User talk:Joppe

From Guild Wars 2 Wiki
Jump to: navigation, search

condition damage tables?[edit]

What is your intent with the tables you're adding to skill pages? Condition duration/damage scale the same way for all skills, so there really shouldn't be any need to document those general mechanics on individual skills. —Dr Ishmael User Dr ishmael Diablo the chicken.png 18:58, 26 September 2012 (UTC)

My intention is for the reader to find out how much damage their skills are doing and what are the breakpoints for dealing more damage. That depends on their Condition duration bonuses (only full seconds count), Condition duration of their skills (depends on rounding) and Condition damage stat. Later on when this wiki has actual numbers how each skill scales with the Attack stat, readers are able to make meaningful decisions what sorts of stats they want on their equipment. You're correct that mechanics of Bleeding are universal, but total damage output depends on skill. I feel this information is not obvious, but rather useful.--Joppe 19:10, 26 September 2012 (UTC)
The duration breakpoints are still generic and could be summarized in a table on the Condition Duration article, similar to gw1:Expertise.
As far as damage comparisons go, there are really way too many variables involved to be documented by a concise table. You've got Power, Condition Duration/Damage, Precision, and Critical Damage attributes; weapon strength; skill recharge times; etc. Also, your weapon skills are all linked together, so comparing individual skills like Impale and Riposte doesn't help very much.
This sounds more appropriate for a profession guide article, where you can go into detail about each weapon and its skill set, then compare different weapons to each other and derive the best attributes to boost for each. That's the sort of thing that would be very valuable to players, definitely more so than isolated tables on individual skill pages. I would recommend you bring this up for discussion on the community portal, to see if other people have ideas or experience to contribute. —Dr Ishmael User Dr ishmael Diablo the chicken.png 20:07, 26 September 2012 (UTC)
I understand your point and I agree with the idea you presented about profession guides. I just had different method of seeking information. I would look at skill like Explosive Shot and have serious troubles understanding where the numbers come from and how should I go about improving its damage. So my first go-to point would be the skill page, since Pistol article mostly lists the individual skills and Engineer is quite far from what I'm looking for.
As far as individual skill articles go, I don't see how more information is harmful, even if it's derived from another article. There isn't much content on skill pages besides the (derived) numbers on damage and condition duration, which unfortunately have very little explanation power. Tables look ugly on this first draft and the values do not yet have comparison point. I'm still optimistic that this can work out into something useful for players, once we have more information about damage calculation and dynamic level adjustment (which affects Damage scaling but not Bleeding scaling). --Joppe 21:04, 26 September 2012 (UTC)
Still, discussing it on the CommPort would be would have been a good thing to farm ideas and opinions from the rest of the community before adding the tables to numerous skill articles that might have to be reverted if the consensus is that they aren't useful.. —Dr Ishmael User Dr ishmael Diablo the chicken.png 21:59, 26 September 2012 (UTC)
I plan to bring up discussion about (physical) Damage calculation, since that is one issue that definitely warrants it. This sort of information is extremely important for competitive gamers. I think the logical place for damage information is in skill pages, since each skill scales differently. That is the major reason why I decided to add Condition Damage tables for concise comparison without the need to swich between articles for each condition.
Thank you for your concern. I'm doing this voluntary work since I'm particularly interested in these mechanics. Those tables I added have edge cases where they wouldn't look stylish, like with skills that add multiple damage-over-time conditions. I'm definitely interested to hear if you have better "road map" plan for this.--Joppe 12:20, 27 September 2012 (UTC)
The skill page is for documenting the skill itself, as much as I would love to see numbers flying around, they aren't appropriate on individual skill articles. Players won't understand the tables, how to use them, or two to modify them as far as I can tell. You could create all the tables on a user page before adding them to skill articles so they are available somewhere with or without consensus.--Relyk 18:51, 27 September 2012 (UTC)

damage formulas[edit]

Also, the damage formulas you added for some ranger skills are wrong. Skills don't scale with attack (ie, weapon damage + power), they scale directly with weapon damage * power (ie, if you could somehow get power down to zero, all direct damage would be zero). Damage formulas are in general not that useful, since they don't give any additional information that the damage tooltip (...when it's accurate) does not already give. ~ Capric 01:15, 27 September 2012 (UTC)

