Arcanaville

I don't doubt Zwill is being told its a vBulletin problem. Because Zwill has to take their word for it.

Anyway, the problem isn't cookie expiration. The problem is that periodically the board software seems to think something has gone totally haywire with your user session, and actually sends DELETES to ALL your cookie information. Including the cookie data that houses your name and password, which is where its stored if you say "Remember Me." The boards should *never* delete that information unless you log out. So the boards must be thinking periodically that you either logged out, or it should log you out for some reason.

Although vBulletin had an issue related to this in the past, it never had it to the degree we do, and every version of the bug had a trivial fix associated with it. I'm guessing someone performed a custom maneuver I'd get modded for describing somewhere in the authentication or state tracking code, and mucked it up. For some value of M.

The only workaround I've found that actually works is to periodically save the cookie information, and when I see evidence of a cookie delete to simply rewrite the saved information back into the cookie. Then I'm back where I was before. I've been too busy to write a definitive workaround, which would be to create a transparent proxy that filters cookie-deletes from boards.cityofheroes.com so that cookie updates work fine but cookie deletes get thrown away. That would, I think, fix the problem. But only in a really annoying sort of way.

Changes to resurgence aren't being discussed to improve the performance of Willpower for Tankers. I don't think anyone believes Willpower for tankers needs any help whatsoever. Revive was being discussed in terms of improving the tanking performance of Regen. Within that context, a rez that dumped aggro is counter-productive to tanking, and a rez that taunted aggro without giving the tanker time to retoggle is suicide.


It defeats the purpose of any of my suggestions to backport them to other archetypes, because the sole focus of my suggestions is to differentiate the Regen set for tankers to be better suited for Tanking. Its not to simply buff Regen everywhere. Regen is fine everywhere else.

Most other sets *automatically* change when they are ported to other archetypes without the devs having to lift a finger. SR jumped from 30% defense SO slotted for scrappers and brutes to 40+% defense SO slotted for tankers. The devs didn't buff the set: tanker modifiers did that. Nearly melee secondaries have a sizable percentage of their strength scaled by increased tanker modifiers: they all have significant defense or resistance that autoscales. The only set for which you can say no large component of its mitigation would scale upward for tankers is Regeneration. That's just a quirk of how health buffs, regeneration buffs, and heals work. But even Willpower has substantial resistances and defenses that scale upward for tankers. Regen only has resilience, and that is only a small fraction of its total mitigation.

That's an unambiguous tanker scaling discrepancy.

At one time, meaning right after launch, Energy Manipulation was actually seen as the *worst* blaster secondary, at least by the more vocal proponents of blaster secondaries. However, it quickly became seen as the most all around good blaster secondaries: the jack of all trades secondary, particularly to support the two main factions of blasters: ranged blasters and blappers. You can blast with anything and you can blap with anything, but having said that the two best blapper secondaries were generally considered to be energy and electric, and the two best utility secondaries were generally seen to be energy and devices, and the best ranged support secondary was seen to be energy. When you show up on all three lists, that's a good secondary. Other secondaries have things that set them apart from energy: Ice with Ice patch for example, or Fire with all the PBAoEs, so its not like energy is the best, period. But in terms of general purpose secondary, energy tends to always be near the top.

You calculations have a fundamental flaw and its due to that "ordered to unordered correction factor" I mentioned above. You're weighting every possible combination of power pool selections equally, but they are not all equal. For example, in the calculations I mentioned above for the 8 power pick search (i.e. picks through level 14) I said there were 60,518 "ordered" picks. However, those were the sum total of all builds with zero power pools, one power pool, two power pools, etc. The actual number of builds with X amount of non-travel power pools and Y amount of travel power pools is given by this matrix:

where the first row is the total number of builds with zero travel pools and zero through four non-travel pools. Descending rows show total number of builds with one, two, three, and four travel pools and corresponding non-travel pools. Why I have to keep them separate is that travel pools have five powers, non-travel pools have four, so they create different possibilities in build sequences.

The total number of builds up to level 14 with zero travel and zero non-travel pools is only 9: that's because there are ten total primary and secondary powers available at level 14, and one of them you have to take (the first secondary). So you're taking eight of nine, and there's only nine possible powers you could drop.

