Arcanaville

If I understand what you are asking, then no.

Suppose you have a global accuracy buff. When you fire an attack, the game presumes that your accuracy is the accuracy of the power plus the accuracy slotted into the attack for the purposes of that attack. So at that moment in time, the game will calculate the total accuracy enhancements, apply ED, then add your global accuracy buff, and *then* apply DR to that final total, and then that will be the accuracy of the attack.

The same thing would be true for Damage. Both Accuracy and Damage work the same way as things like Tohit, the only difference is that Accuracy and Damage are situational based on power slotting, whereas tohit generally isn't.

Put it this way: you *never* add a number affected by DR to another number affected by DR. If you were meant to add them, you were meant to add them and apply DR at the end (of course, that's per attribute: I'm not counting things like subtracting a DR-modified defense from a DR-modified tohit).

So if you are doing calculations and you are applying DR to a number, and then adding another number, something has gone wrong. Of course, DR-modified numbers are *multipled* all the time: buffs are multiplicative to their base value, so when you have +35% damage that is a multipler: base * 1.35. But as far as I can think of at the moment, adding two DRed numbers implies you did something wrong.

That might be the simplest rule of thumb that helps reinforce when DR occurs in the calculations, at least to the best of my knowledge.

The way DR works is that it works on totals. This can be a bit confusing, so an example would be helpful.

Suppose I have two tohit buffs on me: Tactics and Aim. In that case, the buff from tactics and the buff from aim are added together, and the total is subject to DR. The net result is the net buff I get from both.

But suppose both tactics and aim are slotted with tohit enhancements. In that case, the enhancements themselves are *also* subject to DR. But because the +tohit enhancement slotted into Aim is separate from the +tohit enhancement slotted into Tactics, those slottings are hit by DR separately.

So you add up all the tohit enhancement slotted into tactics, and apply DR. That becomes the net enhancement that tactics "sees." You do the same thing to the tohit enhancements slotted into aim.

Then you take the buff from tactics - with the DR reduced slotting - and the buff from aim - also with its DR reduced slotting - and add them together. Then you apply DR to the final total, and *that* is the buff I get.


It all sounds complex, but the rule is: DR affects totals. You total up all the enhancements slotted into Aim, and DR hits that. Does DR hit the value of Aim (with slotting) itself? No, because that value is an incremental buff that could be added to other things. DR only hits the number at the end. If it helps, ask yourself this: what is the buff applied to. In the case of enhancements the buff is applied to the strength of the power itself. So you total up *all* the enhancements and apply DR. What is Aim's tohit buff applied to? A player. So you have to add up *all* tohit buffs for the player before you apply DR.

That's why you do not apply DR to Aim's numbers itself. You are actually applying DR to *the player's net tohit buffs* after they receive *all of them*. Contrawise, you apply DR to Aim's slotting after you tally up all of Aim's slotting. DR affects modifiers as those modifiers are applied to their target.

Hopefully I've explained that clearly enough.

Actually, I never said Tankers would never get Regen. I said it was potentially problematic for Brutes because Brutes are balanced around fury generation and the devs probably saw giving Brutes QR to be problematic for that reason. Now that Brutes have Willpower, that objection doesn't really hold water any more, so its theoretically possible that Brutes could get Regen at some point down the road.

I said it wasn't a good idea to give Tankers a directly proliferated Regen because of the alpha strike issue. *If* Tankers get Regen, it would almost certainly be a retooled Regen that factored this in. Since that might be something they would want to think about how to do in the best possible way, Regen for Tankers might be a semi-permanent backburner project they haven't made a priority yet.

I did say, however, that Regeneration would not be overpowered for Tankers. Its a bit more complex for Brutes, but I wouldn't call it overpowered for Brutes either (since Regeneration's performance varies non-proportionately with incoming damage, whether you act like a tanker or not has a strong bearing on the overall performance you'll see from Regen in terms of survival: that's part of the issue with Brutes, which were not archetypally designed to function as tankers explicitly).

I still do. And I'm sorry but its you that are misunderstanding copyright law. The fact that you can state that a collage is a "derivative work" and "cannot be copyright protected" is trivial proof to me. By definition derivative works are copyright protected by copyright law. The only question is whether the person making such a work had the original right to incorporate the individual elements. If they did not, they would be in violation of the copyright protection of the elements.

As a legal matter: if it cannot be copyright protected, its not a derivative work because its basically not a "work" within the context of copyright law.


Wrong in two ways. First, a collage does not require copyright ownership in the underlying works, only permission to use which I have: the EULA grants that permission to use the works to create characters.

Second, even if I *didn't* have permission, that's irrelevant to the question at hand. Well established precident states I nevertheless own the copyright on the elements of the work I created that still exist in tangible form, and the specific choice and arrangement of the character creator elements is both separate from them and fixed in tangible form. "The aspects of a derivative work added by the derivative author are that author's property, but the element drawn from the pre-existing work remains on grant from the owner of the pre-existing work." Stewart v Abend, Justice O'Conner writing for the Supreme Court.

