League of Legends Wiki:Discussions/Gold Efficiency

Disclaimer: This is the third thread on this subject. The last two threads have had to be closed because nothing productive has come out of them and both have dissolved into users personally attacking each other and not furthering the discussion. I am taking the reins of this discussion and if a comment does not adhere to our discussion policies or I feel as though it is adding nothing meaningful to the conversation I will remove it without hesitation. This is getting out of hand that we can not have a civil discussion without it coming down to something like this

What are we discussing
Changes to the gold efficiency section of item pages.

Why are we discussing this
Thread:1120532 started this. In summary, User:ClariS and User:Willbachbakal had a personal disagreement in regards to what defines gold value, methods of acquiring gold value, usage of gold value for passives and actives, and some basic wording things, for more info please see that thread. The discussion quickly added the debate of if the gold efficiency section should stay or be removed, which second thread can be found here: Thread:1122267. Please use these threads as a way of learning more about the subject, do not continue discussions from them here.

What are we going to do
Obviously there is a lot of disagreement about not only whether or not the gold efficiency section should stay on the wiki but also as to if it does stay should it remain the same as it was, should be changed in favor of including stats that don't have a set gold value or should it be redesigned completely.

What are my thoughts?
As per my post in the second thread I believe that the gold efficiency section should be redesigned to something that isn't so black and white and should use softer terms as to not sound so definite. The section as it was not only misleads users but has brought criticism upon us from a couple of sources.

 To see my post click the show button to the right: Gold efficiency the way we were doing it was too black and white and too emphasized on if the item is efficient or not.

I believe that the section should not be designed around whether or not the item is "efficient" and we should not be throwing around definite terms like that regardless of whether we had values for all the stats or not. Instead I propose that we focus more on just providing a gold value for the stats we can and use less severe yes/no only terms to make the point that the point of the section is just for giving gold values to compare with other items, not for determining the items worth as a whole.

This would allow us to continue to deliver data in its purest form to users and allow them make their own analysis's off of it instead of us providing it for them in the form of "oh, well, this item is 110% cost efficient while that one is only 102%, obviously the first one is the better buy all the time" which is something that some players do. We can still give values to certain stats but we don't have to feel obliged to try and give values to the ones we don't/can't that end up skewing the data with assumptions/estimates.

Some other things to discuss:
 * How we calculate gold efficiency, currently we use the least efficient item as our base, this means using long sword for AD and amp tome for AP.
 * Back in my day (3+ years ago) we used the most efficient basic item for a stat, that's how everyone did it, players and Rioters alike, this meant using B.F. sword for AD, recursive for attack speed, blasting wand/NLR for AP, etc. This would make things like amp tome not "gold efficient," which to us made sense.
 * Should we even list the "total value"
 * We can still call it cost analysis without throwing in a "total value," especially for items where some of the stats we can't give a value for which means we aren't really giving a "total." As I said before we could find different terminology or scrap it all together."

Discussion
I will start this off by referring to my "what are my thoughts" section. A happy middle ground between removing it and keeping it is by changing what it is fundamentally into a pure stat section. To give an example, instead of the current " is 79.5% gold efficient" and "Total Gold Value = 2147g" we'd say something like "Total calculable stats: 2147g." This takes out the elephant in the room, the word efficient, and gets rid of the word total in situations where it really isn't the total due to incalculable stats. 04:35, May 24, 2014 (UTC)
 * All suggestions to remove any 'implied' connection with effectiveness is a good thing in my view. ClariS (talk) 04:38, May 24, 2014 (UTC)
 * I concur, since it's not actual cost efficiency being calculated, but rather Comparative Item Value. Deshiba, the Nitpicker 05:25, May 24, 2014 (UTC)

Collapsible Template Proposal
I propose a slightly different set of changes. Through a productive exchange with Ajraddatz on the previous discussion, we came to a reformatting of cost analyses that could potentially make everyone happy. In addition, these could still work with the changes proposed above, though I'm not too fond of them (I'll explain below):
 * Cost analyses should be turned into collapsible templates on every item article. This allows them to be as long and thorough as they need to be and still take up the same amount of (minimal) space. Users who are not interested in the section or who do not like it can skip the whole thing instantly, and not have it take up any space in their viewing of the article.
 * That same template could also carry a disclaimer, placed within the template, as a tooltip on the template's header or preceding the template (I vote for option 2, the tooltip, if we are to implement this). Gold efficiency is, and always has been, a measure of an item's stats relative to the baseline price of said stats. It does not claim to be a measure of an item's power, nor does it measure the effectiveness of an item, yet the two previous threads suggest this tends to be heavily misinterpreted to the point where it would warrant a disclaimer. It would put aside any potential misinformation in a clearer way than just giving the item's gold value, while still leaving untouched a system that is itself correct, accurate and informative (at least, when the data is being handled correctly). --Willbachbakal (talk) 07:47, May 24, 2014 (UTC)

