User Yoshida Keiji Signature.jpg Yoshida Keiji talk 18:43, 28 July 2012 (UTC)
Not yet. Could you create a sandbox article (or three) of how you see it being used? Thanks. – Tennessee Ernie Ford (TEF) 23:15, 28 July 2012 (UTC)
Either we dedicate a new section to AI behavior at the bottom (above Trivia and maybe below/above Notes), or we just paste this template at the Notes section, these four examples are ready to be copied to each of their corresponding skills. The template by itself displaying all relevant links (skill talk page, Unexpected behavior and Hero behavior) will show users where to discuss appropriately and understand the reasonings behind each Documented research and study. I believe most users are unawared of such pages. The template shall help consisting wiki usage. These four examples are skills I have researched myself, and then we go picking up one by one as users make contributions, we double check and finally incorporate this new template. User Yoshida Keiji Signature.jpg Yoshida Keiji talk 01:38, 29 July 2012 (UTC)
So...? User Yoshida Keiji Signature.jpg Yoshida Keiji talk 05:37, 1 August 2012 (UTC)


Template was included to Formatting Guideline after 3 months of discussion with no strong opposition. User Yoshida Keiji Signature.jpg Yoshida Keiji talk 07:04, 13 August 2012 (UTC)

(Reset indent) I'm irked to find this template on skill pages :/ User Chieftain Alex Chieftain Signature.jpg Chieftain Alex 10:50, 13 August 2012 (UTC)

More skill categories for auto-generated skill lists?[edit]

There are plenty of skill lists on this wiki. They were once created by someone, some weren't finished, maybe skills were overlooked, and then they were invalidated by patches. We cannot possibly maintain all the lists that were once thought to be a "good idea", but are now full of misinformation. While there's an incomplete category, it's evident that nobody bothers to re-check all these lists periodically. The worst offender is probably List of skills by related subject, which aims to be everything, but is in a constant state of half-useful and half-outdated.

A possible solution would be to move the maintenance to the skill's article and autogenerate the skill lists from there. Instead of listing all skills on the Burning page, we might instead have the Categories Skills that cause burning, Skills that prevent burning etc. This way, when burning is removed from Skill X, we just edit the Skill X's page to update the new skill description and categories and we're done. Verification is simple: view the game update page, click all links, and you have all the pages that need to be checked.

Additionally, it would allow on-the-fly creation of other lists. Need a Party healing article? Category: Healing skills and Category: Skills with Party AoE. Easy.

The downside is that this will require a lot of categories. For every game mechanic, there may be Skills that cause X, Skills that prevent X, Skills that require X, Skills that benefit from X, and possibly the distinction between Skills that raise X and Skills that lower X (e.g. speed boosts vs. snares). We would need to create and agree on a list of categories that are deemed useful. A category Healing skills is needed, but do we need Skills that modify healing (small list) or even Skills that trigger on healing (two skills)? And what's the correct name? Skills that cause X would be consistent, but Healing skills sounds better and is more concise than Skills that cause healing.

The second downside is a strict requirement to clone Silver Edge a couple of times, because someone has to update all 1300 skill pages. ;) I suppose it's bearable if we do it one Attribute at a time - there's no need to change all skills overnight; we just can't use the new autogenerated lists until categorization is done. I'd also write some greasemonkey scripts that make it as easy as opening the skill page, clicking on the correct categories from a list, and hitting Save page. Still, I'm not going to do it all alone. Volunteers, anyone?

I have a rough and incomplete draft of categories on my hard drive (more than 100 already), but let's discuss the general idea before going into details. It will be a lot of work upfront, but IMHO it will pay off to give us a better, more accurate, and more easy to maintain wiki. Tub 14:49, 25 May 2011 (UTC)