In that case, the author did not have permission to use the underlying work, but still has rights in their derivative work. In this case, the collage as the unique artistic arrangement of the elements is copyright protected, but the collage author has no copyright to the individual elements within the collage.

What I don't own is the copyright on the entire derivative work including the character creator artwork as a whole, but that was never in dispute because you can't use NCSoft's intellectual property outside of the game without permission either way, and that wasn't the issue. The issue was whether you can reuse your characters minus the artwork and without referencing protected elements of the game backstory. And the answer is still yes.

And I'm getting tired of quoting precident and basically hearing "nuh uh." This case cites a Supreme Court of the United States decision, and it cites the opinion of the court in reference to a case where the author of a derivative work did not, ultimately, have the underlying rights to elements contained within the work. Under those circumstances, the statement that the author of the derivative work nevertheless has rights to their own contribution in that specific case is not just relevant, its essentially legally definitive.

This discussion is quickly becoming boring, because there are no actual relevant legal points being brought up to argue against. Just statements of opinion that are in direct contradiction to the law.


Also:

I'm well aware of the legal principle of dilution. That is why I was very specific in my example to pick one where dilution was not an issue. Profit or revenue is not the issue, and that is not the sense in which I used "non-commercial" but rather in the sense of not being venture exposed to any member of the public. The school has lots of opportunities for dilution. My bedroom does not.

Intrinsically. Well I'll be darned: spell check agrees with you there.


If you're still out there, I have one recommendation; the same one I give to everyone entering a technical field (my undergraduate degree is EE for reference). Get your money's worth. College is not just about textbooks and classes, especially not in engineering. Its about working in labs for suck pay and terrible hours, its about figuring out what that really expensive piece of equipment will do when you push that button when no one is looking. Most importantly, its about learning to become a problem solver. The best thing you can do in college, besides not flunking out of classes, is to learn to be a good problem solver. And there will be *lots* of opportunities out there to learn to be one. Prof needs student hire for his lab? Special project going on somewhere on campus? Something will present itself and give you opportunities to increase your knowledge and experience.

Ultimately, your grades and that diploma can get you a job, but it will be the skillset you graduate with that will keep it for you. Make sure that when your college is handing out skills, you stand in line.


Good luck AC. And if you send me homework assignments, fair warning: I forgot Bernoulli's equation 18 years ago. Otherwise, my door is always open to questions from budding engineers. Plus, I might one day be flying a Boeing with your work in it, and I would rather the wings not fall off.

My characters, which include the sum total of their appearance, their backstory, and other written elements related to them, can be copyright protected as a whole.

Not only wrong, but *really* wrong. Derivative works *are* protected by copyright. The copyright law actually has a *section* on the protection accorded to derivative works: its section 103.

Section 103b:
In copyright law, the "threshold for originality" is presumed to be extremely low. *Any* demonstration of creativity, which would be trivial to demonstrate in making a character with anything short of the random button, would pass. Therefore, my assemblage of character creator artwork is theoretically copyright protected.


It almost sounds like your knowledge of copyright law comes from the doctrines espoused prior to 1976. Prior to the 1976 copyright act, different courts held different things when it came to derivative works, some of which said what you seem to be saying above. However, those legal theories were supposed to be extinguished after 1976.


Oh boy. Where to start.

* First: the visual appearance of a character *can* be copyright protected as a work of art, if its a specific fixed instance. What cannot be copyright protected usually are character designs. Those *can* be trademark protected. A character design that cannot be copyright protected would be something like "tall muscular guy with red and blue costume with cap and big S on the chest." That's a character design: it isn't a specific work of art, but it can be trademarked.

* As an aside, its possible to both copyright *and* trademark a work of art. Trademarks are just that: they are marks intended to identify a particular business or trade. Trademark protection *doesn't* protect against copying or usage per se: it protects against anyone else using the mark for commercial purposes within the same basic business or field.

* Can DC stop me from drawing pictures of Batman on my bedroom wall? No. My drawings are obviously copyright owned by me as the author of the work. I'm not using the character design of Batman for any commercial purpose, so I'm not in violation of their trademark. Ergo, there is no legal rights issue. If I draw Batman on my office building, I may be in violation of their trademark rights. If I blow up a cover of Batman and put that up on my office building, I'm in violation of their copyright rights (assuming DC owned those rights and the cover was a work-for-hire).


Copyright law is applicable in the case of the rights being discussed, because no one is claiming that any player nor NCSoft has ever taken out trademark protection on any character created by a subscriber.


