User talk:Armond/Guide to optimization

From Guild Wars Wiki
Jump to: navigation, search

This post paid for by the Interational Association of TABLE Manufacturers[edit]

Score
(10 max)
Name Explanation Examples
0-2 Useless Actively harmful choices or overshadowed by other choices
3-4 Meh Might be useful for particular situations; avoid them generally.
5-6 Average Cannot go wrong with these, but there are better options
7-8 Good! Optimization begins here; these are the minimum requirements for a good team.
9 Best Think hard before dismissing these options. "I don't like it" is the kind of thinking that can make you fail hard mode.
10 Mandatory These define a build (whether team or individual).

In case the table format is of use to you (I am biased towards presenting data in this mode, especially when a lot of text would be repeated in bullet or paragraph form...but other peoples prefer different styles.) — Tennessee Ernie Ford (TEF) 06:23, 29 July 2011 (UTC)

The color-coding thing is an experiment. I'm mostly carrying it over from the d&d optimization forms, where it's a very common practice started by Lordduskblade, but I'm not sure it'll work on wiki. My main problem was that I needed a way to inform the uneducated as to what skills are powerful, why, and how they all compare to each other, and this seems a good way to do it. -- Armond WarbladeUser Armond sig image.png 13:12, 29 July 2011 (UTC)
The color coding is a very helpful addition to the guide (although perhaps using bold face plus a color overuses the tools for visual emphasis). (You might also want to revisit the colors: silver and gold don't show up as well as the others... so maybe reds (below 3), oranges (3-4), yellows (5-6), and greens and blues for superlatives.)Tennessee Ernie Ford (TEF) 16:32, 29 July 2011 (UTC)
The silver and gold do blend in a bit too much in a white background. --JonTheMon 16:57, 29 July 2011 (UTC)
red and blue and purple look too much like links imo, bolding doesnt help much. I like orange like tef said but i dunno about red and blue. it's a shame silver doesn't work though, that's a cool color 24.130.140.36 17:53, 29 July 2011 (UTC)
and have ou thought of using a template so you can change stuff like this more easily? 24.130.140.36 17:56, 29 July 2011 (UTC)
Might this from Tef's suggestions on color codes, help? 72.148.31.114 18:37, 29 July 2011 (UTC)
The TABLE should follow a "5-star rating system" like PvXwiki; All the score of 0-4 is gonna be able to communicate to the reader is that the these skills/builds are bad. Reason for this could be explained better by the good folks at X-Play.
And the Colors could follow the existing item-rarity coding: 5-Gold = Meta-build & mandetory-skill, uninstall GW1 for failure to bring these ; 4-Green = highly recommnended in 90% majority of PvE encounters; 3-Purple = minimum requirement for optimization; 2-Blue = Niche use, can't go wrong taking one of these for the other 10% of PvE that Green doesnt cover; 1-Red = Useless, overshadowed & harmful... dont even bother buying/learning! Uninstall GW1 for failure to avoid them.
--Falconeye 02:29, 30 July 2011 (UTC)

I'm putting a 15k bounty on a set of 25x25 pictures of roman numerals I-VI. Will respond to other stuff later. -- Armond WarbladeUser Armond sig image.png 18:47, 29 July 2011 (UTC)

Apparently I have issues with time management. Let's see...

  • I like the idea of a template, because I can integrate that into my monobook.js buttons.
  • The more I think about it, the more I lean towards a five point scale. Like I once said of PvX, anything not good or great hardly needs mentioning unless there's a way to make it good or great.
  • While the background on the tables above looks terrific (excluding the red), I'm not sure how I could practically use it in the guide (unless I were to start rating/commenting on every skill using a table format, which might not be a bad idea once I figured out how to read table coding and how to sort by rating, preferably without a semi-redundant "score" column).
  • The more I think about it, the more I want to stay away from blue, red, and purple. They all look almost exactly like wikilinks, and even with bolding I find myself hovering over the text to try to click it.
  • That said, without bolding, I find the rated text doesn't jump out at me very much. Maybe I'm just too used to the d&d forums, but I find an advantage of the bolding is I can quickly and easily find the highly rated stuff (something that's a bit harder when the writer doesn't bold it, as many things are rated higher in their descriptions due to favorable circumstances).
  • I'm kinda leaning towards brown (0-4), dark orange (5-6), green (7-8), cerulean (9), and royal purple (10).
  • Once again, anyone who makes and uploads 25x25 buttons of roman numerals I, II, III, IV, and V will have my eternal gratitude and possibly some semi-material goods.