Comments
I am apposed to this proposal, because of: Deshiba, the Nitpicker 13:47, May 24, 2014 (UTC)
 * Page loading: With the revamp to the item list, I am experiencing page loading issues caused by the images or script. Having the suggested templates/scripts would escalate this issue. Note' I have: 25Mbps;down, 1.5Mbps;up, shouldn't be an issue on this end.
 * Disclaimer fallacy: Disclaimers are notoriously lengthy and when issues arise they need to be expanded. It might grow into a mini ULA, which the majority of people TL;DR.
 * Duplication: Having that disclaimer in a template on all 201 item pages is an amount of duplication that should be avoided.
 * Standards deviation: Having calculations/explanation on the item main page is a deviating behavior to any similarly complicated game mechanic (e.g., AD, Attack Speed, Critical Strike.)
 * I contest the following points:
 * Disclaimer fallacy: A disclaimer along the lines of "Gold efficiency is not a measure of an item's effectiveness and should not be the sole determiner to a build" would be short and concise enough to clear up all the misinterpretation issues that arose in the previous threads. Nonetheless, if you are opposed to the idea of a disclaimer, I'd be more than willing to drop it. Misinterpretation of data is a user issue and not a wiki issue, so the onus should not be on the wiki to tell people how to avoid reading the data wrong.
 * Duplication: Duplication of templates is a common occurrence in Wikipedia and every wiki, including this one (see the "additional info" sections that have integrated editing shortcuts to allow for easy repetition of information). A disclaimer in the form of a tooltip would actually be one of the least space-consuming forms of duplication here.
 * Standards deviation: This is a non-sequitur. Gold analysis follows from items. There is nothing to analyze in terms of AD or AS on an item other than how much AD or AS they have, and the only application gold value has to stats is simply giving the gold value of said stats. Moreover, calculations and explanations actually take up a significant portion of the health, armor and magic resistance, the latter two of which even have a very well-constructed and large equation that help explain the concept of effective health (damage of a certain type required to kill a champion as a function of their health and resistances, indicating a linear EHP scaling with armor/MR), so the initial assertion is also wrong.
 * --Willbachbakal (talk) 16:28, May 24, 2014 (UTC)
 * DisclaimerThe biggest concern here is that the disclaimer will grow over time when more situations need to be accounted for. The example you gave does not cover everything that Gold efficiency could be used for. An example of this would be, that the numbers are not based on anything that riot has confirmed or denied. Which would add another line and 50 characters.
 * Duplication, having a disclaimer that will potentially grow with gold efficiency calculations on every item page, when those calculations are also present on the gold efficiency page is a duplication of data, and needless creation of a template for the disclaimer. Having the disclaimer and calculations on the Gold Efficiency Page would remove the need for the template and the duplication.
 * Standards Deviation Oh but there is, the amount of AD and AS is not equal to the amount of DPS, which you have to calculate using both. And any calculation of the health, armor and magic resistance is done on their respective pages, not on the item page itself. Having "Amount of calculable stats: xx" and then linking to the gold efficiency page is in line with how we handle everything on the item pages, and the entire wiki for that matter.
 * Deshiba, the Nitpicker 13:19, May 25, 2014 (UTC)

Counter Proposal
I propose integrating, gold efficiency/total gold value/total calculable stats/descriptive term, into the trivia section of the item pages. Then placing a link on the chosen descriptive term back to the gold efficiency page where the calculation and explanation would be displayed. This would achieve a couple of things: This is also a golden opportunity to add something new to this wiki. This is like a little side tangent on it's own but relevant to the discussion. I was thinking, could we use links to send variables across pages, something like (different name)|Variable and end up with a link like: This way we could catch that variable and display any calculation going into that item, and also provide additional information like being a base item and having the LS value derived from it. You could have the rest of the items and their calculations hidden in a hide/reveal template so that would still be available. The item that redirected you to that page would then still take center stage. Deshiba, the Nitpicker 13:47, May 24, 2014 (UTC)
 * Making item pages a quick read and concise by turning: "Although not a basic item, is the item from which the life steal gold value is derived. From this item derives a value of 44 gold per percentage of life steal."" into "Total calculable stats: 800g.
 * Bring back Trivia sections which could encompass/absorb "Strategy" sections. That would in turn make the item pages look more uniform.
 * Bring use of templates down from and  to a simple link.
 * Increase traffic to the gold efficiency page
 * Provide a single place to include a disclaimer instead of duplicating it 201 times
 * Bring the gold efficiency in line with other game mechanics and information we currently display on item pages.
 * http://leagueoflegends.wikia.com/wiki/Gold_efficiency&item=Vampiric_Scepter