It just occured to me that at least one person has a rather unique insight on this question. Castle created his character prior to becoming an employee for Paragon Studios. Technically speaking, by some people's interpretation around here, he had no rights to his character the moment he added it to the game as a player. I'm now curious to know if Castle was asked to specifically sign a copyright transfer document to that character when it was added as a contact to the game. Technically speaking, I'm pretty sure anything Castle makes as a result of being told to do so by his employer is covered as a work-for-hire creation, but I'm curious if the NCSoft legal eagles realized that a pre-existing character of Castle's would create a potential loophole to that agreement which they decided to close. The actual intent of the 1976 copyright law was to force people to always do that so that the courts wouldn't have to keep deciding fuzzy ownership cases like this. I'll have to ask him that next time I have a chance, although I'm not sure Castle would be comfortable answering contractual questions in a public forum (although this is a rather harmless one).

You cannot copyright a name alone, but you can copyright a character design which includes a name as a feature of the character.

How I assemble the costume artwork is considered a separate protectable design. That is why NCSoft specifically requests non-exclusive license to it: without that license they could not use my character design even if they own the individual artwork components that comprise it.

In any event, I mentioned that case not as a case that would be operative in determining copyright ownership, but as an example of author rights separate from copyright ownership, since some suggested that there was no such thing (the preface to the citation states this directly).

In terms of cases that might be specifically relevant to the issue of whether or not a player actually owns their characters, separate from any intellectual property components they might have incorporated into them, try:

In Dumas v. Gommerman the Ninth Circuit Court of Appeals asks and answers the question of what is required to be considered a work for hire under copyright law post 1976. In particular it lays out the history of the Copyright Act and the intent of the provisions of the act with regard to defining work for hire and the protections afforded content creators. I recommend those not familiar with copyright law to read the thought process laid out in the decision; its an excellent summary of how the various laws interact.

Both that case and Community For Creative Non-Violence v. Reid address the issue of whether its reasonable to consider players work for hire when creating their characters. Indirectly, they also state the requirement that copyright transfer must be written and signed by the original owner or author.

Now, is it possible to argue that the EULA supercedes the requirements of section 204 of the copyright act? Almost certainly not: courts have held that the intent of section 204 isn't a formality, such that something like an EULA could substitute for it. The intent of 204 is to protect authors from giving their rights away, *and* forcing others to negotiate with authors. See Effects v Cohen:

(emphasis mine)

It would be impossible to argue that the EULA creates the same environment intended by section 204, and in the absence of that courts would - and do - uphold the literal requirements of the law: that copyright transfer must be signed and in writing. Effects v Cohen goes so far as to say that it doesn't matter if an industry operates by a consistent mutually agreed upon standard that differs from copyright law: neither custom nor alternate contractual context supercede the law. In fact, court decisions and legal analysis consistently state that the *intent* of the law was to eliminate these types of options, specifically to eliminate questions about authorship. If the argument is being made that the EULA is asking for my copyrights in exchange for the privilege of paying to subscribe to the game, any intellectual property attorney will tell you that the 1976 act was explicitly intended to eliminate that type of deal-making. The 1976 act makes that form of agreement illegal to protect me from people who would try to make me agree to it. This isn't just guesswork: its a very explicit intent of the law (see the cases above).

One more for the road; this time a case a few more readers might have heard of, that both cites and then amplifies Cohen above:

SCO v Novell


It was mentioned above that perhaps NCSoft only asserts copyright on characters *created* with the game client: that if a professional author previously created a character it is protected by copyright, but if its first created in the game then it doesn't have "prior protection" and NCSoft can then assert ownership of it.

This was only true prior to 1976. Prior to 1976, the legal theory could be made that a player making a character in CoX could choose not to copyright the character, and then allow NCSoft as the publisher of the content to copyright the character instead, by mutual agreement. In fact, some state courts held this to be perfectly valid under the copyright laws of the time. After 1976, this legal theory becomes immediately invalid. After 1976, authors aren't required to copyright their works by overt act (such as publication or registration). The law recognizes them as owning the copyright on their creations immediately upon creation. In fact the law goes further and states that authors own the copyrights on their work during the process of creation. Because of that, NCSoft cannot ask me to forgo copyright protection: the law gives it to me whether I want it or not. The moment that happens, whether the character existed prior to me playing the game or not becomes irrelevant: in both cases copyright protection extends to me as the author first. NCSoft must now take it from me in either case: the law doesn't distinguish those two cases.


So far, I haven't heard anyone mention the "best" legal angle NCSoft would have to attempt to assert ownership of my characters, at least to the extent of preventing me from using them in commercial ventures, which I guess tells me no IP lawyers in the house. Attempting to assert the EULA automatically grants them ownership has zero chance of working. Zero point zero zero chance. Attempting to assert the characters are work for hire creations has an only infinitesmal chance of working: it probably has a better chance of being considered frivolous and subject to claim of attorney fees and court costs.

