Arcanaville

What you're saying is that Shield Charge is so unique of a power that any target the devs aim for would be totally subjective. I disagree that its that bad, but for the purposes of this discussion I'll accept that as at least a reasonable perspective.

What, then does trial and error do? If the devs have no idea what the power "should" really do, then what should they be looking for in each "trial?" If you have no idea what the power should be doing, you have no way of evaluating if the power is doing the right thing at all. And really, you shouldn't be adding things to the game that you have no idea at all what its purpose should be in the first place.

Its entirely possible the devs could *aim* incorrectly, in which case the mathematics of implementing that aim would lead to bad places. But when that happens, you're not supposed to throw your hands up in the air and say "oh well, lets try another completely random set of numbers" nor are you supposed to say "well, lets just keep fiddling with these knobs until we tune into a station we like." You're supposed to figure out *why* your aim was bad, and correct that bad judgment. And that too is mostly a numerical analysis (at least for the parts of the power that have numerical performance).

When Shield Charge left the bow, it could have hit something the devs didn't aim at, or it could have hit what they aimed at, but the devs aimed at the wrong thing. In both cases, that represents an error either in implementation fidelity or design judgment. In both cases that's an opportunity for learning: either for improving their design judgment or their implementation craft. Trial and error says just randomly guessing is better than trying to become a better game developer, and its always always the wrong choice.

The only time trial and error makes sense is during brainstorming. When you are trying to extend your craft significantly beyond where it has gone before, so you don't have an intuitive feel of either what your own design rules will do out in far or what your own numerical models will do in those regimes, some initial random trial to probe the new game space makes sense. But that should very quickly transition to more systematic testing, and it should never extend all the way to actual game implementation.


Just because there is subjectivity in game design, doesn't mean its without purpose. If it was completely purposeless, then when designing Elude's defenses if the devs accidentally set it to have 4.5% defense instead of 45% defense you're suggesting that they should try it out at that level first before deciding that 4.5% defense is "wrong." Because the choice to set it to 45% defense or 4.5% defense is completely arbitrary, and if its arbitrary who knows which decision will ultimately be "better."

That's not how it works. If Shield Charge had a design purpose, it should be possible to hit it, or get within the game's margin of error of it. If Shield Charge didn't have a design purpose that could realistically be aimed at, it was broken at birth. Once you say "I don't know what I want it to do, I'll just keep changing it randomly until I see something I like" you're not a game developer anymore: you're a random number generator that draws a salary.

The bottom line is that if Shield Charge was intended to be the best nuke in the game, the numbers say they hit that target in spades. If Shield Charge was not intended to be that kind of attack the numbers say they missed badly. That is not a matter of opinion. That's objective fact. Which one they were aiming for tells you if SC is bugged or not. Based on what Castle said, its objectively bugged. There's no subjectively whatsoever in that analysis.


On the subject of how to "balance" shield charge: its not actually a totally unique situation. There are other powers related to it that, if I was implementing the power, would act as guidelines for the limits of what the power should be capable of doing.

First of all, the two powers related to it are thunderstrike and lightning rod. Thunderstrike has a design similar to SC's original implementation, in that it delivers significant damage to its target and splash damage (and KB) to surrounding targets. Its design intent appears to be very similar to that of SC in that regard. Lightning rod, on the other hand, seems to be practically a clone of the original implementation of SC. It deals scale 2.4 at its target and 0.7 to surrounding targets, plus has a knockdown. My guess is that when the original concept of SC couldn't be implemented, LR was used as its alternate model.

Both of these powers deal high damage to a single target and a much lower level of splash damage to surrounding targets. And the reason for that is that high radius spherical AoEs tend to have less than scale 2.0 damage.

Then we have burn. The devs at one time thought Burn was overpowered if critters didn't make an attempt to run out of its radius. Burn deals a total of scale 3.0 in damage ticks if all of them land, and it does so in a much smaller radius than SC does: 8 feet vs 20 feet. If Burn is problematic dealing scale 3.0 over time within 8 feet, its not a matter of opinion on whether SC is problematic dealing scale 3.6 instantly within 20 feet.

Without being told, I would assume that spherical AoEs dealing 3.0 or better without extra restrictions are verboten for anything outside of blaster nukes, and anything that does more than 2.0 damage across the entire radius is questionable. Anything that does Lightning Rod damage is at least reasonable. My target for SC would be something between those two limits. And that's before even knowing what SC's actual design intent is. Is it intended to be stronger than LR, or similar. Is it intended to be focused, or distributed in its damage (more at the core, less all around, or more evenly distributed damage)? Is it intended to be used rarely, frequently, or something in between. Those things would inform my target for SC, before writing down a single number.

So no, I don't think SC is so unique of a situation that its completely beyond numerical analysis to implement. Ultimately, saying SC is beyond the ability for mathematics to handle is like saying its so unique its completely beyond the English language to handle.