Comments
I don't like this idea: --Willbachbakal (talk) 16:17, May 24, 2014 (UTC)
 * Space issues: Cost analyses can be quite lengthy. A single cost analysis can take a minimum of five lines, and pushing everything to one page would bloat it horribly, to the point where it would be several times bigger than pages like the List of Runes and the List of Champions put together.
 * Readability issues: As the logical continuation to the above, a page that's horribly bloated will also be a pain to read. Just trying to find the right item in a sea of 178 different cost analyses would be needlessly difficult.
 * Item information belongs to the item's article: Gold value, being an expression of an item's stats, is a property inherent to that item. It makes no sense to ship that information off to another page when on the item's article we have sections such as trivia and even lore, both of which are lot less relevant to gameplay yet just as relevant to the item as gold efficiency.
 * Specialization beats generalization: As attractive an option as it might seem to put everything with a running theme onto the same page, databases work best when they concentrate all the information regarding a non-thematic topic to the topic itself, and a set of secondary articles on the side if the topic itself is too big. Wikipedia and even other wikis regularly use copy-paste templates across multiple articles that, although seemingly inefficient, work better than one single template on a mega-article grouping everything relevant to the template.
 * It doesn't achieve what it suggests: The above proposal states that moving cost analysis down to one article would put it on par with other properties, yet that is not actually correct. The armor and list of champions' armor page, for example, list items and champion abilities that increase or scale off of armor, plus every champion's armor, but does not remove that same information from the original articles. If we were to truly integrate gold efficiency in the same way as every other property, we would keep cost analyses on the item articles and duplicate the information on the gold efficiency page.
 * I might be able to remedy your dismay a bit.
 * Space issues will not disappear depending on where they are, either the item pages or gold efficiency pages. What my solution would actually allow is a visual restyle similar to the Template:Items page, where the overall bloating to the Gold Efficiency page can be limited. Where as per the way it is now, item pages like Trinity Force get bloated while item pages like Frost Queen's Claim use it to flesh out the page and the Gold Efficiency page is basically worthless because you can't find what you need.
 * Readability issues would be solved with said restyle
 * The information that Pickaxe has the "Total calculable stats of 900 gold", would not disappear from the item pages. It would just move the formula for getting at that number to the gold efficiency page, where it is duplicated from as it is now.
 * As per the trivia and lore sections of items, the reasons they have not been forked as per any champion page I can only guess at. Not every item has this section, but in my opinion they should have been forked a while ago.
 * Specialization beats generalization: <- you're going to have to dumb this one down. It's too complicated in the wording.
 * It doesn't achieve what it suggests: It would achieve exactly what it suggests if we follow the letter of the suggestion.
 * That the forking of that particular list from the armor page doesn't follow the conditions set when it was first decided to do so is an issue that needs to be resolved. And a different matter to this discussion altogether. (For the record, I was part of the effort for forking that data at that time. Whoever duplicated that should be dealt with.)
 * Deshiba, the Nitpicker 15:12, May 25, 2014 (UTC)

Restoring sections
Though not a part of the exchange, I also propose that we restore the sections of the cost analyses that were removed, namely the ones pertaining to item auras (e.g. ), item actives (e.g. ), and conditional passives (e.g. ). Because gold value is a measure of an item's stats, and that all of these effects produce more stats, they tend to increase an item's gold value in a given situation, and often by a large amount. Because these extra stats are only conditional upon certain situations, they deserve to have their own subsection in the cost analysis dedicated to those occurrences (i.e. the gold value of an aura's stats per person affected, the gold value of an item active while it is in effect), separated from the baseline gold value but with an additional line listing the item's total gold value when its effects are put to full use. This should also be reflected in the item's gold efficiency: the gold efficiency of its base stats should be listed first, but it should also have a line on how gold efficient it becomes when its conditionals are fully in effect. --Willbachbakal (talk) 07:47, May 24, 2014 (UTC)

Comments
The issue that arises when giving gold values for stats acquired from passives, actives, and aura's is that we our largely ignoring the mechanics, triggers, and conditions that were required to acquired those stats. The prime offender is, an item that grants 200% movement speed that decays over 8 seconds. The item was given a gold value of 11,900 for the 200% movement speed it granted, even though the 200% movement speed is only there at most for half a second upon leaving the fountain, while being completely lost if you perform any offensive actions or taking any damage. Even if there was a disclaimer there to remind people that it's only a temporary stat, the gold value presented has very little usage because there is so much restrictions that needs to be accounted that vastly changes the actual meaning of that gold value.

Therefore, at best, the gold value is estimation for few instances, but at worst, it's an lie since the mechanics are devaluing those stats. As a wiki, our main tasks are to archive information and provide facts. Numbers are facts, but numbers ignoring context has no meaning or worth. We shouldn't be providing information as 'facts' when we know it's not completely accurate.

I'm gonna start listing 'theoretical' situations on how the gold value is skewing the information falsely. All of these situations does not exist in game, but to a common person, the glaring problems should become evident on how misleading gold value would be if we were to use the same method we have been using while ignoring the mechanics, tiggers, and conditions. Now, let's take this a step further and asks, what makes this any different from the other items. Real items may not be to this extreme, but they fall within the same workings.
 * An item grants 300 armor at the instance you take damage from a small monster and only for small monsters.
 * Another item grants you 1000 health regen and 1000 mana regen for 1 second upon spawning.
 * Another item grants 1 Health for 2 seconds each time you auto-attack an enemy champion, stacking up to 1000 times.
 * And for a real stupid theoretical item, you gain 1000 of all stats only while you're dead (occurs only after all Resurrection effects are done).