But the angle I think (and the intellectual property lawyers I've chatted with seem to think) has the best chance of succeeding (meaning: its better than 1%, but not great) is the argument that my characters are a joint creation where both myself and NCSoft are the joint legal authors. The problem with this argument is that joint authorship requires certain conditions and considerations which are very unlikely to be seen as occuring in CoX (or most MMOs). Basically, there is no consideration for me in this equation: I have no special rights to NCSoft's "contribution" to my characters separate from the one I'm presumably granting to my contribution, and there is no shared direction or artistic control. The relationship is so one-sided in terms of artistic contribution (I do all the work, NCSoft offers nothing in exchange except usage of its tools) that this would be a very difficult assertion to demonstrate.

If I were an attorney and was asked to pursue this sort of claim, this is the angle I would take. And I would give myself about a 2% chance of surviving summary judgement against me. And a one in a thousand chance of winning at trial, unless opposing counsel is a moron and the presiding judge is sniffing glue. And I don't think this has any chance of surviving appeal at all, unless the appellate courts feel like making new law that day.

No, you can't.


No way, not even close. Work for hire is not a colloquial term: it has a fairly specific meaning. While there is some grey area that has been litigated, I believe you would be laughed out of court if you attempted to argue that a customer *paying for service* was actually operating as an employee of NCSoft.

Title 17 Section 101 contains the definition of a work for hire for the purposes of copyright:


That's also unlikely, this doesn't make it past summary judgement unless I get a judge that is suffering from a brain tumor.


Except the EULA goes on to state that where that is not possible a different set of conditions apply.

Without that escape hatch, the EULA would be no more enforceable than if it were to say:

By submitting Member Content to or creating Member Content on any area of the Service, you acknowledge that you are no longer protected by the laws of the United States of America and can be summarily executed by NCSoft for typing your password wrong.

It is *literally* that ludicrous.


Technically, then, murder is not against the law in this country either.


This particular situation has never been tested to my knowledge, because no one is stupid enough to attempt to test it.


I don't know how you could have read section 204 and not read section 203. Section 203 specifies under what conditions exclusive transfer of rights can be revoked or terminated by the original author. These rights cannot be transferred contractually as with copyright rights themselves: it can only be passed on by act of law, which is essentially the case where it transfers to my estate or is transferred as part of my assets during a situation like bankruptcy.

But I cannot otherwise transfer those rights, nor can I be compelled to give them up contractually.


Once again: this is black letter law. Anyone asserting otherwise is manufacturing ambiguity where there is none. If I don't explicitly transfer my rights, Federal law doesn't recognize the transfer. NCSoft doesn't even have a case to make: they cannot say "I know what the Federal Copyright statute says, but my EULA says something different." Basic contract law: contracts specifying illegal or impossible conditions are automatically unenforceable.

NCSoft's attorneys are more aware than you are, because they specifically left escape hatches designed to ensure their EULA is *not* unenforceable under US law.


For those that are interested in this subject, this is the best relevant example I can find on the subject of author rights over and above copyright owner rights:

Marvel vs Simon. Although this judgement involves some fairly subtle legal reasoning surrounding the issue of estoppel (in this case, the Appellate court ruled that a settlement which includes a declaration by an author that a work was actually a work for hire *doesn't* preclude that author from revisiting the question of authorship rights when attempting to assert a right that didn't exist at the time he surrendered that right), it is noteworthy for demonstrating in an actual court case an author attempting to exercise the rights granted in section 203: namely the right to revoke an exclusive transfer of copyright by the author, overriding contractual agreements to the contrary. In fact, the appellate court said something more significant: it said that even if an author claims in a settlement that they did the work as a work-for-hire the author can challenge their own assertion when it comes to attempting to assert rights which did not exist for the author at that time. In other words, the court recognizes that the purpose behind copyright law (or parts of it) are to protect authors from being strong-armed into giving their rights away to cheaply (the decision states that rather specifically). As a result, the court's view was that if copyright law grants a right to the *author* and not just the *owner* an author has the right to attempt to assert their authorship to gain those rights, even if he was asked to surrender those rights earlier as part of a settlement agreement.

A summary of the decision is here.


Again: this is some of the blackest of black letter law: you cannot be asked to surrender ownership of copyright without a written and signed agreement unless you are operating as a work-for-hire, and work for hire always requires actual employment in some capacity as a mandatory condition.

The US is *not* one of the countries that allows for rights to be transfered in the manner described, which is why NCSoft does not have exclusive rights to my characters.

In the US, copyright is presumed to attach "at birth" meaning the moment I create the work. NCSoft cannot then demand either ownership or exclusive rights to my work because copyright transfer must be done specifically and in writing, and exclusive transfer of rights is considered to be a copyright transfer under the law. Section 204(a) in particular governs this, and it states that copyright transfer must be signed and in writing. Furthermore, courts have consistently ruled that 204(a) should be interpreted to mean that the transfer must be explicit - it must essentially name the works in question. The EULA doesn't come close to meeting the burden of 204(a) which is why NCSoft specifically asks for a non-exclusive license. Doing so prevents them from attempting to claim something Federal Copyright law forbids, which would make the contract invalid.

