Arcanaville

Custom critters use a special temp damage scale. Its supposed to be roughly comparable to what the critters would do if they could slot damage enhancements. In fact, minions use exactly the same temp damage scale for melee damage that players use.

In absolute terms, at level 50 a scale 1.0 melee attack will do the following amount of damage for a minion, Lt, Boss, EB, and AV custom critter in the AE:

Minion: 107.09
Lt: 160.63
Boss: 267.72
EB: 374.81
AV: 481.9

Note this is what standard critters deal with a scale 1.0 melee attack:

Minion: 120.48
Lt: 180.71
Boss: 385.52
EB: 481.9
AV: 874.56

Interestingly, standard critters have higher effective damage modifiers. In fact, from Boss and higher the modifier is a lot higher. So why are custom critters more dangerous offensively? Because they have *our* powers, and our powers are actually a lot more dangerous.

For one thing, there is a design rule that says that except for special circumstances, no critter deals more than scale 2.0 damage in one attack. No one: not Statesman in the LRSF, not giant monsters on monster island, not even Hamidon has an attack that deals more than scale 2.0 damage. Whereas we players have lots of those attacks. Total Focus deals 3.56. LRSF Statesman's total focus-looking attack deals 2.0. Energy Transfer deals 4.56. Be thankful custom critters don't get literal assassin's strikes, which deal 7.0.

So while the strongest AV attack for conventional critters is going to deal about 1749 damage, for a custom critter wielding energy transfer that will be 2197. And the devs usually aren't giving AVs powers that buff damage by a lot. Rage provides a 40% damage buff. Heck it used to provide 80%, but the devs eventually divided all damage buffs by two to approximately equalize them with player buffs**. Still, rage boosted Energy Transfer for a custom AV deals 3076 damage. That is Lord Recluse while buffed by the towers kind of damage, and he's usually +4 higher than the players when he hits for that much. This is an even con AV hitting for that without being externally buffed. If you faced this AV at +4, that attack would hit for 4430.

The ultimate, though, is Ninjitsu. To simulate being able to crit from hide, Ninjitsu Hide grants a 100% damage buff to the critter. So an SS/Nin custom AV that lands KO Blow (3.56) while Rage is up (+40% damage) and while Hide is unsuppressed (+100% damage) will land 4117 points of damage in one attack. This is like four yellow mitos hitting you simultaneously. You only make SS/Nins when you just plain want to kill the players, period, unless you're not using Hide and Rage from those sets. This of course ignores the old stacking bug Dispari mentions.



** I argued that since mission maker powers are supposed to simulate player powers, and critter damage scales are supposed to simulate critters enhancing powers like we do, damage buffs should be diluted to about half strength like ours are. When we buff damage by +80%, that is usually on top of our ~ +100% damage slotting. So we go from 200% to 280, an increase of 40%: 280/200 = 1.4. So the critters, who don't really have that enhancement, should get half the benefit from damage buffs. Before they made that change, an SS/Nin custom AV could have landed a 4800 point attack on a player. That comes close to one-shotting a scrapper at the resistance cap.

I believe it probably is against the vast majority of strong solo builds. Things like Level 54 Master Illusionists are things I stayed away from for a number of reasons, not the least of which was the primary target for the Scrapper Challenge was scrappers, and the combination of melee ranged debuffing and ultrahigh psionic damage would be problematic for most scrappers. Also, I specifically avoided things like filling the entire mission with mostly bosses.

Also, the scrapper challenge was specifically intended to help me practice building custom critters, so its full of those. I know there are some really nasty standard ones. But it seemed a bit unfair to stick things like Hamidon in there (until they took Hamidon away from me: a lot of things not in the AE are not in there because of me, no kidding).

I think a couple of things could take it on, none of which I currently have. A Dark Armor tanker with a soft-capped defense build could probably do it (there are such builds floating around). A really powerful Willpower build could probably do it (my DB/Will isn't 50 yet). And I'm guessing a perma-dom Mind control dominator could do it. The ingredients for a strong build to take that mission on seem to be: strong to smash/lethal/psi, very high defense in general, mez protection, and either ultra high control or enough health to take the occasional very large hit.

Interestingly though, I'm not sure if yours is harder than mine verses teams. If you scale the Scrapper Challenge to +4, the difference between the two might equalize verses an eight player team. Still, anyone who manages to cruise through the Scrapper Challenge should definitely take that on: its definitely a step up in difficulty.