I'm in favor.
I think we can reduce the amount of work required in the following ways:
  • Limit the number of things we list to things that we categorize; if we don't see a need to cat, let's not list.
  • Limit the number of things we categorize to things we have discussed as important.
    • We can do this for a class of cats, e.g. condition-causing skills, so we wouldn't by any means needs to have 100s of discussions.
    • We can further reduce the amount of discussion by grandfathering in any list/cat that has more than x number of page views (e.g. 1000).
  • If we decide on a list, a bot can do the work for the current skill lists that are +/- 100% accurate.
Even if we decide not to use any of the above tricks, I still agree with Tub that we should start switching to cats and auto-listing, even if it takes us a long time to finish the conversion process. — Tennessee Ernie Ford (TEF) 16:02, 25 May 2011 (UTC)
I'm somewhat against this (similar to the issues i've brought up regarding Category:Dark damage), but if it goes through, I would prefer if they were hidden categories, as to have less effect on pages. --JonTheMon 16:09, 25 May 2011 (UTC)
Can you please link the discussion for the Dark damage category? It's not at the page you linked and I can't find anything related in your recent contributions.
Falconeye seems to love creating new pages and categories, but mostly leaves the pages unstructured and unfinished, which is why Category:Fire damage is as useless as many of the skill lists this wiki has. This approach won't work, we need to decide on categories first, then go through all skill pages systematically. This includes consistent category names and hierarchies, of course.
I consider visible categories more useful, because they provide a function similar to the "Related Skills" section. I estimate 2-4 new categories per skill, so it shouldn't blow up the pages too much. If space is an issue, I'd rather hide the categories "Energy Skills", "Prophecies Skills", "Fire Magic Skills" etc, because they're redundant with the infobox. Tub 16:54, 25 May 2011 (UTC)
I have started to re-read this monster wall of text from the beginning. I will stress that too much information on the skill page is just as bad as too little. My approach to create a small list was ultimately doomed to fail because it couldn't be applied to every skill. What I will add now is this: "How do we decide which categories are important, which are redundant, and which are both?"
Obviously when dealing with this we will run into some redundancy. It is almost impossible to provide complete coverage without some overlap. Now, like you said, we don't need categories like "Prophecies Skills", "Energy Skills", etc. Categories of this type should be separate from the sort of functional relationships categories. Then we will run into smaller categories, like "Converts Damage to Health". This is limited to a very small number of skills. Would we want that to be a category, or would we want to throw a small list into the notes of each skill?
I don't mean to just poke at this renewed attempt without offering suggestions, but these are the problems which will arise for certain skills. Hopefully we will be able to figure out an approach that is both consistent and concise. Edit: I will be compiling a list of functions for categories I find important. Perhaps we can compare and contrast and derive which are truly informative and which are redundant and which are trivial. FleshAndFaith 01:36, 26 May 2011 (UTC)
imo, include every conceivable category, and let them remain hidden, as was suggested earlier. This way, there's complete documentation for every possible purpose. — Raine Valen User Raine R.gif 1:43, 26 May 2011 (UTC)
I'm against every possible purpose as a rubric for deciding what to keep; we've seen a lot of unnecessary and duplicative categorizations recently. I'm also against creating new categories or new lists during this effort: (1) let's start with what we already document and have trouble maintaining; (2) find a way to better maintain these lists (auto-magic, suggested above, is one such system); (3) make the necessary adjustments to fit the new system; (4) consider whether we are missing other useful lists or categorizations.
I don't really care if there are 10 or 100 categories listed at the bottom of an article; I try not to rely on categories to find information and instead prefer to see appropriate navbars, see also links, and so forth. (Lots of category pages don't offer a useful way to compare/contrast/filter, since there are too many articles listed.) — Tennessee Ernie Ford (TEF) 02:04, 26 May 2011 (UTC)
I agree with Raine, to a certain extent (still a first). Making every plausible category would help when updates roll around and change things up drastically. That said, how and why do we "hide" the excessive categories, and which categories are considered excessive? How do we decide which are worthy of being shown and which should be hidden? Do we have the autogenerating robot choose the 5 largest categories? And will we have a page with a list of every category? If so, do we hide the excessive categories at the bottom? How would we present a list of every category, if that is indeed in the plans ahead?
I think it mostly comes down to the coding, which I am no help with (I'm retarded). FleshAndFaith 02:14, 26 May 2011 (UTC)

Skill quests[edit]

It might be nice if we added profession icons for eligible professions next to skill quests (e.g. Mesmer A Mesmer's Burden). It would be nicer if we had campaign icons as well (e.g. {{proph}}Mesmer A Mesmer's Burden). Are there any campaign icons lying around? -- Armond WarbladeUser Armond sig image.png 20:17, 26 December 2011 (UTC)