The only major exception to this is the work-for-hire exception. But no player of an MMO is likely to be seen by any US court as operating as a work for hire employee. Any attempt to assert that would border on ludicrous.


It is black-letter law that the assertion that NCSoft "owns" your characters is a myth. The only kernel of truth in this is that while I own the creative work I generated, I don't own, nor do I have a license to use, any of NCSoft's intellectual property. So while I can take my characters and use them elsewhere, I cannot use the specific artwork from the character creator, I cannot use the specific descriptions of the powers in the game, nor can I use any of the backstory elements that NCSoft owns the copyright on, such as the Rikti. I can only use *my* parts of the characters. In some cases, characters may be so encumbered by IP that the player doesn't own that it is impractical to use them elsewhere. But they do own them nevertheless.

And just to be 100% clear: if I made them up, I own the names of my characters for the purposes of them being a component of an artistic work I created. Whatever the EULA says, it would be against Federal law for NCSoft to attempt to assert they own them. Once again, they can only demand a non-exclusive license to use them themselves, which is why the EULA has that escape hatch in the first place.


Trivia: there are some rights that as the creator of a work I *can't* surrender, no matter what. That's because copyright law recognizes the creator as a separate entity from the copyright holder. Even if I were to sign my rights away for a creation, there are laws that provide me with some rights to reassert or reacquire my copyright rights in certain circumstances, no matter what I've signed. These rights are granted to me when I create the work, and they stay attached to me until I die.

The problem with statements like this is that they create cherry-picking opportunities; its not really a quantitative statement.

Here's another one: SR is the *only* mitigation set that does not have *any* means to mitigate an entire class of attacks (non-positional psionics), whether by direct damage mitigation or by attack debuffing. That means while people like to talk about "holes" SR is the only mitigation set that actually has an actual hole: one that you can look all the way down to the bare health of the character.

How do you counterbalance these two "only"s? Let me know when you come up with a way.


And how would you counterbalance that against the fact that Regen has far more endurance to burn on offense than sets like Invuln and Dark Armor? The game is balanced around endurance as a constraint, and Regen has a far lower constraint on offense intrinsicly.

The calculations for endurance constraint on offense are not difficult to do. Most people throw them out, however, because "everyone" can build around it. Which is another form of balance cherry-picking: counting what is perceived as a problem, while ignoring all countervailing effects that are not perceived as independent problems.

I've done this calculation myself, by the way: its not favorable to making a case for Regen.


Now keep going and compare SO SR to SO Regen, and see what the mitigation advantage actually is. I'll give you a hint: its not 15%. Since you got this far, I'm rather surprised you didn't go all the way and calculate by how much Regen "lags" the other sets, based on this theory of animation time exchange (which I'm not saying I necessarily buy, by the way: if that theory operated reasonably well we'd then have a theory that would allow one to exchange offense for defense and compute, say, the value of a damage aura relative to a defensive power, which is the Holy Grail of balance calculations; unfortunately, this doesn't work in the general case).


By the way (and this is not your fault) your Dark Regen numbers are slightly off. Dark Regeneration's cast time is listed as 1.17 seconds, but it actually roots for 40 frames: about 1.33 seconds (its one of those powers that is on my list of root discrepancies). With server clock considerations, DR's activation cost per cycle is actually closer to a whopping 5.3% all by itself (if you're lucky on clock alignment, its possible for it to occasionally be 4.8%, but that won't happen very often). My guess is that eventually that will be fixed downward, towards the 1.17 cast time, but at the moment Dark Armor is surprisingly "expensive" in terms of animation time, given it has only one click to deal with (it is, however, probably the singularly most problematic defensive click in existence).

Actually, I've never done that. I've already specifically looked at Regeneration compared to other mitigation sets in much more detail than the spreadsheet can: the spreadsheet was originally created as a power proliferation comparison tool.

According to the spreadsheet, Regen is actually *stronger* than most powersets out there strictly in terms of damage mitigation strength, for most temporal circumstances. Anyone who wants to factor in potentially countervailing issues is free to do so. Anyone claiming that the numbers "don't match the game" is also free to do so; however as was the case with the Scrapper Comparison threads, I'll point out that I've actually played essentially all of the defensive sets and thus I have at least a comparable amount of first hand experience with all of the sets as anyone else. If I observed a very strong discrepancy between my analyses and actual in-game performance, I would have noted them. My spreadsheet is *not* an analysis: its a simplified quick reference sheet for damage mitigation (and what-if tool) more than anything else. Within that context, its extremely accurate.

And I have, in the past, at least six or seven times, analyzed factors outside of direct damage mitigation. In particular, I wrote extensively about the role of endurance drain on survivability that is normally not factored into survivability calculations, so its not that I'm incapable or unwilling to investigate them.