TL:DR: The real question I implore all to consider is, how far can you ignore these mechanics before they become a glaring issues, and once you break that point, what reasons do you have that does not also apply to all the other items with active and passives. ClariS (talk) 13:11, May 24, 2014 (UTC)

Deshiba, the Nitpicker 13:47, May 24, 2014 (UTC)
 * I am partially opposed to restoring those sections because of:
 * Temporary Statistics: Using the least cost efficient base items gives the value constant for a statistic (buy once, always get the effect). Measuring temporary actives with that value leads to ridiculously cost efficient items. Example: The temporary regeneration of a is  gold efficient.
 * Auras could be calculated for the benefit it gives the user, because this is the effect it gives 100% of the time the item is in the game, so not suffering from the unrepresentative hyper efficiency of temporary actives.


 * I would be fine with bringing temporary benefits in if we remove % gold efficiency and just list the gold values associated with the effects. Like Aegis of the Legion would be something like:


 * 200 health = 533g
 * 40 magic resist = 800g
 * 10 health regeneration = 360g
 * ​Total calculable stats = 1693g


 * 20 magic resist = 400g
 * 10 health regeneration = 360g
 * Total calculable stats = 760g


 * 40 ability power = 870g
 * 30 attack damage = 1080g
 * Total calculable stats = 1950g


 * 4% attack speed = 133g
 * 4 ability power = 87g
 * Total calculable stats = 220g


 * 20% attack speed = 666g
 * 10% life steal = 440g
 * 10% spell vamp = 275g
 * Total calculable stats = 1381g


 * 72 ability power = 1566g
 * 30 attack damage =1080g
 * 52% attack speed = 1733g
 * 10% life steal = 440g
 * 10% spell vamp = 275g
 * Total calculable stats = 5094g

Now, we could also add a section for total stats with 8 stacks if need be but that's a bit long for this discussion section. 16:06, May 24, 2014 (UTC)


 * Let me pull up the analysis to Guinsoo's that I had added, as it will better explain what I'm trying to get to:

The whole cost analysis is split into three sections: one giving the gold value of the base stats (unchanged from what it is now), one listing the gold value per stack and one listing the gold value of the Surge passive. Following each sub-section, there is a line indicating the combined gold value of that passive in full effect plus the base gold value, and at the very end another line lists the item's maximum gold value with every passive at full effect. All of this is accurate, because the gold values listed are specific to the situations they cover. The cost efficiency section also covers this, indicating the starting value of Guinsoo's stats, how its gold efficiency ramps up and how its gold efficiency varies depending on the passives. It covers every contingency in just four lines, on one of the items in the game with the most possible contingencies. It is far more informative than the current cost analysis, which doesn't even address any of these conditionals, while remaining just as accurate. I wish to apply the above to every other item with a conditional: aura items would list the value of the aura per person, and indicate the gold value of a 5/3-person aura, item actives would list the value of the item's stats while the active is in effect, and this could combine in cases of AoE item actives like the one on.

I also wish to contest the following points: --Willbachbakal (talk) 18:25, May 24, 2014 (UTC)
 * Temporary Statistics: This is an argument from personal incredulity, which is a logical fallacy: just because an item looks ridiculously gold efficient while it is in use does not mean that such a massive gold efficiency is incorrect. A health pot gives 10 health regen while it is in effect, which is worth 360 gold and therefore makes the item gold efficient while it's giving you health. The data is correct, the interpretation is clearly not. Similarly, the fictive items created by ClariS also follow this model:
 * An item that grants 300 armor against small monsters grants 6000 gold's worth of stats against small monsters. Since this assertion only holds in the specific situation that you are taking damage from a small monster, it is completely accurate.
 * An item that gives 1000 health regen and 1000 mana regen for 1 second upon spawning gives 96000 gold's worth of stats for 1 second upon spawning. It does not matter if you spawn at full health and do not make use of these effects, what matters is that the item's giving you that many stats in that specific situation. This is a problem of conflating an item's gold value with its effectiveness, a fallacy that was resolved two threads ago.
 * The gold value of each stack while it lasts is 2.67 gold, causing the total gold value of the sum of the stacks to go up to 2670 gold. The feasability of this cap is not salient to the gold value analysis, again because gold value is not a measure of an item's effectiveness.
 * That "really stupid" item's passive would be worth [insert massive amount of gold] while you're dead. Yet again, gold value does not measure an item's effectiveness, merely the stats it provides at a given time, therefore the fact that this item has a massive gold value only while you're dead is neither wrong nor contradictory to anything.
 * Refusing to accept 's gold efficiency at the precise moment when it provides 200% movement speed is, again, an argument from incredulity. As with the above, the cost analysis never claims to say that the item is always that gold efficient, nor does it assert that the item's gold efficiency translates to as much in-game power. Again, this is a misinterpretation of gold efficiency. It is still worth keeping because it informs readers of how much gold's worth of stats they're getting at the specific time when the item provides that many stats, something all cost analyses should be doing.
 * On a side note, auras tend to be massively gold efficient when around other players, and for the above reasons I think that's a perfectly valid thing to say: in fact, it has very real repercussions in-game. Rioters keep talking about how auras are actually massively gold efficient (and don't feel as powerful as they are) and thus can't be made too strong, and supports frequently pick because of the massive amount of stats it provides when near the rest of your team. This is one of the cases where an item's gold efficiency very clearly correlates to its effectiveness in-game.
 * How long can you ignore these mechanics before they become a glaring issue?: The mechanics you mentioned are not actually a problem. The purpose of gold efficiency is to inform readers of the stats that they are getting from an item at a given time and how much these stats are worth, period. How you choose to interpret this information is up to you, and you clearly have been misinterpreting the data on quite a few articles. The real question here is: what makes you doubt that an item can be so massively gold efficient in a specific situation?
 * What reasons do you have not to apply the same rationale to every item with and active and/or passives?: Truth be told, if I could accurately measure the gold value of every passive and active in the item roster, they would have been added months ago. As much as I'd like everything to have a precise gold value, there are some effects that I don't have the tools to measure, since they do not hinge on stats and are thus uncountable in terms of gold value. The same goes for certain stats such as flat armor or magic penetration, which are present on some items but whose measurement would include too many variables to get an accurate value. Percentage penetration could get assigned a gold value, considering the only items that feature it are advanced-tier items whose only other stat is the one featured in their components, but doing so would not make much sense since those are the only two items to feature such a stat.