I think the one I'm working on now is substantially harder than Aeon's Army Extra Hard, although its hard to say this early. Its *intended* to be. My rough target is for the one I'm working on now to be significantly harder than AAEH at *even con*, and scale up to team-crushing difficulty at +4.


Edit: I should clarify what I mean by saying I'm not sure if AAEH is harder than the Scrapper Challenge vs a strong team. Against a solo player, the goal is to present enough different offensive challenges that the player cannot counter them all: one leaks through and they die. But against a team, you have a very different problem. A team is going to turn the tables on you and present your mission with a wide range of offensive challenges that can simply neutralize them, making them unable to deploy any of those neat offensive tricks you've placed into the mission.

Against a player, you're thinking of giving your critters high offense, exotic offense, strong debuffs. In a team setting, you have to do that *and* simultaneously protect your critters against chain knockdown, superstacked mez, extreme tohit debuffs, mega sleeps, nukes, etc. Right now, my soft-capped SR scrapper testing the mission with invulnerability turned on is showing melee defense between 38% and 34% - down from 47% and with 95% DDR**. Even FF bubbles are going to have problems here. But if a team of eight Earth controllers just earthquakes everything in the mission, most of those debuffs go away, along with most of the offense, and it becomes a cakewalk.

I could just literally give everything status protection, but that's boring. I would basically be doing the powers equivalent of making everything AVs. So while I will have to include some status protection or the mission becomes controller snack food, I'm trying to do so in a measured way - even in a "killer" mission. That's the challenging part. I want it to kill super strong teams, but in an interesting way.

I really hope to have an alpha version of the mission set to lower than max difficulty ready for people to play around with Sunday or Monday. I just have to figure out where to publish it: I don't want to depublish any of my missions just yet and all three slots are full. I'll probably have to publish on a secondary account, which means making sure an alt on that account has everything I need to publish the mission. The alpha version is going to probably be weaker than the final version while I think through all the ways a strong team can neutralize it.

I *am* coming to the conclusion that there is no such thing as a threatening minion or LT to a team. To a solo player, yes. To a team, everything will likely be bosses, EBs, or AVs most of the time. An LT is too easy for a team to vaporize.


** PS: standing there letting them pound on me, the critters managed to take my defense down to a record low of 26.5% defense, down from 47%. Punching through DDR, that's 210% defense debuff. But its unlikely any team will see that, because I doubt everything will be able to stay alive long enough to see it. Either the critters go down, or the team goes down, before the debuffs stack that high.

In terms of powers and abilities, no. The one thing I do consider out of bounds, and only in the sense of being trivial rather than being unfair, is to just fill a mission with nothing but unkillable AVs.

Actually, there is one effect I've struck from my mission as being unpalatable even for a mission of this difficulty: foe intangible. I don't intend to cage the entire team indefinitely, which is theoretically possible. That would make the mission difficult in a trivially annoying way.


The problem is less figuring out a way to do that, and more of a way to make sure the mission *appears* winnable while still being that. Even for an ultra killer, I think the mission should look like it could be beaten in theory. A mission that people take one look at and say "screw that" is less interesting than one that people say "I wonder if we could build to defeat that?"

Herding will not be a good idea in the updated mission.

The main bosses job is really to stay alive for a while and trigger the ambushes: the ambushes themselves are the real threat. I like the idea of "constellations" of critters creating the threat rather than one big target. The right set of bosses can be nastier than any standard AV, but in a more interesting way - I think - than just one big target.

I've tweaked up the main minion group for the killer mission, and I'm starting to work on the main LTs. I'm going to split up some of the threatening powers between them because minions run out of endurance too quickly. One thing about designing nasty critters: if you give minions too many powers ironically they can become less threatening because they tend to burn up all their endurance fast, and not have enough to ever use some of their stronger powers.

Also, you have to be careful balancing things like summons. Make too many critters spawn off gang war or mastermind pets and you can blow the aggro caps, making most of your threat stand around doing nothing but be targets. I'm trying to dial that in.

That's just for the minions and Lts that will be spawning in the background. I'm still thinking about the main threat. I'm cranking up the difficulty on those by an order of magnitude, and using some of the stuff I normally don't use because its usually too nasty for the normals, only the crazies would want to face them.