Debuffs have never been a factor I've "ignored." Rather, I've explicitly considered them secondary factors in the value of a mitigation set**, in the same bucket with powers like Quick Recovery which also has a survivability and other benefits which are also non-trivial to account for, and which are typically presumed to be secondary advantages of the set outside of primary mitigation: things which are qualitatively and not quantitatively balanced for (because in CoH those things have no quantifiable model by which they are granted, unlike damage mitigation which had an implicit one even if the devs were not in possession of an explicit one).


** In the I4 version of the Scrapper Comparison thread, I posted a specific article on this subject detailing just exactly what the challenges are in attempting to quantify such things, and giving reasons for why quantifying them might not even be consistent with the design of the game at that time. Briefly, some things seem to be granted to some sets even *knowing* that at some level they are an unbalanced-for advantage, but since numerical advantages do not always translate to measurable performance (levelling) advantages, some things simply are better with the knowledge that its an acceptable advantage if its not performance-skewing. The (original) Claws modifications seem to be an example of this, as is quick recovery. Positional defense itself seems to fall into this category, at least in intent if not in implementation. The powers design seemed to be (at least at that time) analogous to trying to give everyone the same amount of vanilla ice cream, and then giving everyone a different set of topings on top without being all that careful to make sure everyone gets precisely the same ounces of toppings, so long as everyone's toppings are different enough.

Those guys are specifically there to present the choice of clearing them, or working very hard not to draw them. Or, if you are a maniac, trying to complete the challenge while pulling them into the fight. They can rapidly increase the difficulty exponentially with their sonic buffs and debuffs.

There have been a few players that have left comments stating that on even difficulty levels it was actually sufficiently difficult to clear the sonic engineeers themselves they left to find a better respec before actually doing the core challenge itself. And to be honest, if they are giving you trouble before the challenge starts, that's probably for the best.

All it takes to swap two powers around is a simple copy and paste. Actually, you don't even need to do that: you just have to change a couple numbers in the base spreadsheet that set the MinLevelAvailable setting (that's not precisely what its called, but you get the idea) for the two powers.

However, that's something that they have a lot of discretion for when they are copying a powerset to a new archetype, because that is actually a new powerset being created where there was none before. *Any* change is fair game if its justified. But changing power order on an *existing* powerset is something the devs *can do* but *will not do* unless there is some absolute proven necessity for doing so, because that would be a change to a powerset that has an established power order that players are already using and potentially relying on. You don't just get to say "well, my way is better: I'm sure they will understand."

The hurdle to overcome when suggesting power order changes on an existing set, as opposed to a new (even if proliferated) powerset is extremely high. To put it another way, you could probably sooner get Castle to make Brute Evasion into a damage aura than you could get Castle to swap the order of Dodge and Agile.

Just FYI: I've been testing adjustments to that arc off and on via the test server, but I haven't had the time to really do a completely new version justice. I'm finding its extremely difficult, under the rules of the AE, to make a version that hits that fine line of "really hard, but not impossible" that the current version does, in a different format other than ambushes (i.e. a more conventional mission).

I *think* its just the right level of difficulty that hard-core players who try it believe they *could* beat it with just the right level of effort/insps/temp powers, so they know its not mathematically impossible. And that's the key to making the challenge at least nominally credible.

For now, the ambush version is only version I've designed that is viable. The comments I get for it as people run it continue to be just about where I wanted them to be, namely just at the edge of "are you crazy?" and "I'll be back." A few people claim to have beaten it without too much difficulty, but those are rare outliers (A Willpower tank claimed to finish it without deaths or inspirations back when it was more of a smoke-fest).

On the subject of making more missions like this, I'd encourage anyone that wants to try their hand at it to give it a go. If nothing else, make one that specifically challenges your own build. But if you can make one that other people agree is a worthy challenge to undertake, and its possible to beat but not easy to beat, that's a great target to set for yourself. It takes a lot more knowledge about how the powersets work and how people play than it first appears, and its definitely a learning experience (mine is not version 3.X for no reason: I went through three major revisions and about ten minor ones between beta test and the live servers).

I don't think the game needs a "PvP dev" much less me as it. Moving forward, I think the main thing CoX PvP needs is a middle ground between the test server and the live game that can be used to test PvP evolutionary changes over longer periods of time than beta tests normally allow for, and to allow for changes to be propagated in a manner that doesn't instantly displace people from one set of rules to another.

A combination of a cross-server PvP arena plus an expansion of the arena settings could allow the devs to introduce PvP changes as optional settings, and see over longer periods of time which ones are more acceptable to the PvP community in general. And moreover you'd have the additional option to detect segmentation in the PvP players, in terms of whether different groups of players prefer different PvP "rules" than others.

Over time you could in a more graceful manner introduce and migrate PvP mechanics to something that both caters to the established PvP community while looking for changes that would make it more palatable to players outside that community.