Missing stats
As for the existence of stats and effects without a current gold value, like armor/magic pen and 's active, I think the current gold efficiency system already accounts for that, to the tune of "This/these stat(s) and/or effect(s) must be worth at least X gold for item Y to be gold efficient" or "Item Y is X% gold efficient without this/these stat(s) and/or effect(s)". Since it already manages to be accurate and informative in its current state, I don't think it needs to be adjusted in this respect. --Willbachbakal (talk) 07:47, May 24, 2014 (UTC)

Comments

 * As I mentioned in my post in the last thread and in this thread the "% gold efficient" part is a big problem that should be removed, not worked around. The big problem is that the entire scale we use is something we made up.  Like I mentioned in the last thread, what we calculate the base stats off of, most efficient or least efficient, does have a factor in these sections.  It's fine for just listing gold values on stats to compare to but it messes up whether or not an item is gold efficient, granted it's only by a small amount but it can still be the difference between whether an item hits that big imaginary benchmark known as "gold efficient" which people take as more important than it really is.  In addition to the stats themselves there's things like uniqueness in stat combinations and slot efficiency that we can't calculate and do affect whether an item is "efficient" to a player and not including these but still using the terms we do is, as I keep saying, misleading and something that we shouldn't be doing as a fact and stat-driven site.  15:45, May 24, 2014 (UTC)


 * A lot of your comment is based on a misinterpretation of gold efficiency: the system in place cares only about the gold value of the stats on an item, and its sole purpose is to inform the reader of these. Whatever it cannot measure, it takes into account as an effect that either can make the item gold efficient, or that doesn't prevent the item from being gold efficient on its measurable stats alone. We cannot calculate the increase in effectiveness brought on by stat combinations or slot efficiency, nor should we attempt to, as these are not variables that gold efficiency sets out to measure. What you identify as misleading is simply a misinterpretation of data that sets out to achieve goals completely unrelated to what the reader is looking for.


 * Nonetheless, this raises an issue of clarity. The gold efficiency comes with a good definition, but needs to define it more strongly and more tightly, and so does the gold value/efficiency tooltip in the cost analysis sections. In the worst case, we should even add a disclaimer that specifies that gold value is only based upon an item's measurable stats, and not its effectiveness or impact on a build.


 * I strongly disagree that percentage efficiency is an arbitrary stat. It is a comparison of an item's gold value, i.e. the worth in gold of all of its measurable stats, to its price, and thus is accurate for the stats it measures, particularly as the gold values of said stats are determined from the items themselves. Since the benchmarks are the least gold efficient items, which represent the highest baseline cost for a stat and thus has a one-to-one representation with the associated cost, the standard of 100% gold efficiency serves to measure an item's gold value in stats relative to its cost. If we were to switch from the least gold efficient items to the most gold efficient items, we would just slide the gold efficiency values along the scale, and the standard would thus shift to a value under 100% gold efficiency. The comparisons that you'd be able to make between items now would be the same as after the change, and the entire system would remain internally consistent. Again, this standard is not a measure of an item's effectiveness: items that are gold inefficient on stats alone tend to either have an associated active or passive, conditional stats that make the item gold efficient in certain situations, or a synergy effect with other items, examples of which would be or . Again, the fact that these items are gold inefficient in the current framework is an indicator that there are effects to them that make them gold efficient in the right conditions, and this would be the same if we shifted the whole of gold efficiency to one end or the other.

--Willbachbakal (talk) 20:49, May 24, 2014 (UTC)

This is like arguing between two political parties--the opponents seem absolutely mental but neither side budge in their position. Will and Shaw have said everything and more I'd like to say regarding this. I am sure opponents of gold efficiency feel that its supporters are equally ridiculous, but I would like my vote to count for keeping gold efficiency.

Blaisem (talk) 05:29, May 25, 2014 (UTC)

The Willbach issue
I've figured out the big reason why these discussions can not move forward and I've decided to call it "The Willbach Issue." Thinking back there were some other users who acted this way but none so good at it as Willbach here. He completely blocks further discussions on an issue by not only being unmovable in his opinions but also by not just stating his opinion and letting that be that. Every single post in every single thread he has to reply to with mountains of text that are 90% full of him either dancing around with words or repeating himself from other posts. There's no room for actual discussion because he derails every post by making it about him, his opinion on the subject and how everyone else is wrong not only on a factual level but a conceptual level, they just don't see the issue in the right way, the way that Will sees it. He basically becoming like an ulting Galio in the middle of a jungle pathway, taunting all who come near to stop what they're doing and fight him in an uphill battle.