I'm not advocating it specifically, but an example of a progressive penalty that would encourage players to bail out of Granite eventually without forcing them to do it explicitly, and could be balanced around offense vs defense, would be for Granite - after some set amount of time - to apply a progressive, asymptotic debuff to player recovery. So the longer Granite was toggled on, the closer player recovery approached zero. The lower recovery got, the slower your activity rate would have to get for any power that burns endurance (like attacks).

Interestingly, that means it doesn't impair the power at all as a weird form of out of combat costume option. But it applies a debuff very hard to escape from: only respites and endurance recovery (true recovery that actually adds endurance points like transference, not things that buff recovery rate like stamina or speed boost) would allow you to function indefinitely under Granite.

I'm not sure if this is "extreme" or not, but its not the only way to get creative with mechanical balancers. Its just an example. And there is another "advantage" of sorts to this kind of balancing mechanic: its extremely difficult to trivially min/max around. You could argue about whether that sort of penalty is better or worse than Elude's, but in general it would be very hard to prove, because the qualitative differences swamp the quantitative ones (or rather can be made to do so with the right selection of numbers).

That's a dangerous game to play, though, because you never want to deploy something you yourself can't figure out. I have an idea of how to quantitatively balance such an effect, but I wouldn't want to do so by simple trial and error like was done with Instant Healing.

Hmm. Think I've spotted the problem here, but just to be sure I'll ping Castle about it. It seems that Troll Integration might have been broken since practically the beginning of time, and was recently fixed to start working again. But its not a toggle, and isn't coded to suppress when held. In effect, although its called Integration, Trolls now currently have Practiced Brawler than never expires. That's probably not intended.

Back in I2 when Elude was changed to the defense click, it was perma-able with perma-hasten and 2-slot recharge (and quickness). I want to say it was 300s recharge and 120s duration. Back then, base tohit scaled with rank:

Minions: 50%
Lts: 62.5%
Bosses: 75%
AVs/Monsters: 90%
Pets/Turrets: 105%

Elude itself gave about 45% defense**, which could be four-slotted to about 81% defense. Passives were giving about 7.5% defense, so if you detoggled yourself (or respeced out of them except FF) you could have 6-slot passives and 4-slot Elude for about 16.5 + 81 = 97.5% defense.

Sounds like a lot, but remember: pets and turrets had 105% base tohit. Also: back in I2 we were running invincible full team missions that could spawn +5s. Each level higher added +5% tohit to the critters, so a +5 mission would be spawning:

Minions: 75%
Lts: 87.5%
Bosses: 100%
AVs/Monsters: 115%
Pets/Turrets: 130%

Even passives + Elude wouldn't soft cap you against +5 AVs or pets, pseudo-pets, or turrets (i.e. don't dance in front of the Malta Engineer). But you could be floored against practically anything else, and you could, if you were really serious about it, stack power pools like combat jump. Or you could go the extra mile and keep all the toggles, driving you to 127.5% defense (assuming the toggles were slotted for endred).

Which would be fine except in things like the Eden trial, or anything else with a lot of DE in it. Those annoying +100% tohit buffing crystals really used to tick me off. I actually tested once and discovered back then (before Real Numbers or even accurate measurements of tohit) that a minus twelve DE could hit right through normal SR toggles almost as if they were not there when buffed by one of those.


** There is some room for error here. At one point we were told Elude was granting 60% defense, which would imply Scale 8 with the scrapper 0.75 modifier. But its also possible it was Scale 6 and misreported without the modifier taken into account. There is conflicting information about it, and it was a long time ago. However, the 45% defense number better matches my observations of the time, particularly as perma-Elude went up against things even higher than +5.

There actually is no such rule. That's why Granite Armor is allowed to exist at all. However, powers generally have conceptual reasons for existing as well as numerical/performance reasons for existing. They are not necessarily etched in stone (har har), but they are independent reasons. In other words, no numerical argument can directly overturn a conceptual decision, and no conceptual argument can directly overturn a numerical decision.

Elude's conceptual design states that its a temporary, emergency use overdrive. Making it perma with any numbers would therefore be a violation of its conceptual intent. And you can't give a numerical reason for overturning its conceptual intent. You would have to give a conceptual reason for overturning its conceptual intent that was convincing to the devs. That aint easy.

Right now the conceptual intent of Granite Armor is to be a power you *could* use all the time but has sufficient downsides to make that decision non-trivial. In other words, its *supposed* to be something that players will choose to sometimes use, and sometimes not use. Some players could choose to use it all the time and others almost never, but across the range of players it should be used some percentage of the time that isn't 100% and isn't 0%. Traditionally, of course, the devs have not had a lot of success crafting powers that generate that situation (cf: Instant Healing, Rage, every version of Elude, Unstoppable, and MoG prior to I4, Granite Armor).


My own opinion on Granite Armor itself and Stone Armor in general is that I like conceptual diversity in principle. I don't think there is anything wrong with a powerset that has a tier 9 that can be toggled on all the time as long as its balanced appropriately. So I don't think its true that just because Granite can be made perma it is intrinsicly flawed. But I do think the penalty for running Granite is too easy to trivialize. However, changing Granite to have a penalty that *isn't* too easy to trivialize is likely to anger players used to the way the power works now. To be blunt: based on the kinds of suggestions I've read over the years, not just for Granite Armor but for other powers like Rage, I believe that most players (who have any opinion at all) believe that an "appropriate penalty" for anything is something that looks like enough of a penalty on paper to pretend to be a penalty without actually being something they personally couldn't completely negate in actual play.