This would lower the "cost" of adopting player-driven suggestions for improving PvP, because no suggestion would automatically orphan other PvP players and the devs would have a way to know if that player's suggestion is in fact acceptable to the players (because they would increasingly use that setting).

It is in many ways highly suboptimal, and probably more confusing, but it also seems to be the safest way to proceed given the current circumstances. The main point is to create a marketplace of ideas where the devs' ideas of how PvP should work and various players' ideas of how PvP should work would have a chance to compete on a level playing field, and hopefully open the door for more experimentation with player-promoted ideas.

In many ways, I don't think CoX's PvP problems are mechanical problems at this point. I think they are community-based problems. I think any solution has to address the player community more, and in a way that is still acceptable to the devs given their design requirements.

Fixed. I dunno why it did that. The forums seem to do odd things to my posts at times. Sorry for the incorrect reference.

So far, I haven't seen this behavior to a greater extent than random chance predicts, although the last time I checked was about a year ago.


Actually, if you make 100 rolls, and each roll has a 5% chance of being a "miss" then the odds of getting exactly 5 misses are only 18%. Without getting into the complexities of calculating standard deviation, if your expectation is that anything outside of 2 to 7 misses is "statistically unusual" then by my calculations the probability of seeing 2 through 7 misses in a sample of 100 combat rolls is 83.5%. The odds of seeing something else is therefore 16.5%, or about 1 in 6. In other words, by your definition, about one in six stretches of 100 combat rolls is going to be "unacceptable."

If I'm remembering my probability and statistics 101 correctly, this is the table for the statistical probability of observing N misses out of 100, if the odds of a miss are 1 in 20:

Even something "extraordinary" like seeing twice as many misses (10) will happen 1.67% of the time, or about one time in 60. My guess is that if you were to solo from level 1 to level 50, and somehow stayed pegged at 95% chance to hit the entire way through, you're very likely not just to have seen 10 misses out of 100, but to have even seen 10 misses out of 100 happen twice in a row back to back. You probably even have a decent chance of having seen the incredible 15 misses out of 100 (this ignores the effects of the streakbreaker which would complicate the issue of observing these results directly).


Now, if someone has a chat log of 20 combat rolls in a row rolling 0.95 or higher, even after simply throwing out the streakbreaker induced hits, that I would like to see. The odds of that happening are about 12 million to one against, which is possible (and in fact likely) across the entire playerbase, but would be comparable to winning the lottery.

If memory serves I came in second to Pilcrow in the second poll. I used to tease him that it was not unexpected that the copy and paste forum master managed to get the most votes in a forum poll.

I'm not certain they are, for a couple of reasons:


My memory of past measurements done by other people in trying to quantify archetype populations is that sometimes it generated significant differences between servers. Whether that represents an actual skew between servers, or simply the large statistical fluctuations inevitable in these kinds of measurements is an open question.


Ditto for the time, but moreover I think there is another potential for skew here that is larger, and again may skew significantly by player type. Player populations may have significant cyclical beats to them congruent with things like timezone and work hours. The problem here is that since some players play for less time than others, the more casual players might cluster more.


Ah, I should have elaborated. In my head I was wondering if placing alts in each zone and using /whoall might be a simpler way to collect the statistics than using /search, as /whoall would allow an automated script to capture the statistics from chat logs.


My lean is towards thinking that 24 people playing 24 blasters for one hour each represents a higher level of popularity than one player playing 24 blasters, for reasons similar to Zem's. You're looking at popularity in terms of representation: given a single player looking at the current character population, which archetypes do they encounter more. I'm inclined to look at popularity from the perspective of "for each player, which archetype do they prefer to play more." In other words, I think I buy Zem's "one player, one vote" position more: regardless of the amount of time that each player plays, if they play at all they get one vote. If someone plays blasters 90% of the time, and tankers 10% of the time, they generate 0.9 votes for blasters and 0.1 votes for tankers, regardless of hours of play.

I'm not sure that's the "right" perspective, but it feels "righter" than the time-weighted one to me at the moment.


Technically, I think the devs (at least theoretically) collect that statistic, although its not a stat that I have any insight into, except insofar as I believe the datamining they do intrinsicly contains that information, given what Castle said about blasters (that they underperformed in or out of teams).

Theoretically, we know how fast archetypes level: the devs hinted at it with the Defiance changes. Except for blasters (at the time) no other archetype could claim that all of its powerset combinations underperformed the average of all archetypes by more than some (admittedly unknown) margin. From a meta-analysis perspective, I can theoretically combine the analysis of the information mentioned above with the answer to this question, which the devs might be willing to answer separately:

Does X sxignificantly underperform relative to everyone else at level Y?


Of course, more information is better, but I was hypothetically thinking about having the opportunity to run essentially one report out of the dataminer.


This has a different set of potential skews, and I'm not claiming to have thought about them all carefully either, although I think it would be a worthy data collection project also, especially since it seems to be within the capabilities of a player (a crazy player, perhaps).