Now, there's also only one way to have no power pools: there are no permutations to consider. But there are many ways to have one power pool, or two. And since I'm keeping them separate in my build computations, we have to consider all the combinations of ways to take zero through four non-travel and zero through four travel pools. The total number of combinations given six non-travel and four travel pools is given by a different matrix:

Since the build totals include the power pool constraint, impossible combinations are already zeroed out (as you can see from the first matrix). So the actual total number of possible build combinations for each possibility of travel and non-travel power pools is:

The largest number - 447,660 - sits atone travel power and two non-travel power pools, or three power pools total. In other words, if we are counting *all* possibilities, and not just "good" ones, then up to level 14 the power pool case with the highest number of possibilities is the case of three power pools: one travel and two non-travel. And the reason is basically that there are lots of power pools to choose from, and the total number of powers in primary and secondary is still low at level 14.

However, this is generalizable to higher levels. You cannot just examine the total number of possible ways to create a build with N number of power pools, because that presumes each N is equally diverse: it contains an equal number of options. But that's not true, and those numbers shift around as you change the constraints on the system like by adding or removing inherent fitness.


Incidentally, for anyone who cares:

Level 16: 48648888, 344529

The numbers are increasing so rapidly I'm now wondering if I will have to reoptimize for the search to complete in a reasonable amount of time. I did not expect to be searching fifty million build sequences just to get to level 16. Its entirely possible the explosive growth levels off as the number of options begins to fill the build space, but I doubt that's happening before level 30. By which time I might be searching billions of sequences and tens of millions of possible distinct builds.

Actually, as specified by Starsman, it would be a death sentence, since the rez would also immediately taunt everything in range. Without a Soul Transfer-like mez or temporary immunity from damage, rezzing in the middle of a fight and taunting everything is impractical. That would have to be addressed if a literal rez power was being leveraged in this way.

It would be easy to do, but you'd still be dead and still collect debt to use it. I think most people would go along with the notion that if Revive somehow protected against receiving debt, then its not really dying: its a bounce-back.

Also a real death in game terms would be potentially problematic for a tanker because they would instantly lose aggro. Dark Armor solves that problem by giving a rez that pretty much allows a tanker (and anything else) to mez everything in the room short of an AV with triangles up. Its a legitimate tanker act to die and pop back up with essentially the most powerful player mez in the game. You're still "tanking." But dying, dumping all aggro onto everyone else, and then having to rez and retoggle is probably less practical from a tanking perspective than Soul Transfer is.

The purple triangles offer protection against confuse, but confuse has the longest duration of any mez and a permadom mind dominator can stack a lot of confuse. Stacking mag 6 per application under domination, you can break the triangles with 9 stacks of confuse, which means you basically have to get your cycle time for confuse (which you can easily get below six seconds) more than 9 times shorter than the duration of the confuse (which you can easily get above 60 seconds for the non-domination component). Factoring in level differences, with enough recharge you can confuse any AV that isn't explicitly immune or highly resistant to confuse. In other words, don't try this on Reichsman in a 'cuda. On the other hand, it works against Romulus in the ITF.

Most AVs do not have any protection against sleep in the triangles, so a mind dom can perma-sleep an entire group of AVs with AoE sleep powers without having to stack quite as many applications as confuse. But the main thing is to make sure every AV is hit by the sleep, and stacked enough to cover an occasional miss.

So in the LRSF specifically, you mass sleep everything, then you stack up a ton of confuse on a single AV, and away you go.

If the justification for changing these powers for tankers is that tankers don't scale health, heal, and regen, then there's no specific reason to backport the changes to scrappers, brutes, or stalkers. They all share the same modifiers for health, heal, and regen. Backporting them would put us right where we started: with tankers having the same numbers as scrappers, and not scaling.

As to the dull pain change, at 180s recharge you could perma Dull Pain with DOs, which I think is not unreasonable for a regeneration tanker. This one change is compensating for the fact that a Regen tanker would have the exact same strength in fast healing, integration (regen), reconstruction, and instant healing as a Regen scrapper. Rather than scaling all those numbers upward by some tanker fudge factor, I'm focusing all of the improvement on Dull Pain which offers burst damage resilience that is most valuable, and therefore appropriate, to Tankers.

Keep in mind I don't think Regen is a bad set: I just think its problematic for tankers. And since there's a legitimate avenue for asking if the fact that nearly all of Regeneration's strength is not scaled by tanker mods relative to scrapper mods, it seems to make sense to me to try to kill two birds with one stone: acclimate regeneration to tankers and address a potential scaling problem simultaneously.

Technically, that would be identical to the situation where the Regen had 2500 health, and the regen had a power that self-mezzed them below 500 health (and their regeneration rate was 80% of normal). Of course, if you're held you can't use reconstruction normally.