Even with bosses off, beating it is still not easy. Its certainly a far harder challenge than the RWZ challenge.

As to ways to make it harder, its actually been tweaked downward many times. I removed excessive confuses from it, and I took out smoke when too many players noted that without +perception, it turns the thing into an impossible mission. At one time there were a lot of fire guys in there with fire damage auras, but those were bugged to do double damage, so they got removed also.

I'm using this mission as the foundation to make the Ultra Killer Mission for this thread. I've been rebuilding the minions. They are already plenty scary, before I've added the upgraded LTs. And believe me, I'm using all my knowledge of making nasty critters from the ground up: the minions aren't throw aways.

Yeah, the names are not very tricky. That's going to change in this mission. They won't deliberately lie, but neither will they say what the thing is so blatantly.

If I'm lucky, I might have an early alpha of this thing up this weekend some time. Depending on how many ITFs I end up on, of course - still got those two badges to work on.

PS: Glad you enjoyed the mission. Its not intended to blatantly assassinate players. It is intended to be beatable by a strong enough player or set of players. But its intended to make people think it took effort to do it, even for strong players.

PPS: I gotta fix the contact. Its not supposed to be a boulder, but the devs keep patching things in ways that alter my missions in weird ways.

I believe the tentacles have a damage aura that isn't typed. I don't know if that is intentional or not. Next time I'm in Apex I will try to check it out.

In this case, you are 100% CONFUSED. I said nothing about the NA EULA. I responded to Shadowe who said that the English statute doesn't say its limited to only discrimination that occurs within England itself. I said it doesn't say that because it doesn't have to: all laws in England only affect the jurisdiction in England. The EULA between an English player and the EU company NCSoft does business as *is* within the jurisdiction of England. But the terms between NCSoft and anyone outside of England fall outside its jurisdiction. English law cannot dictate what those can be.


Apparently your definition of the word "likely" is as fluid as your definition of "nationality." This case wouldn't even make it to court to be laughed out of it.

My goal would not be to make a mission that tried to outlast the players. That would be boring, and there are trivial ways to do that. My goal would be to make a mission that killed the players, in tricky ways. That would be more entertaining and more in the spirit of the challenge.

By the way, the Scrapper Challenge mission actually has some design rules in it, which I set for myself to test a theory. I theorized it was possible to make an extremely difficult mission without resorting to the conventional way the devs do it: by scaling numbers, especially health, aka the big bag of health syndrome.

So the rules were:

1. No AVs. The mission actually contains no Archvillains.

2. No indestructible critters. None of the critters in the mission is designed to be ludicrously difficult to kill. They are all individually defeatable, but extremely dangerous.


This time around, I would use AVs. Even very tough EBs would probably not be strong enough to take on the strongest possible teams out there, with purpled builds and alpha shift. But the goal would be to make a mission that was dangerous in an interesting way. The Scrapper Challenge is dangerous in I think an interesting way. If you want to see what I consider an interestingly dangerous mission, that's the place to start.

Was it from actual attacks, or could it have been from a damage aura? If it was from the ranged attacks, I don't have an explanation. If it could have been from a toxic damage aura, I think there might be ... an issue.

Japan is under electrical power rationing and many areas have rolling blackouts to maintain electrical power. Between that and the other disruptions in the country, some online gaming facilities and data centers in Japan have chosen to shut down completely in light of this. NC Japan is stating that they are going to continue operating their Japanese data centers, but at reduced electrical power and with reduced staffing, in order to continue to provide services to their customers.

Given the fact that this is a one time deal, I think it is reasonable for the devs to expend effort to figure out if there is a reasonable way for EU players to merge NA and EU accounts if they own both. Of course, I'm going to want a way to earn passport in exchange, but that's a small price to pay.

Do you actually believe NCSoft, a Korean company, cares more about North American people than Europeans? NCSoft, a Korean company, folded the tents of the Korean expansion of City of Heroes when it failed to attract enough players, without a second glance backward. If it were North American players and servers being merged into the more popular European version of City of Heroes, NCSoft would have not blinked an eye in changing the less popular NA players and preserving the state of the more numerous EU players.

You can bet if there were three hundred thousand subscribers in India right now both the NA and the EU players would be getting merged into their server farm and we'd all have to learn to advertise the ITF in Hindi.

Also, Atlas Park would be even more bizarre than it has ever been.