The solution to this problem, as I see it, is to realize that all of these posts come from one person, you don't have to reply to every single post of his, hell you don't even have to reply to one. The points he brings up are pretty much the same points he brings up in every post, just worded differently. Also realize that a discussion isn't about winning or losing or us vs them, despite how much Will wants to make it so, it's about the sharing of ideas and collaborating to make a better final result. As is Will is not only stymieing progressive discussions but also derailing all potential discussions on issues with the way he not only talks to other users but the way that he just takes up more page space than anyone else with his replies. 22:03, May 24, 2014 (UTC)

Comments

 * I'm really saddened by this. Of all the people here, you were the one I thought was least liable to make personal attacks. This is completely out of line, especially coming from an admin.


 * I'm not making this discussion about me. None of the posts here have been about my person, and I have taken great care to justify my viewpoints with structured argumentation and reasoning. I may be the most vocal member of this debate, but I am not derailing it in the slightest. All of the sections here, save for this one, are clearly relevant to the issue at hand, and I am proposing changes that could both move this discussion forward and help the wiki as a whole. Please also note that these changes were discussed in collaboration with another user (an admin, in fact, whom you can contact for confirmation). I have shown in my above posts that, although I have strong opinions about the subject, I am willing to compromise on certain aspects (a disclaimer for cost analyses, for example).


 * You accuse me of repeating posts, yet a lot of arguments you and others like you put out had clearly been answered in said posts. The issue of diminishing returns on resistances, for example, had been addressed multiple times, yet it was Deshiba repeating himself without taking heed to contrary arguments, when his position (that armor and MR provide diminishing returns) had not only been disproven multiple times by my points, but also by the health, armor and magic resistance articles themselves. It is your side that is guilty of repeating themselves without paying attention to opposing viewpoints and arguments, not mine.


 * Moreover, the statement that I have not moved the discussions at all is completely untrue. Some of the points I helped develop:


 * Gold value/efficiency being a measure of an item's stats and not its power.
 * Gold value/efficiency not determining an item's effectiveness on champions.
 * Gold value/efficiency not being valid as the sole determiner to an item build.
 * Gold value/efficiency having strong in-game justification and informativity more than sufficient to keep it on the wiki.


 * All of these have become the majority opinion, despite your objections. I have clearly contributed to these discussions and moved them forward. On the other hand, there is a mountain of evidence showing some of your proponents, namely ClariS, Deshiba and even TehAnonymous, acting in a completely uncivil way, derailing discussions, making them about their person, and generally contributing negatively to the conversation. The post you just added is more fuel to the fire. I can pull quotes for each user on demand. It is unnerving that so many of these people are in positions of responsibility and power, where exemplary conduct is a must. If you have a problem against me, I would rather we move it to the relevant thread or message wall, since putting it in this discussion really would make part of it all about me. A personal attack like this does not belong in a discussion you yourself wanted to remain civil. The discussion was, right up until this point, civil, and it should stay that way. I'm playing by the rules of the game, and so should you. --Willbachbakal (talk) 22:57, May 24, 2014 (UTC)