Another strange way to get this sort of behavior is to change the way debt works so that debt is awarded ten seconds after you die instead of immediately. Then add 100% debt protection to Revive for ten seconds. If you revive within 10 seconds of dying, you get no debt, and dying without debt is essentially not dying.

However, this could easily be abused, so you have a refractory period when the debt protection cannot be reapplied, say four minutes, even if revive's recharge is lowered below that threshold. You'll rez, but with debt as normal.

A possibility I suggested to Castle in another context, but would require a significant UI change, would be to change reconstruction so that it is less all or nothing on recharge. Change reconstruction so that its base heal is 12.5% instead of 25%, and its base recharge is 30 seconds instead of 60 seconds. Then alter the mechanics of the power so that it recharges in 30 seconds, but if it is not used it continues to gain in strength until at twice its nominal recharge it is twice as strong. So using it at 30 seconds would grant a 12.5% heal. Using it at 40 seconds would grant a 16.7% heal. Using it at 50 seconds would grant a 20.8% heal, and using it at 60 seconds would grant the maximum heal of 25%. This would have to obey slotting in both heal and recharge, which would require new mechanics and new UI code, which would make it an expensive change. Not to mention it would require educating players about the new mechanics.


I still think, though, if you cut the recharge of Dull Pain from 360 to 180 seconds (allowing it to be essentially perma at low levels with slotting) and make the +maxhealth fully enhanceable instead of half enhanceable for tankers specifically, the set would have no problems at all. And both changes are nominally justifiable as a way to work around the fact that tankers have neither higher regeneration modifiers nor higher health buff modifiers. That's fine when regeneration and health buffs are not central to the powerset's performance, but the special case arises with Regen when nearly all of the set's performance is tied up in effects that tankers do not scale.

Maybe when they add a wrestling powerset.

Perhaps I should have written an optimized C search rather than just whip up something in python. I'm looking at the maximum number of possible builds with inherent fitness for each level. However, it might be a while before this program completes, since its currently only up to analyzing level 16, and its already examined about forty million different possible ways to select powers given a fixed primary and secondary. I didn't think there were that many.

Computations through level 14:

Level 1: 2,2
Level 2: 6,5
Level 4: 18,9
Level 6: 144,59
Level 8: 1386,342
Level 10: 15390,1833
Level 12: 191952,9425
Level 14: 2905002,60518

Those two numbers represent the total number of possible ways to pick powers - i.e. the total number of power selection *orders* - and the total number of unique builds that are generated by those sequences of power choices. Also, the numbers are low: to reduce the search space I'm only looking for ordered sequences where the total number of power pool picks in each power pool are sorted in descending order. So I have to then apply a correction factor to add combinations of power pools. The ordered to unordered correction factor increases the total number of valid builds through level 14 to 2,194,455.

You might think that includes a lot of weird builds, like builds that have all four power pool selections at level 14 or three travel pools:

Travel pool choices:
0: 555,681
1: 1,050,172
2: 504,954
3: 80,328
4: 3,320

Non-travel pool choices:
0: 13,539
1: 307,476
2: 851,085
3: 797,580
4: 224,775

(total number of *pools* used, not powers)

Most number of options: one travel, two non-travel pools: 447,660
Least non-zero number of options: zero travel, zero non-travel: 9

To be honest, I don't especially care how many builds options there are at level 14. What I am curious to know is at what level are the number of options maximal. In other words, is there an inflection point beyond which the number of possible different builds drops. I would guess that it would at some point probably around 38, but I'm not confident in that guess at the moment.

And interestingly: Level 14 does have less choices than Level 12 in one respect: number of build options with *zero* power pool choices. Level 12: 35. Level 14: 9.

I'm also curious to know the answer to the literal mathematical question: to what degree did inherent fitness reduce the total number of possible builds if any, and at what levels, assuming stamina and its prerequisites were taken as soon as possible.

That may take a lot of run time unless I rewrite this code, and I'm kind of busy right now and inclined to just let it run for a week or two. I'm actually a bit surprised no one has been crazy enough to do this yet. I mean, the computer is doing pretty much all the work: I just have to wait for it to finish.

This also doesn't count slotting.