The legislation doesn't specify that because its presumed: English courts do not have jurisdiction over people outside of its territory nor of conduct that occurs outside of its territory.

However, Part 2 Chapter 1 of the act specifically lists the protected characteristics that the act protects from arbitrary discrimination as defined in the act. Those characteristics are age, disability, gender reassignment, marriage and civil partnership status, pregnancy and maternity, race, religion and religious belief, sex (gender), and sexual orientation. The act further refines the term "race" to refer to color, nationality, and ethnic or national origins.

Neither the account name change nor the global handle change explicitly target any particular color, nationality, or ethnicity. It treats all persons with accounts on that system identically, and in particular does not distinguish between nationality. The act doesn't specify continental location as a protected characteristic.

The act also creates an exception for indirect discrimination, which the renaming system would theoretically be. It states that a provision or practice is discriminatory if it cannot be shown to be a "proportionate means of achieving a legitimate aim." It does not have to be "fair" as people on the forums have been defining fair. It only has to be shown to be aimed at a legitimate purpose, and the practice must not be excessive. Merging the authentication systems is a legitimate aim. Changing the EU accounts that collide is a proportionate means of achieving that aim.

The merge therefore fails to implicate one of the specified protected characteristics, and fails the definition of indirect discrimination as defined within the act.

The optimal attack chain problem can't be solved with linear equations, or any equations for that matter. It has to be discovered through search: given a set of powers P1, P2, P3... look at all possible sequences Pa-Pb-Pc... and find the one with the highest damage within a fixed amount of time. There are ways to reduce the scope of the search that eliminate options that are extremely likely to be bad ones, but it does ultimately come down to searching them all. Given a set of six attacks with an average cast time of 1.5 seconds, there will be 6^10 possible options to look at for a 15 second window, and 6^20 for a 30 second window. That is 3.7 quadrillion different combinations (3.7 x 10^15) for the 30 second case. A brute force search that checks a million combinations per second would take a hundred years to complete.

I will note that I believe the optimal attack chain problem to be NP-complete, because it appears to be a variation of the Knapsack Problem. That means as a practical matter there is no solution to the problem of finding the optimal attack chain whose running time doesn't rise more or less exponentially (mathematicians will correctly nit-pick this as only requiring super-polynomial time). If it is essentially a knapsack variation as I suspect (if I'm not writing another searcher, I'm not proving its NP-complete either) brute force search with some heuristic modifications is likely to be the best we can do.

Your algorithm essentially generates a chain based on a heuristic guess: it presumes that heuristic actually generates the optimal chain which it is not guaranteed to do. Heuristic guesses are usually in the ballpark, but are often suboptimal which is why some people have thought about doing it the hard way: actually searching the entire attack chain space for the best of all possible attack chains.

I would say that the fact you don't have to succeed in the Incarnate trials to make significant progress in the end game system, you just have to participate, makes it more inclusive. You don't have to win, you just have to play. And the more you play, the more you advance, the more Incarnate power you acquire, the easier it will be for you to eventually succeed.

For some people, the lack of a guarantee of success will itself be unpalatable. But that doesn't mean the system is less inclusive. It may simply be less appealing to people who must always succeed.

You don't know that. I don't know that either. But the danger in wildly guessing is that people never seem to be aware of any possible way of doing things except the way they thought of, even when there are real world examples of situations that would make their hypothetical incorrect.

Here's the far more likely possibility.

There are two account databases, one for the EU servers and one for the NA servers. This makes sense since at one time the EU servers were actually in Europe, and if the authentication server for those servers were in the US an internet disruption could cause the EU servers to be impossible to log into, even if they were otherwise reachable and functional.

NCSoft has decided not to operate two different authentication servers any more, and are collapsing them. Unfortunately, both server login name and global handle are not just required to be unique, they are designated as keys in the database. EU is being exported and imported into the NA server, because there are less EU records. The EU records are being checked against the NA records and when global handle or account name collide, the EU records are being modified before importing. This cannot happen in reverse because the NA records are already in the database, and key fields cannot be changed in the NA records without a special process occurring. They can be changed in the EU records because those records are being exported and then imported. To change the NA records, the NA record would have to be specifically changed by special process. This would take significantly more time and effort**.

I don't know if it *is* like this, but this should come as no surprise to any dba out there. This is so obvious that anyone not taking this possibility into account cannot possibly have a background that makes their technical suggestions credible.