Thanks for this info. Are you saying that each skill scales the same way (e.g. naked traitless Guardian with X power and PvP Sword deals same damage with skills as naked traitless Warrior with X power and PvP Sword)?
I borrowed style of the formula from Hammer Swing article. --Joppe 12:45, 27 September 2012 (UTC)
Well, Warrior sword skills and Guardian skills are different skills, so they wouldn't deal the same damage. Note that the damages recorded on the wiki are uniformly at 916 power and wielding a PvP weapon (which is an exotic in terms of weapon damage).
What I'm saying is that if you increase your power by 20%, then the direct damage of all your skills will also increase by 20%. Same thing with weapon damage: if you switch out your mainhand sword to one with 20% higher average weapon damage, then all your mainhand sword skills will do 20% more direct damage on average. etc etc
Now, some skills (typically utility but weapon skills as well) aren't weapon attacks in they don't scale at all with weapon damage. I have theories on how those work, but I haven't gone through all the skills to verify that there's a general rule as opposed to skill-specific rules. However, as a general rule, the great majority of skills that do damage scale with your weapon damage.
I hadn't noticed that there were damage formulas involving attack showing up on other skills. That's less than good, since skills don't work like that. We should really revamp the damage article, since how damage and damage tooltips scale with everything except level is actually very well understood at this point. ~ Capric 13:35, 27 September 2012 (UTC)
If I recompile everything you said about weapon skills, we get the following formula for weapon skill damage: Listed damage = weapon skill coefficient * weapon damage * power. It respects your finding that increasing values of Power or Weapon Damage increase skill damage in multiplicative manner. Since we know listed damage, weapon damage and power, we can find out what the weapon skill coefficient is! It's probably something comparable to Condition Damage coefficients of damage-over-time conditions. --Joppe 14:53, 27 September 2012 (UTC)
Close. At level 80, it's tooltip damage = (skill-specific coefficient) * (average weapon damage) * power / (2600). How the tooltip damage changes with level is very murky, and if you have insights into that please share them. The short version of level 80 tooltip damage is "this is how much damage you'll do to a heavy training golem in the pvp lobby."
I'll reiterate that due to how things are standardized, there's no additional information that the coefficient gives you that the standard damage does not: it neither makes things easier to compare nor easier to analyze. And for skills that do not scale with weapon damage, it is completely impossible to separate out the "weapon damage" for that skill from the coefficient. I do not find direct damage coefficients to be useful in GW2 due to the existence and general superiority of the damage tooltip. ~ Capric 15:07, 27 September 2012 (UTC)
This post [1] has promising insight on how level scaling works. You might want to contact the poster for more details. The motivation for finding skill coefficient is itemization. Let's say you have a skill like Explosive Shot where you can see at the skill page how much damage it does at level 80 along with standardized Bleeding damage. As a player I'd like to know what gives more bang for the buck: going with Power / Precision / Critical Damage stats or going with Condition Damage / Condition Duration stats. Coefficient gives you the basic idea how each skill scales regardless of what values (Power / CD, etc.) you currently have. We know that conditions such as Bleeding do not reduce their scaling with levels, but Power -based attacks probably do (i.e. 100 Condition Damage always gives you 5 bleeding damage per second, but 100 Power's impact on skill damage will depend on level). So the answer whether to go with Power or Conditions could in some cases even depend on what level character the player has. (edit: Condition damage scales with level like Power does; what I said about importance of coefficient still holds). --Joppe 16:22, 27 September 2012 (UTC)

That particular link has more to do with stat scaling, which—although not perfectly characterized to my knowledge—is a fairly straightforward quadratic progression. GW2 loves it some quadratic progression with level. During BWE2 and BWE3 there were fairly comprehensive dataset of tooltip damages collected around here. I looked at them then and wasn't able to figure out tooltip damagelevel scaling; scaling seemed to vary by skill, which is really annoying. I haven't gone back and re-collected the data post-release nor have I seen any even moderately comprehensive dataset that does have levelling data.

Here's my issue with coefficients:

  • They ignore weapon damage. Not all weapon types have the same average damage for a given level. Rifles have a very high average damage, while shields and warhorns are much lower. Thus comparing just the coefficients between two skills based on different weapon types is going to be misleading.
  • It is not clear that coefficients are constant across levels. Here's a link, as an example, to some very basic BWE2 tooltip damage info for elementalists: [2]. If coefficients were constant across levels, one would expect that the ratios of tooltip damages from any two given levels should be constant across skills. This was not the case.
  • The damage already gives you a very good indication of how the damage scales. And while I have a lot of gripes about the tooltip damages (they rarely take into account %-based damage increases and are flat out wrong more often than I would like), damage tooltips are very conveniently located directly in the UI and are updated in real time there.
  • What stats to itemize for will depend greatly on your traits and sigils. Sigils of Earth, for example, greatly alter the precision vs condition damage vs power balance, as do a number of traits that proc on crit.