A long while ago, someone suggested "reactive heal" for Regen: if you hit zero, a heal would fire instantly to bring you back above zero without dying, and this could happen once per some interval. The problem then and now is that the tech doesn't exist to do it. In fact, back then the best we could say was "the tech doesn't exist to do it" and now I can say "I'm not sure how you would do it even if the devs wanted to add the tech." "Dying" is a difficult thing to override.

Having it trigger before death is problematic in terms of picking just the right spot to do it that isn't too early and isn't too late, especially because its likely that this mechanism would have at least a quarter second delay in firing. The game engine doesn't really do "instantly" very well.


There was a related discussion back then about adding scaling regeneration to Regen, in fact before SR got scaling resistances. Although I didn't originate the idea, I did do the math to show it was possible to balance this, albeit in a complex way related to damage spikes. Here, the idea was that when health was high, regen was low so damage could easily bring health down. But when health was low, regen was very high, so health tended not to stay down there and bounce back up quickly.

Keep in mind this idea was formulated not long after toggle IH went away and before I5, so it was being balanced at a time when scrapper performance was still historically sky-high. Back then, the only thing Regen was *supposed* to be vulnerable to was quick burst damage: anything slower than that simply wouldn't be fast enough to take out toggle IH. The intent was to replicate some of toggle-IH's performance in sustained damage while creating a higher vulnerability to burst damage than being at full health with perma DP would imply. Doing that by granting super-high regen but only at low health meant Regen could have high alpha resilience (relatively) and high sustainable damage through regen, but not both at the same time.

For me, the precise way in which they have an advantage is significant. I tend to feel the opposite: I care less about the attuned enhancements, but I would care more about purples. For me, "pay to win" implies a significant advantage towards progress, and the exemplar effects do not directly accelerate progress. The amount of time the average player is exemplared makes this advantage normally minimal. The real advantage to me has always been the leveling *upward* of the attuned enhancements, which does in fact offer a progress advantage. However, that advantage is not one of power, but more of cost and convenience. At no time do the attuneds "outlevel" comparable IOs of comparable level, so at no time do they generate better results than theoretically possible in-game. What they do is offer identical results to the best possible results at all levels within their range. To me, that's a grey area, but clearly one where the store-bought IOs do not surpass what's achievable in-game.

The purples, however, are a special case. Their scarcity and (usually) cost prevents them from being acquirable at all in-game except through an enormous amount of in-game effort on a relative basis. This is also true, and sometimes more true, for certain specific IOs, but purples were *designed* to be this way, which makes them a special case. In that case, the devs would be selling a different form of pay to win: pay to achieve something that was explicitly designed to be difficult to achieve beyond the normal levels of difficulty in the game. It would be like selling Master badges. That has no affect on power or ability, but it is a variant of pay to win: its paying to bypass the difficulty curve for a special task.

To me, these are both grey areas, but I think the attuned Mako's walk the line, while selling purples would be hopping back and forth across it. I would be much more likely to object to the latter than the former.

Unless you sold them for an astronomical amount of points, and that creates a different set of problems.

We don't generally handwave away mechanical concepts except where absolutely necessary. Its one thing to hand wave away the scaling resistances, and another to hand wave away, say, adding cold damage to fire blast. Handwaving away scaling resistances in a presumptively regeneration toggle is more the latter than the former. Particularly because it isn't necessary: regeneration already has an association with +health, which offers comparable benefits to +res.

Also, the "concept" for shields is not Captain America: that is an exemplar, but not the concept. The concept for shields is shields: their version of what a shield can do. Powerset concepts do not have specific strengths or specific characterization packages associated with them. Reconstruction is not based on a specific ability of Wolverine.

In City of Heroes, there are two kinds of psionic attacks. The first kind has no physicality in the conventional sense. Its typed "psionic_attack" and would pass right through Shield defenses. The other kind has enough physicality to be evaded by defenses that work to either place the target out of the path of the attack or to place an obstacle in the path of the attack, typified by positional defenses. Shields work on the latter and not the former, which is a reasonable concession to the limits of the combat system in the game.

The scaling resistances were not my first choice either: I had long advocated +health in practiced brawler myself. The scaling resistances were sprung late in CoV beta, so they couldn't even be discussed openly on the standard forums. By the time I figured out what exactly what they were doing and posted a description on the beta forums, beta itself was almost over.

Actually, a part of me would be pleased to think the rules no longer apply to anyone, so that includes me.

Alternatively, a special invention converts all your temps into a single wand of wonder.

"Proved" is a bit strong. What I believe I can show is that if the nukes were treated like any other power, factoring in their long recharge and crashes, that the current power balancing rules would specify that they do something between somewhat more, and massively more damage.