You specifically responded to someone saying:

The fact that you are interpreting that as a trivial reply tells me you are in fact unaware of the possible complications to this task that could make your perspective that the decision to change EU accounts was purely arbitrary completely false. The bottom line is that you're inviting comments like this by making assumptions about the mechanical process that, whether you realize it or not, are neither obvious nor reasonable. You assume capriciousness because you are unaware of any other possibility. Those other possibilities exist, which makes your conclusion faulty.

The fact that you are in error is not in and of itself the issue. The issue is that you seem so certain about something you are wrong about. That certainty breeds incredulity.


** Obviously globals can be changed and so can account names because it actually happens, but the process isn't necessarily one that could easily be incorporated into a bulk automatic transfer and import.

Which nationality are you claiming is being discriminated against under the Act? I know the legal system in the US has issues, but if the British system would honor a claim of discrimination against everyone in the country I need to tune in to Law and Order UK more.

The problem with all legal arguments for discrimination is that even if you were to accept the notion that EU City of Heroes players are a protected class of person, in every country CoX operates NCSoft is not specifically discriminating against any one group of people over another which means there isn't a local claim of discrimination possible, and even within the EU as a whole there is still no discrimination against any subset of EU citizens.

If you can find a law that states a private company must treat Europeans as a group identically to US and other citizens, and a jurisdiction that can apply that law to NCSoft, then you might have grounds for a claim. You'd almost certainly still lose, but you'd at least have legal grounds to make the claim in the first place.

The irony about all of these pseudo-legal claims of discrimination is that they are actually counter-productive to their own cause. So long as NCSoft changes *all* EU accounts that create a collision, its *impossible* to run afoul of any EU discrimination law. However if they change *some* but not *all* EU accounts with a collision, then the precise manner in which they do that *can* subject them to a discrimination claim in theory. They'd still be crackpot claims, but legally viable crackpot claims. These same arguments claiming NCSoft is discriminating against EU players can easily be turned to claiming that changing them based on, say, veteran status is a form of discrimination because it potentially targets the economically disadvantaged because they are less likely to have long term sustained subscriptions. Or using a first-come first-change global system discriminates against the employed. Or using any complex system discriminates against the elderly, who could be more easily confused by such a system.

These are all, I should emphasize, silly assertions. However, they have stronger merit than the claim of ethnic or national discrimination because they actually have one property the EU-wide claim does not have. It actually *mentions* a *subset* class of person being discriminated against. The EU-wide claim is essentially claiming that the class of person being discriminated against, within the jurisdiction being discussed, is everyone.

However unfair or unwise the NCSoft change might or might not be is a customer service issue. It is not a legal issue. Pressing the legal issue cannot help, but it can backfire. No one is going to find a law that says a Korean company has to treat US and other citizens no differently than EU citizens when it comes to a server merge. But I wouldn't bet against someone finding a law that on paper says changing by veteran status is legally questionable. I wouldn't actually take that bet.

There is a notion that the best way to advocate is to throw the kitchen sink at the issue, hoping that something sticks. There is no cost for deploying a losing argument. That's false. Advocacy often falls victim to its weakest argument.

One day, when I decide I want to take on the huge workload required, I have an idea for evaluating attack sets and defense sets by doing exactly that: creating a model that requires that combination to both survive and defeat a scaling critter group. The complexities are huge. But it factors in a few things we currently do not: overkill on offense, for one thing. Optimal chains actually aren't necessary the optimal way to defeat a group of targets. A more accurate AoE model for another, and a proper survivability model that factors in decreasing incoming damage. I fiddled with that in my scrapper comparison simulator, but that was a very simplified linear model.

At that point, you're now simulating the actual game and not just its mechanics.

Although, these reactors were apparently bought from the US.

Its easy to overgeneralize, but this I can say from first hand knowledge. Japanese culture emphasizes duty, and that tends to cause the Japanese to not cut corners as much as you might find in other cultures. On the other hand, it also causes them to be less likely to admit failure or ask for help immediately. Just like every other culture, Japanese culture has its strengths and weaknesses.

When the chips are down, the true professionals transcend culture. A culture of risk taking and pushing the envelope to keep up with the west was partly responsible for the Chernobyl accident, but the reactor operators, engineers, and fire and safety crews risked and in most cases gave their lives to prevent the accident from getting worse. I don't believe any culture changes that sense of duty when people think other people's lives are in their hands. They don't always do the right thing, but they do it selflessly.