To me, itemization strategies are more relevant in a profession guide where a holistic approach can be taken, rather than in individual skill articles where you will necessarily only paint a very, very small piece of the picture. So small as to be misleading, imo. ~ Capric 18:18, 27 September 2012 (UTC)

additional note: In case I wasn't explicit, it's only my opinion that coefficients are not useful in evaluating skills. If ultimately people want them anyway, I'm not really one to say they shouldn't. My bigger issue here is just making sure that any formulas and numbers are accurate and reflect what's in the game. ~ Capric 18:37, 27 September 2012 (UTC)

I agree with the point three of your concerns. Now that you've told me how damage works, it indeed seems like a very useful number. Condition damage numbers were bizarre when I started playing the game. Nothing in the game tells me that 2¾ second bleeding is as good as 2 second bleeding. The game even makes distinction here. I wish I had read an article that told me I need to increase my condition duration by 50% or 48% or wherever the rounding point is and then my damage would jump said 50%, or else I'd be wasting my money on duration runes.
I'll speak frankly here. I don't understand the motivation to curb mechanical ideas. We have people asking on forums questions like "what is a good weapon?" and getting answers like "who cares?", "every weapon is good", "I pwn with axe". We have articles like this one that are nothing vague, false, outdated, sensationalist marketing talk. Linking straight from front page. Is this what people look for? "We'll read the backside of the DVD case for you". That's not how I view the wiki project. Value of wiki is in deep understanding how the game works. It's about discovery what's beneath the surface. There's an old game, Diablo, that was practically solved by a person called Jarulf [3]. The guide answered every possible question about the game and then some. Nobody got inconvenienced about variables, coefficients or breakpoints. We've come a long way here where adding a table to skill page is too difficult to understand and worth wiping. These comments seem to imply that the skill pages are finished as they are, but I'm afraid they won't explain themselves to uninformed readers. I'm done ranting.--Joppe 23:38, 27 September 2012 (UTC)
I'm not trying to discourage you, and I greatly apologize if I am coming across that way; also remember I'm just another (somewhat opinionated, of course) user, and one who likes numbers but has no particular skill with—or much interest in, to be honest—presentation. I agree that the skill pages could use more information... I just don't agree with you on what precisely :) However, there are a lot of other categories that skills fall under which are all very useful that are at best incompletely captured on the wiki. For example:
  • Cast time.
  • Channel time, and when specific effects come out in multi-hit channeled skills or other complex channeled abilities.
  • Interrupt priority; some skills you can interrupt other casts with while others will be queued in.
  • Stationary abilities.
  • True instant cast abilities that may be activated during other skills cast time and when knocked down etc.
As for some pages being weaksauce, oh yeah. Like there is massive confusion pretty much everywhere about how damage works and what armor means etc etc. And the damage page is, imo, pretty terrible and badly in need of a re-drafting. The fact that the game UI lists "attack" as a stat on your character sheet but that stat isn't actually used anywhere does not help at all.
Uh, anyway I'm not really sure I had a point to all that. ~ Capric 00:01, 28 September 2012 (UTC)
"they won't explain themselves to uninformed readers" That's my issue: I don't think your tables of seemingly random numbers are explaining themselves very well. And that's why I'm arguing that we need high-level guide articles to explain everything. Seemingly random numbers scattered around individual skill pages don't provide any value on their own. —Dr Ishmael User Dr ishmael Diablo the chicken.png 00:49, 28 September 2012 (UTC)

"As a sign of courtesy, also leave a message on the talk page of any involved user. "[edit]

Sign of courtesy: [4]. Thank you very much for your attention. Pika Fan 17:54, 29 September 2012 (UTC)

accusations[edit]

Don't accuse someone of vandalism simply because they revert your edits. If they don't leave a clear reason in the summary, ask them for an explanation. —Dr Ishmael User Dr ishmael Diablo the chicken.png 21:08, 29 September 2012 (UTC)

attack speed[edit]