Which means that the crashing nukes do less damage than that because of a special case rule that limits the maximum damage they can do. That's not unreasonable, but that has to be balanced against the question of whether we actually need to be reducing blaster performance, when blaster performance has always been datamined to be lacking.

My best guess analysis from all available numbers was that CoH had somewhere in the neighborhood of 70k-80k subscribers prior to Freedom's announcement. My guess is the current number of VIP subscriptions is somewhat higher now, but I have no idea if its higher by a tiny amount or a larger amount. I suspect that the number of Premium players is substantial, but I doubt we suddenly jumped to 200k active accounts this quickly.

Both quoted posts don't state, or even imply that builds are more diverse now. They both attempt to assert that builds are not less diverse now. The first one by Sam:

asserts specifically that in the days before ED, we may have had more theoretical choices for slotting than after ED, but as a practical matter players were funneled into "right" choices both by the way the slotting worked, and by the fact that the overwhelming recommendations floating around pushed people into them. He's analogizing that to the pre-inherent fitness period when there was the theoretical choice to take or not take fitness, which has been technically removed, but for a very large percentage of the player population that was not really a voluntary choice as they saw it. With inherent fitness, those people see four power pool choices available, not three.


Uberguy's quote is much less easy to misinterpret:

He *explicitly* states that we don't have less diversity. That does not mean he's stating we have more. Its not even an implication we have more. He's saying we do not have less.

I've been saying I do not believe inherent fitness significantly reduced build diversity. That does not mean I believe inherent fitness increased build diversity. It means I do not believe it reduced it. By in large, I believe it was more or less neutral.

ED + Inventions + power pool extensions + lowered level thresholds most definitely combined to radically increase the amount of build options available. ED by itself not so much, but ED combined with Inventions did, and Inventions added additional diversity options beyond realizing the options ED theoretically had but didn't generally exercise.

You said you could tank with it. I responded to that by saying I could also. I don't know where I'm required to then post a second post after that one saying, "having said that, here's a completely separate thought related to the first one but falling along a different line." I thought saying "but for me, the question..." captured that thought sufficiently well."

The devs could do a statistically valid survey of builds. Players would have a much more difficult time doing that. However, what a player can do is what I did earlier in this thread. First, guestimate the number of total player builds that took fitness, given statements by the devs that a large majority of players did when they made fitness inherent. That number would have to be much higher than 50%, but it is unlikely to be very close to 100% or the devs would have said words to the effect that "virtually everyone" took fitness. A reasonable guestimate is somewhere in the neighborhood of 75%.

Its not important what the precise diversity elements are throughout the entire game to determine if overall diversity increased or decreased to a first order approximation. First there is the observation that it is highly unlikely that those players without fitness would suddenly change their builds in a way that made them more self-similar. They would still have the same power choices. They could theoretically divert slots into fitness, but that's not likely to occur in a way that is exactly identical. Or they could decide to redo their builds entirely within the context of inherent fitness but its also unlikely that given inherent fitness that group would coalesce around a more similar set of builds. Whatever diversity they had, they would likely retain. And since fitness basically increases run leap, regen, and recovery, and since those fundamental changes could theoretically be replicated by simply increasing base stats, I do not consider adding fitness powers to every build to be a decrease in diversity. Characters with fitness slotted in the same way just have higher stats, and we wouldn't say they were more similar if the base stats themselves were increased. Contrawise, characters with fitness not slotted the same way are different.

The critical group are the characters that had fitness, and theoretically gained inherent fitness with a respec. These characters gained a number of power choices, probably three, and another power pool option since fitness no longer burned a choice. The question is, for this group of characters, is it more likely that for a given powerset combination that same high value of players all chose the same build modification: the same set of three powers from primary, secondary, and power pool. And since we're assuming that number is much higher than 50%, that would mean there would have to be one single dominant choice for *everyone*.

In other words, most energy/energy blasters with fitness all took the same three powers: they all took burst, combat jump, and maneuvers, say. Or alternatively, most energy/energy blasters were members of a dual build: one set took A, and the other set took B, and after fitness each took the other's powers, so they both had A and B.

Just *looking* at the opinions available suggests that the burden of proof is on the person making that assertion, because there are a wide range of valid options. A wide range of valid options for a specific case was presented in the thread. The fact they exist means this line of argument is only true if the vast majority of players don't consider them valid options. And there's no reason to believe that to be true.