Having said that, regardless of the earthquake magnitude models, the lack of a plan for portable backup electricity and alternate sources of cooling water still seem like odd preparedness errors to me. And I'm not just saying that idly. I was actually involved in a risk assessment of a nuclear power plant where one of the things I specifically focused on was the emergency cooling system. I believed that even though that system was not necessary most of the time (if ever to that point) it needed to be protected to as high a degree as the reactor itself. Prior auditors failed to make the same recommendation. I have no idea why.

I wrote a program to do exhaustive search a long time ago. Exhaustive search takes a surprisingly long time even for 60 second and 30 second chains. I tried to bound search scope with heuristics, but those started to get more complicated than the searching algorithm itself, and I decided at the time I didn't need to know the optimum chain that badly.

One of the heuristic checks though was something I thought would eventually speed things up a lot, except I never got around to finishing the code for it. Basically, it was a loop detector. If you're looking for the best 30 second chain, there's no point in looking at the chain A->B->C->A->B->C->A because the powers just keep repeating. That chain is just A->B->C. So A->B->C->A->B->A is a valid search item, but A->B->C->A->B->C is not. That part of the tree could be pruned.

Furthermore, I tried to see if this rule made sense, and under what circumstances. If you think about it, if A->B->C->A->B->A is a good chain, when why isn't A->B->A->B an even better chain? In the above chain C is following A and B, but A is also following A and B. Assuming recharge is not the issue, it cannot be true that A is the best choice sometimes and C is the best choice at others when confronted by the same situation. So the rule I was trying to create algorithmicly was this one: once you've decided to pick X in a certain situation, you are required to pick X whenever that same situation recurs. However, that turned out to be a sticky thing to code algorithmicly, due to the complexity of defining "situation" in a way that would work. It has to be based around the limits of recharge, but its not necessary that the same situation involves identical time to recharge. It could be different, but still close enough to make the powers become available on the same schedule.

It was specifically at that point that I decided I didn't need to know that badly. I've already written *two* simulators. I wasn't all that crazy about writing a third.

Even if you could, which you cannot in this circumstance, the treaty in question affects member states. Unless NCSoft joined the European Union when I wasn't looking, that treaty does not apply to NCSoft. It applies to the member countries of the EU, which are obligated to enact local laws consistent with the treaty. NCSoft can only run afoul of those local laws, if they exist and are relevant to this action. They can no more violate 2000/43/EC than they can violate the Treaty of Versailles.

The treaty in question does refer to discrimination of the kind recognized by most discrimination laws: gender, religion, ethnicity. As the NCSoft change affects all EU citizens equally without regard to age, gender, ethnicity, or religion, it would fail the basic test for being a covered type of discrimination which means no member state is required to enact laws which would cover the right of all EU citizens to equal protection of global chat handles relative to citizens of countries within North America. You'd probably have a better chance of arguing this to be an illegal tariff. But honestly that only converts the argument from nonsensical to ludicrous.

That is a completely arbitrary point of view. You're defining the only important groups of players to be NA and EU players, and anything that spreads things out between the two groups is not discriminatory.

If I choose to divide the players into two groups: recent subscribers and veteran subscribers, then changing only EU accounts spreads the changes out equally among recent and veteran subscribers.

You self-identify as an "EU subscriber" so you believe that set definition deserves special consideration. But its a completely arbitrary choice to afford EU subscribers special protection from inequitable discrimination higher than all other groups of players. But if you don't think there are players who self-identify as veteran or recent subscribers and have differences of opinion on what should the equity be between the two groups, then you were not here when veteran rewards were announced.

Except for the case of changing all accounts, all options discriminate. Its just not illegal or unethical discrimination. We also discriminate against people who don't pay subscriptions. But short of changing everyone, all other potential changes discriminate against someone. And if we change everyone, we're saying we're willing to inconvenience everyone unnecessarily to satisfy the fairness doctrine of a few. You get to pick your poison. Target some and not all, and you're discriminatory. Target all and not some, and you're unnecessarily inconvenient. Whichever one you pick, there will be a lot of people who will know in their hearts that you are an idiot for picking badly.

Of course, one of those groups is unreasonably wrong. The other one.