 * Armor and Magic resist have diminishing returns without them 100 armor would give 50% resistance all the time, so 200 armor would give 100% resistance. Read what | Diminishing Returns are before claiming they don't come into effect. I was repeating myself because you keep spreading information that is WRONG. Even by the examples you yourself provided it is the exact definition of what diminishing returns are, which is a mistake that, to this day, you refuse to admit. Man up. Deshiba, the Nitpicker 13:34, May 25, 2014 (UTC)
 * Please read this. Your post is clear proof that it is you, not me, who refuses to acknowledge other people's points, and are more concerned with saving face than moving the discussion forward. Don't take my word for it, or even my math, just click the link and read the section. I did not create it. Your assertion that resistances have diminishing returns is incorrect, as clearly stated by the section I linked to you: "What this means: by definition, armor does not have diminishing returns, because each point increases the unit's effective health against physical damage by 1% of its current actual health whether the unit has 10 armor or 1000 armor." I want to work with you, I really do, but you also need to work with me. I'm not your enemy, or even your adversary, I'm someone who wants to improve this wiki, which is presumably what you're trying to do too. --Willbachbakal (talk) 15:00, May 25, 2014 (UTC)
 * Again, you trade Gold for Armor to achieve Damage Reduction, the amount that Damage Reduction increases becomes lessened with each point of Armor you trade. The person that made that video and the person that added it to the Armor page are mistaken because they take the effect it has to the Health pool as the return. If you increase the Health pool however, the effect instantly shows up, where 1000HP*300AR=3000EHP;2000HP*200AR=4000EHP, proving that increasing armor does not equate to an equal return in EHP.
 * I would also like to point out that I've already admitted wrong doing in the last discussion for some harsh words and apologized for that. But that only caused me to get attacked personally even more, and harassed on my wall outside of the discussion. Your claim of me wanting to save face is absolutely false, and will not get me to give up on the purity of data which is getting corrupted in this case. Deshiba, the Nitpicker 15:31, May 25, 2014 (UTC)
 * Again, your calculations are wrong. 300 armor gives 75% damage reduction, so you get 4000 EHP from 1000HP. 200 armor gives ~66.7% damage reduction, so you get 6000EHP from 2000HP. You calculation is missing the 100 you need to add to the armor for the equation to work, a value clearly indicated in the section I linked. The proper calculations would be 1000HP*(100 + 300AR)=4000EHP;2000HP*(100 + 200AR)= 6000 EHP. 300 armor increases your EHP by 300% of your base HP, 200 armor increases your EHP by 200% of your base HP. The people who took the time to make those sections are therefore entirely correct.
 * The question you need to ask yourself here is: is every percentage point of damage reduction worth the same? Not worth in terms of gold value, but worth in terms of the protection it gives you. The answer is no, and can be easily demonstrated: a flat +1% to damage reduction does not have a consistent effect. Going from 0% to 1% damage reduction means you take less damage, but going from 99% to 100% damage reduction goes from making you very, very tough to literally invincible against that damage type. The amount of damage you can take goes from a finite amount to infinity. Because of this, one arbitrary unit of protection translates to a smaller and smaller percentage value the more damage reduction you have, and eventually reaches a value close to zero, never quite pushing the whole of damage reduction to 100%. The impression of diminishing returns you're getting from damage reduction is, therefore, an illusion, as the smaller percentage amounts you're getting from armor represent the exact same amount of protection. This is a disproval by counterexample, and is therefore valid irrespective of opinion and point of view. The only way you're going to get your way out of this is simply by admitting that you are wrong, or simply not posting in this discussion. Insisting that you are right will only further discredit you. --Willbachbakal (talk) 15:54, May 25, 2014 (UTC)


 * As someone who's posted here seldom, I am shocked to see this kind of aggressive, personal attack coming from a mod.  I find it extremely unprofessional and unfitting behavior for a mod.  Whether you agree with Will's methof of expressing himself or not, there are many better ways to communicate it and pursue a positive course of action.  A Wiki requires healthy guidance from its leaders.


 * Blaisem (talk) 05:07, May 25, 2014 (UTC)

A Potential Issue of Falsifiability
This is a point I'd like to address because a lot of what I've been suggesting would tend to look like gold value/efficiency would be impossible to disprove (namely, the baseline values of stats not being too important to gold value as a system), which is one of the most easily recognizable indicators that a system is bunk. I'm surprised that this hasn't been brought up earlier by the opposition.

To confirm that gold efficiency, as an evidence-based system of informing readers as to the gold's worth of stats they are receiving on an item, is falsifiable, here are a few examples of what would instantly falsify gold value/efficiency: While bringing up potential issues like these looks like I'm shooting myself in the foot, I think this is necessary to prove that gold efficiency is not just a theorycrafter's pipe dream. It runs on good assumptions, is consistent with itself, and manages to deliver the information it seeks to find accurately and correctly. It is a system that is based on evidence and solid math, and is therefore capable of providing concrete information to the wiki in its current incarnation (plus the sections that got removed). I invite others to find ways to falsify gold efficiency, and will argue which propositions would genuinely provide a point of falsifiability and which would not.--Willbachbakal (talk) 21:21, May 24, 2014 (UTC)
 * Variable stat values: Gold value is based on the assumption that one point of a stat will always have the same gold value, regardless of what item it's on or what condition it's bound to. That's what allows us to identify the gold value of stats from basic components and extrapolate to the rest of the item roster. If a Rioter came along and told us that one point of attack damage was worth, say, 36 gold on but 50 gold on, the entire gold value system would fall apart: every item would have to have its value determined on a case-by-case basis, with nothing to base ourselves on (or, at least, a set of rules for gold efficiency completely different to the one we're using). Since there is no evidence to suggest this yet, gold value is sound in this respect.
 * What about AD and attack speed/crit chance, or health and resistances? Don't they multiply off of each other and increase their value?: While the synergy between these stats certainly makes them more effective, effectiveness is not a quality that gold value seeks to measure. Even though some stats work better in tandem with each other, their gold value remains the same. Otherwise, the gold value of a stat would also completely depend on the scalings of every champion and every potential build out there, which would be completely inconsistent.
 * Inconsistent gold efficiencies: One of the assumptions that gold value makes is that the gold efficiency of an item is dependent solely on its countable stats and its price. Ignoring uncountable effects, an item that offers comparatively more stats for its price than a another item that offers less of the same stats relative to its price would be the more gold efficient option. Suppose there was an item that simply gave a certain amount of a few stats, and another item that gave the same stats in the same or lesser amounts for a greater price. Neither has effects that gold efficiency does not assign a gold value to. Gold efficiency would immediately tell us that the first item would be more gold efficient than the second. However, if a Rioter were to tell us that it was the second item that was, in fact, more gold efficient, there would be a fatal contradiction to the gold efficiency system, as it would be inconsistent on its own assumptions. Since no such situation has occurred, gold efficiency is sound in that respect.