I don't understand what this number refers to. Skills have a listed activation time, and if that's flat-out wrong it's a bug. Vili 点 User talk:Vili 06:48, 14 October 2012 (UTC)

Attack speed is the cycle between attacks. For example Bleeding Shot has activation of 3/4 seconds, but takes around 0.96 seconds between each shot. This means that 0.75 seconds is spent in firing animation, then 0.21 seconds on preparing for next attack. Trick Shot on the other hand has lower activation at 1/4 seconds and longer cool off period. Both of these skills have similar attack speed, but Trick Shot's benefit is faster break from animation when moving and attacking. Attack speed is essential statistic for determining damage-per-second value. --Joppe 07:04, 14 October 2012 (UTC)
(Edit conflict) :Many skills have inconsistent duration tooltips and listed activation times (Renewed Focus, for example). It wouldn't surprise me at all if there was more to skills than activation time and recharge. Shouldn't be hard to test how long it takes to use either of those skills 100 times. Manifold User Manifold Neptune.jpg 07:04, 14 October 2012 (UTC)

math[edit]

Heya just responding to your main page stuff. First of all, all that math stuff looks right and I'm totally not trying to say or imply that you're doing it wrong. But just to be a pain in your ass (and feel more than free to tell me to just shut up if you don't particularly want to hear it), I'll illustrate how I would do the same thing, but without messing with those icky coefficients :)

So you got 2 skills, one does 155 direct and 255 bleed (henceforth 155d;255b), and the other skill does 134d;128b. To double the amount of direct damage you do, you add 916 power (ie, double it). To double the amount of bleeding damage you do, you add 850 condition damage. So to calculate marginal increase of those 2 skills with each stat, you just divide the respective base damages by amount of stat needed to double their respective damage. So you get 155d/916;255b/850 and 134d/916;128b/850, or 0.169d;0.3b and 0.146d;0.151b. And then you can (as you did) adjust the direct multiplier for crit, passive damage boosts (which do not affect condition damage), expected enemy armor etc etc.

Anyway, I'm not really arguing that you should do it one way or the other, since I've learned that how different people approach the same math problem will vary wildly. And that that's okay, since it doesn't really matter how you do it so long as you get the correct result. But it's like I'm a moth to the fire when I see formulas :) ~ Capric 05:49, 17 October 2012 (UTC)

Thank you for clear response. Your method is faster than mine. When I began to write that page, I thought skill coefficient is the same as power-to-damage coefficient. Only after testing it turned out not to be the case. Skill coefficient is built into the number listed on these wiki pages, so there's actually little ambiquity which skill is going to be more damaging. Only attack speed (which is not the same as activation time) is needed in comparison. I've been clocking some ranged abilities with simple clock and dodge-roll dummy at Heart of Mists, since the dummy is invulnerable. 100 seconds and 104 attacks is how I came up with Bleeding Shot's number. I wish there was a way to test melee and underwater weapons with similar large sample.
Critical chance increases expected damage by half of the value (with default critical damage 1.50). Critical damage is the multiplier for this multiplier. Fortunately value of criticals does not depend on skill damage, attack speed or armor. I think it's safe to create a table where X-axis is critical damage, Y-axis critical chance and cell value is the expected damage presented as multiplier. Such a table would be helpful for players and there's no need to go into coefficients to understand it. --Joppe 20:15, 17 October 2012 (UTC)
Oh yeah that attack timing stuff is pretty useful to know. I haven't done too much testing, but there seems to be a delay that's added in between the completion of a skill's activation and the initiation of the next skill. No idea if that's constant per skill or not, though. I have basically 2 ways that I get timing information on stuff. The first way is to do what you are doing, manual timing with high N. The other way is to get video capture of whatever I'm doing, and get times via frame counting. You still need replicates, but with the added precision from video capture you can get good numbers with a much lower N.
If you're looking for good way to get high N on attacks, I would recommend steady weapons. Steady weapons with lower power takes your damage down so low that you can get a lot of hits out of a single heavy golem. It's still annoying though, especially since the distribution of attack times is not well-defined, you can't really take any shortcuts to get a good estimate of the "real" timing. And obviously any skill that applies damaging conditions will still do way more damage than would be ideal.
Hm, now that I think about it, you know what else might work? There are target dummies in Lion's Arch near the PvP portals. If you're just looking to autoattack something with a stopwatch, those might work well too. ~ Capric 21:01, 17 October 2012 (UTC)