Comments on any or all of the above? -- Armond WarbladeUser Armond sig image.png 06:02, 2 August 2011 (UTC)

On a five-point scale... you have 10 points ;-) Those colors read ok...and are not terribly intrusive when emboldened. (I still think you only need one level of emphasis...but you'll see after you put more together). I'm assuming you won't have any trouble creating templates ...but feel free to impose upon me. — Tennessee Ernie Ford (TEF) 07:42, 2 August 2011 (UTC)
You and your silly math! It only looks like ten points. (You have to squint a bit.) -- Armond WarbladeUser Armond sig image.png 12:58, 2 August 2011 (UTC)
Maybe you can rate the PvE-only skills as a example/excercise in what your table might look like? 70 skills is more manageable then 330. ^_^ --Falconeye 03:46, 13 September 2011 (UTC)

Health vs armor[edit]

Short story: I find that health improves survivability in nearly all PvE situations. I can't necessarily explain why this should be so, but my teams with higher health do better than otherwise identically kitted teams with higher armor.

Long story:

Before I had seen Raine's analysis, I did a similar one and came to the same conclusion. I compared situations in which you had various damage amounts coming in a 10s burst, in a 90s battle, and in a 3 min battle. I mostly considered armor-respecting damage (since that required no additional assumptions). And, in theory, more armor was more better: the more you get hit, the more better. In a short burst, more armor protects you from needing more health; in a long battle, your health has time to recover (or be restored) and, again, the more better more armor would be.

The problem is: when I switched my teams from +armor mods to +health mods, PvE survivability increased substantially. (I can't speak about PvP, since I have little experience there.)

I think this is because the typical armor vs health comparison makes too many simplifying assumptions. For example, Raine's "team" doesn't really compare to any PvE group I've seen that causes folks difficulties. Second: timing is important: the AI simply doesn't try certain types of mega-spikes on high-health targets. Third, disruption and control (recommended by your optimization suggestions) causes the AI to lose focus. Fourth, prots/heals/reflection also affects the AI's targeting. Finally, the AI almost always chooses low health targets over high-health ones — in a team with lots of spirits and minions, giving more health to your casters almost ensures that the AI frontline will ignore them entirely.

I haven't done any testing since 7H and WoC came out, so it might be that things have changed. But when I was comparing all-armor vs all-health mods, I found that situations that were near-wipes/wipes with high armor became low-risk battles with more health; low-risk battles w/high-armor became cake walks with increased health. I'm not convinced that my explanation above is accurate, but I haven't ever regretted going for more health instead (even though it would have been cheaper to go for more armor).

There are specialized situations when I still go for armor (certain solo builds require it), but, in general, I prefer +30 health over +5 armor mods. – Tennessee Ernie Ford (TEF) 06:53, 12 October 2011 (UTC)

It's not a very neutral presentation as it is now. When it's not, y'know, almost 1 am, I'll work on that - because you're right, there's a lot of external factors like minions and spirits. (I have some opinions about minions, too, but eh.) -- Armond WarbladeUser Armond sig image.png 07:37, 12 October 2011 (UTC)

References[edit]

So, i don't know what it says about me, but I can only get the last 5 profession handbook references. Any clues for the first 5? --JonTheMon 14:08, 12 October 2011 (UTC)

Four of them are songs (one is from a video game, and one is [most of] a line from a song that happens to include the song's name). The last is a reference to a great archer teaching his pupil. -- Armond WarbladeUser Armond sig image.png 15:46, 12 October 2011 (UTC)
disenchanted isn't exactly their best work but it fits 24.130.140.36 15:47, 12 October 2011 (UTC)
Well, the last half is the better half. That said, I don't like the ele one as it is, but the only alternatives I could think of was "But Those Were Link's Spells" or just "Din's Fire". Also, I failed at Clannad reference. -- Armond WarbladeUser Armond sig image.png 16:07, 12 October 2011 (UTC)
Well, that might explain my issues; my musical knowledge is rather limited. Yeah, I'm not sure what other elementalist options are good. Maybe go with some sort of Equilibrium joke with "Tetraelementon". And I have no idea how to get a Clannad reference in. Maybe something with starfish? --JonTheMon 16:12, 12 October 2011 (UTC)
More Explosive Than Sanae's Bread. Vili 点 User talk:Vili 19:22, 12 October 2011 (UTC)

(Reset indent) Longshot dervish references (the current book title...leaves something to be desired): So, Dervishes are inspired by Sufis, whose movements also influenced Raï music. Reda Taliani is among the better known artists of the genre and has a song called, Les algériens des kamikhazes, i.e. the Kamikaze Algerians (well, the Algerians of the Kamikazes, to be literal). (Among the reasons I thought this might be worth giving whirl: Raï is considered controversial in a lot of the world (it's to Islam as Flashdance is to fundamentalist Christians) — more than a few singers have been killed for playing it.) – Tennessee Ernie Ford (TEF) 18:51, 2 January 2012 (UTC)