The real problem is you'd need to measure all of the servers, or at least several of the most populous ones, and most or all of the zones, at least several times a day, for probably at least a week, to get a reasonable sample. But there is some automation opportunities for the data aggregation at least, which might make it a less insane project than it first appears.


As to the question of "time played is all that matters" here's a hypothetical question, which I don't claim to have a definitive answer to either. Suppose we have exactly 2 CoH players: one plays a single Tanker 24 hours a day. The other has 24 blasters that they play for about an hour each, 24 hours a day. If we were to collect hourly statistics, we'd discover that at any one moment in time there was exactly one blaster and one tanker playing at all times: Blasters and Tankers would each represent 50% of the total number of characters. However, if we were to collect the actual *names* of the characters, and aggregate on a 24 hour basis, we'd discover that in the average day 24 Blasters and one Tanker were being played.

Now replace the single player playing 24 blasters with 24 players playing one blaster each for an hour. Statistically speaking, the two situations are identical (in terms of the statistics we can collect). My hypothetical question is: are these situations statistically identical in actual fact: do they represent the same fundamental information about the popularity of the two archetypes?

There's some very subtle biases that I think are embedded in /search-style analysis that I'm not sure are insignificant. The biggest one I can think of is whether its very subtly biased against casual players, and as a consequence underrepresents their archetype choices. Ironically, I think the only way to know if such statistical skews exist is to actually collect the data and see if they are present in the first place.

If I had access to the devs' datamining tools, the statistic I think I would most want to pull (related to this matter) would be total XP earned per archetype per day per level, normalized as a percentage of the total for that level. In other words, how much bars of XP is each archetype earning per day, summarized per combat level. Analysis of that information would probably reveal a lot about cumulative patterns of play.

I think buffs have only indirect effects on archetype populations, although those indirect effects can be large. Which is to say, someone that has never played the archetype before isn't likely to fully understand or appreciate those buffs, while those that are already playing those archetypes are, well, already playing them and represented in the statistics already. I think the main way in which buffs can increase an archetype's popularity is by word of mouth: other players playing the archetype can start recommending it more strongly to other players, or contrawise players can reverse bad impressions of them by observing their (improved) performance in-game.

Of course the small publicity-buzz that comes from being buffed can attract some attention to the archetype. But I think one of the conclusions of the statistics that is still probably valid today is that no archetype completely fails to attract people to play it: the larger factor to whether an archetype will be "popular" or not appears to be player-retention. And because of that, archetype populations themselves are driven by relatively slow forces.

I'm somewhat hopeful that the stalker buffs back then improved matters for stalkers. I'm somewhat more hopeful that the recent buffs to dominators will likely improve their retention ratio over time.

Everything in this post is potentially very much dated, at the very least for stalkers and dominators. My main motivations for reposting it were that a) people kept asking, b) people still reference it, and c) its the only place where BaB's original data can still be referenced, since his posts went bye-bye as well. If we are ever given updated data, I'll certainly update this post. As it is, though, I recognize that some of the conclusions and analysis here might be moot in terms of the current game, although they might still be interesting to people in the historical sense. Its also the best information we currently have, dated as it is.

In any event, while game changes may have changed the current situation, I don't think any new information has come up which invalidates the analysis in terms of its commentary of the state of the game at the time.

Actually, the activation penalty (separate from the issues described above - the delay of the effects, and the fact that they can affect a power without lasting for its duration) is higher than that. The issue is that while BU is casting you don't just lose its buff, you lose all of your damage entirely.

*IF* we're talking about full attack chains, so the cast time is a real cost to damage, then with BU recharging in 30 seconds its cycle time is about 31.17 seconds (the mechanics of arcanatime do not seem to increase recharge significantly). That means for 1.32 seconds out of 31.17 seconds (4.2%) your damage drops to zero.

Under this approximation, BU's total damage buff is, assuming 1.95 damage slotting:

[ 10 * (1.95 + 1.0) + 19.85 * (1.95) + 1.32 * 0 ] / 31.17 * 1.95 = 1.122

12.2%. Alternatively, that is 1.95 * 0.122 = 0.2379, or +23.79% damage compared to Assault's +10.5% damage. BU is about 2.3 times stronger than assault in terms of damage over time. But interestingly, neither is *that* great in increasing damage over time. BU is a burst monster, but not an especially great damage over time buff due to the cast time offensive penalty. In terms of damage over time, +25% damage strength from invention bonuses would beat it out (of course, the +tohit of BU is a separate significant benefit).


In a certain sense, BU is a +100% damage buff for 10 seconds, and a -195% damage debuff for 1.32 seconds. When you look at it that way, it makes it more clear just why BU's long-term benefit is not as high as it might first appear: the penalty is neutralizing about one quarter of BU's total benefit (of course, if your attack chain is not full and you only use BU when nothing else is available to fire, BU's cast time cost can be made very low or zero: it then becomes all buff and no penalty).