Comments

 * Well, I can think of a few points:
 * How about the magically more inefficient AD items when the cost of a got changed a while back? AD used to be valued based on a 400g Long Sword, now it's valued on a 360g Long Sword even though the cost of almost all the AD items remained identical - hell, even . Although on that note, life steal magically became more efficient in that the identically priced Vamp Scepter now had a larger portion of its cost redirected into LS rather than AD under gold efficiency analyses. By this token,  suddenly became more inefficient because the gold value of the 100 AD took a massive hit, whereas the value increase of the lifesteal was small in comparison. This despite the fact that all items that built out of Long Sword had their prices increased to compensate for the component price decrease, thus none of the items building out of it became cheaper - finishing your Black Cleaver is an equally good gold investment as it was previously, but according to gold efficiency, it isn't. What happens if Riot decide to do a similar thing with Amplifying Tome? Do almost all AP items become much less efficient?
 * "If a Rioter came along and told us that one point of attack damage was worth, say, 36 gold on but 50 gold on, the entire gold value system would fall apart." - given that we don't assign a value to ArPen or MPen on the wiki, this is sort of already true. Ah, you might say, but this is just a result of these stats being difficult to give a gold value, the AD is still valued the same, we just can't value the ArPen! Well... is this not indicative of there being a flaw in the system somewhere when the items Riot themselves have stated are far too gold efficient are just given the tag "gold inefficient - passive must have X gold value for the item to be efficient"? (I know that's not how we phrase it exactly, but you get my point, I hope.) Every single pen item in the game is inefficient according to the crude analysis of "gold efficiency" or "gold value".
 * On that note - a lot of the Advanced, Legendary or Mythical tier items are "gold inefficient" according to stats alone. Gold efficiency is really only useful at all for items that are just bundles of stats with no major actives or passives. Aside from basic tier items, that's actually a surprisingly small number of items in the game, thus I'd dispute that gold efficiency is really a worthwhile stat for us to religiously calculate and then list on item articles as "x% efficient" (to decimal places, no less!) with just a note saying "passive/active must have Y value for the item to be gold efficient".
 * At the very least, I think we need to reword how we phrase things. "% gold efficient" seems bad to me: "Value of calculable stats" or similar would be infinitely preferable. Remember, we want to provide information to new players, not just the advanced players who also like their theorycrafting. To me, saying something "isn't gold efficient without its passive" - and giving no hints as to how valuable that passive is - does not help us in that regard. Listing the gold value of the stats we can calculate though, seems fine to me. If people are really obsessed with a %effectiveness measure, Google Calc will then quickly answer their question - Unevent (talk) 12:47, May 25, 2014 (UTC)

--Willbachbakal (talk) 15:36, May 25, 2014 (UTC)
 * The issue of components changing their gold value is actually evidence for gold efficiency, rather than a point of falsifiability. It initially sounds ridiculous, but here's the reasoning: we both know that, before and after the price change, remained exactly the same item. It kept the same stats, the same effects and the same price. You'd think that there would be no difference, but the critical difference here is that the stats on the item were not worth the same amount of gold before and after: because the price of a Long Sword decreased but the price of the Black Cleaver did not, you're paying relatively more gold for the AD you're getting, hence the drop in gold efficiency. To put it another way: if the cost of a beer can decreased but the cost of a six-pack remained the same, the six-pack would suddenly become a lot more expensive relative to the price of individual cans, despite no change in its price.
 * It is true that we do not know the value of penetration stats. That's why we don't assign a value to them or even presume a value. "This stat must be worth at least X gold to item Y to be gold inefficient" means that we don't know how much the stat is actually worth, but if it's worth X gold or more the item would be gold efficient. We're also saying that the item is <100% gold efficient on the stats we've measured alone, and since every item like that is usually accompanied by one or more additional effects, the "missing" gold efficiency is pointed at those, whose value we cannot accurately measure but for which we can indicate a minimum value for them to make the item gold efficient. Also, variable stats would mean the same stat would be worth different amounts of gold on different items, which even when accounting for the gaps presented by penetration has not yet occurred.
 * While it is true that a lot of items with uncountable effects tend to be gold inefficient on stats alone, gold efficiency is itself merely a measure of an item's countable stats relative to its price. Even when it does not give a value for the special effects on a top-tier item, the gap in its gold efficiency is itself informative of the effect's importance relative to its stats., for example, is extremely gold inefficient on stats alone, and that is an indicator of how much its Spellblade passive factors into its value. This is accurate, too: if the item had no additional effect and just the stats it listed, it would be completely gold inefficient as it would not be giving your gold's worth of stats. It is because these items have effects on the side that they are more than worth the gold investment.
 * The current cost analyses already list the total gold value of an item's stats, so it's already providing that information. The percentage efficiency value simply compares that value to the item's price, a calculation that is both simple to do and informative enough to warrant inclusion (see the bulletpoint right above this one). It does not give the value of effects it cannot measure, and I don't think it should unless it has concrete information regarding the effect's value (a value given by Riot, for example), as that would otherwise make gold value and efficiency abritrary.