Recommneded reading...[edit]

Relyk's Guide Index - Top 10 OP Skills

--Falconeye 05:37, 19 November 2011 (UTC)

Silver[edit]

I know you love using Silver for the highest rating, but it looks washed out on my laptop and desktop screens under FFx. The other four colors stand out nicely (maybe too much, but I know you prefer the bold face, so pretend I didn't mention it)

I know you will likely need to find your own solution (or leave things alone), so the following is merely meant to help you brainstorm:

  • Use a Roy-G-Biv rainbow, from the worst to decent to best.
  • Red means stop, Green means go, e.g. worstdecentbest
  • Hot colors are dangerous, cool colors are, well, cool: worstdecentbest
  • If you want to get fancy, my toolbar has a link to some sites that help you pick colors, notably a color blender that allows you to pick a start and end color, from which it attempts to pick intermediate steps, e.g. worstdecentbest (you can pick up to 10 midpoints, far more than you'll need).

FYI Despite my posting multiple times about format, I do spend more time paying attention to the content. It definitely influences my team builds. Keep it coming! – Tennessee Ernie Ford (TEF) 21:49, 5 December 2011 (UTC)

I've been meaning to redo the whole color scheme, in fact. I may pick up that color blender and give it a whirl; it looks pretty handy, but I won't have the time to properly play with it tonight.
I've also been trying to convince myself to work on rating new stuff instead of reformatting the old, but ocd etc. Eventually I'd really like this to be mainspace-quality (either moved to the mainspace our linked from there, like your nick predicts), and that's a big undertaking just on the formatting side. -- Armond WarbladeUser Armond sig image.png 22:48, 5 December 2011 (UTC)
+1 on linking from mainspace (I think it's more useful as an opinionated guide, esp. if others create parallel tools). Let me know if I can help in a more substantive way (i.e. something more than sniping at color choices). – Tennessee Ernie Ford (TEF) 22:59, 5 December 2011 (UTC)
I do appreciate the color choice sniping (I'm thinking of settling on a red, orange, yellow, light green, green... thing), though if you have opinions to offer on my ratings or questions on why I rated things the way I did (or even "hey you said x and it's actually y), I'd be glad to hear it. -- Armond WarbladeUser Armond sig image.png 00:41, 6 December 2011 (UTC)

Corrections[edit]

Isn't LBGaze armor-respecting (not ignoring)? --JonTheMon 03:58, 7 December 2011 (UTC)

Words! They are hard for me. -- Armond WarbladeUser Armond sig image.png 06:52, 7 December 2011 (UTC)

LoD[edit]

"(Also, for the record, you are probably better off taking another skill and clearing the dungeon faster than you are hoping for a good drop from a hidden treasure.)"

With the current fix to Light of Deldrimor treasures, this is no longer even remotely true. With vague familiarity of locations (and remembering to skip most/all on L1), you extend your dungeon visits by 5 or 10 minutes at worst (if two people carry it, there's even less delay). In multi-level dungeons, you can expect 5-15 treasures per floor. I keep forgetting to count them all on Frostmaw's, but it's 20 or more (even Oola's has about 10 on L3). I pretty much expect a ruby or sapphire about every other visit now (and multiple gems are common). It's rare now that the end chest has better loot.

Mind you, that doesn't make the skill important on an optimized build (although the AoE holy damage comes in handy, sort of like a "Finish Them!" vs undead). – Tennessee Ernie Ford (TEF) 06:19, 7 December 2011 (UTC)

That's good to know. I've updated the page accordingly. -- Armond WarbladeUser Armond sig image.png 06:52, 7 December 2011 (UTC)

Le "pull"[edit]

Pull can be used by players to mean three different things:

  • Drawing aggro by pulling the enemy towards your (flagged) group.
  • Defanging enemies (by making them waste their initiative on debuffs, moving, and casting on the puller, who is well protected). (i.e. pull their first punch)
  • Controlling the enemy by forcing them into choke points or balling them up.

All three are useful tactics, but they take differing amounts of player skill (the first takes none, the last takes practice by the puller and patience by other players in group). Most players (at least in my alliance) don't seem to understand the difference (even those that regularly engage in Urgoz/Deep runs, that make use of strong pulling/tanking tactics).

Even if you don't care to make the distinction within your optimization guide, it's worth keeping in mind that not everyone understands the advantage of pulling beyond separating two groups of foes. – Tennessee Ernie Ford (TEF) 19:33, 31 December 2011 (UTC)

I guess I really meant "engage" - when I play with my friends, we usually pull with the assassin's face. -- Armond WarbladeUser Armond sig image.png 21:03, 31 December 2011 (UTC)