I'm okay with this. One possible campaign icon might be the concept art for mish completion, e.g. "Factions mission complete" concept art.jpg. (cleaned up to remove the background, o/c). The presentation might look something like...
"Factions mission complete" concept art.jpg Mesmer A Mesmer's Burden
Tennessee Ernie Ford (TEF) 20:29, 26 December 2011 (UTC)
Where exactly do you want these icons to appear? On the Acquisition section on skill pages like Conjure Phantasm? The campaign is already mentioned, and Mesmer as a requirement is a given, since it's a mesmer skill. Can you give an example where the added information would not be redundant? Tub 21:58, 26 December 2011 (UTC)
Yeah, I was thinking of the acquisition section. As for the profession skill icons... Yknow, I'm not sure what I was thinking there. Probably confused my brain after looking at Foible's Fair. The campaign icon is really the only useful part, and then only for indicating that a quest requires your character to be from that campaign. -- Armond WarbladeUser Armond sig image.png 18:33, 27 December 2011 (UTC)

Activation time[edit]

Unless I have self-destroyed my eyes spending too much time in front of the computer... I am seeing several skills like for example: Blessed Light displaying 3/4...but ingame its 5/4... I tried to fix it myself but the template breaks... Do I need to go to the ophthalmologist? User Yoshida Keiji Signature.jpg Yoshida Keiji talk 13:48, 18 January 2012 (UTC)

Sort your skill list by activation time. It's 3/4ths. Manifold User Manifold Neptune.jpg 16:57, 18 January 2012 (UTC)

Related skills: professions[edit]

I would like to suggest an improvement to the formatting recommendations for related skills: add the profession icon before the skill icon, at least in the cases when this list contains the skills from more than one profession. If all related skills belong to the same profession as the skill in article, icons are not necessary. An example below illustrates the reason of suggestion.
Current look of Soldier's Fury:

Suggested look:

As you see, in this case most of related skills (which should give the suggestion about similarity / replacing / synergy with other skills) are warrior skills, so a direct substitution without switching the secondary is somehow limited. Or, if the related skills are from more than 2 professions, it can help with choosing the appropriate secondary. Such approach is already used on the pages about skill effects, like Stance#Related skills. --Slavic 09:19, 22 April 2012 (UTC)


Skill progression[edit]

This template was replaced with {{STDT}} tables in three primary attribute main pages: Expertise, Leadership and Mysticism.

{{skill progression
| attribute = [[Full attribute name]]
| var1 name = [[Damage|+ Damage]]
| var1 at0 = 5
| var1 at15 = 20
| var2 name = [[Burning]] [[duration]]
| var2 at0 = 1
| var2 at15 = 7
}}

The STDT template when highlighted with: bgcolor="lightblue" is not same color and doesn't look a whole unique table. I was told that light blue is default highlight color...should the Template's default green background color be adjusted to a single color as well? User Yoshida Keiji Signature.jpg Yoshida Keiji talk 12:04, 21 May 2012 (UTC)

You mean light blue is a highlight for breakpoints? If your goal is consistency, you could create {{breakpoint color}} so that everyone uses the same color for highlighting breakpoints. I don't think it should be the same color as skill progression tables, because the two tables are demonstrating different concepts. One calls out ranks 1, 2, 15 and the other compares neighboring cells.
If you want the two tables to look similar and not like {{STDT}}), you could create {{breakpoint table}} by copying just the formatting from {{skill progression}}. Editors will still have to manually set the background color for each breakpoint cell, unless you program the math to determine which cells are different from their neighbor. 75.37.19.128 03:59, 24 May 2012 (UTC)

Acquisition -> Campaigns[edit]

Currently the bosses from Beyond campaigns are specified under Prophecies and Factions subtitles. Would it be better to recommend include them separately under War in Kryta and Winds of Change subtitles? Reason: usually there are two categories of players who need to capture skills: firsts attempt to do this ASAP for builds, while seconds (much more experienced) are working on Skill Hunter title. Beyond is much less interesting for the firsts, but for the seconds it's a good idea to capture skills while moving along the storyline. --Slavic 12:50, 14 July 2012 (UTC)