Everyone has a gripe, and this is mine. Choices are supposed to have consequences, or they aren't really choices. While some players feel that consequences (ne penalties) aren't fun, the lack of them is the largest lost opportunity for fun for me personally. Had Champions Online managed to incorporate that one singular thing about the original Champions power system, instead of setting the entire thing ablaze and then claiming they had always intended on making a fire-walking game, I might have ended up spending at least as much time there as here.

Again: that is my personal preference opinion. Its not necessarily what I would recommend to the devs given the current state of this game and its playerbase.

The latter. This has also been tested in the past, so I would recommend carefully reproducing your tests. If Neutrino bolt really is recharging immediately upon casting, that's either a bug or a change in the way the game works. If I have the chance I will try to test this myself tonight.

Boy, that was a very defective envelope. I'm switching to napkins.

I think the low end is at least a decimal order of magnitude lower than you're estimating for manual review.

Hmm, yep, you're right the resistance is split between the two powers. That changes the calculations to 571/693 for LGTF and 6540/7660 for the Raid. 700 dps is still well within the limits of a team of SO-slotted damage dealers, although its now approaching the upper limits of such (an intermixed set of damage dealers and damage buffers or resistance debuffers should also be able to reach that value, probably more easily).

Holding the LGTF Mito reduces the numbers to 71/224, so being able to hold the mito reduces the damage you need to somewhere in the general vicinity of 1/3rd what you would need otherwise. So holding the mito is still very meaningful, but its still possible to overcome the lack of holds with reasonable - but plentiful - damage.

I actually think those numbers are even better than the ones I miscalculated originally. The represent a better value for controllers in that one specific phase of the fight.

Back of the envelope calculation:

Level 50 Lt Health: 855.1
Max Regen: 4.0088
Total Regeneration at the cap: 57.132 h/s
Resistance buff due to passive defense: 70%
Total damage required to match regeneration: 57.132 / (1 - 0.7) = 190.44 dps
Total damage required to defeat LT in seven second window between mito heals**: 190.44 + 855.1/7 = 312.6 dps

So basically, I think Castle's "unkillable unless held" statement is probably more appropriate for the raid version than the LGTF version. Unheld they can still be taken down in quick order by a team of damage dealers in theory. At level 50 a damage dealing character should be able to reach at least 100 dps in a team. Holding the mito cuts the difficulty by more than half, but its not unbeatable unheld.

(In a raid, its different because the mitos become special boss-but-better entities: the same calculation generates 2179 dps to neutralize regen and 3300 dps to defeat a mito in seven seconds. That's quite a bit higher, even with a lot more raiders).


** You don't have to kill them that quick, but if you can you can definitely overpower any heals that come your way. Its mostly a simplification for the calculation.

Or they could switch from the Ones table to a rank scaling table and picking appropriate numbers, but I would not recommend either in this case. In my opinion, AttributeRequires is probably abused more than it should be already.

Terminology. First, CoD doesn't list an activation time it lists an activation period. Activation Period is the rate at which passive and toggle powers "tick." For click powers, its ignored and usually set to zero. CoD is basically noting that whatever it might be set to, its not applicable for a click power.

Real Numbers calls "Cast Time" "Activation Time" just because its probably less likely to confuse casual players. They mean the same thing.

Neither CoD nor Real Numbers currently tells you what the Animation Time of the power is, which is itself a colloquial term for the technical phrase The length of time during which the power's animations will root you, making you unable to activate any new powers or move.

This may sound like a hair split, and for most casual observers of the game it probably is, but for technical people there's an important distinction. Jump Kick plays a three second long animation. Its "animation time" therefore you'd assume is three seconds. But only the first half of that animation is tagged as ROOTED, the second half is not ROOTED and interruptible, so its actually possible to cast another attack in the middle of Jump Kick's animation. So Jump Kick's rooted time is only half that, and that's the part that is significant to people attempting to figure out how fast they can attack with Jump Kick in their attack chain.

The important thing to remember is that cast time and recharge are governed by the powers system. Being ROOTED is governed completely independently by the animation system. Cast time determines how long the powers system will prevent you from activating any new powers, and when the power should start recharging. Completely separate from that being ROOTED will prevent you from actually doing anything, even if the powers system would otherwise let you. The combination of the two affects attack chains. But only cast time and recharge affect the actual availability and cycle time of powers.

We can't judge a powerset based solely on its IOed performance. For the purposes of performance-based balancing of powersets, two data points have the highest weight:

1. What the designed performance of the set is when slotted with SOs.

2. What the actual performance of the set is when datamined and averaged across the entire playerbase, from the absolute min/maxers to the completely incompetent (including important subslices, like the performance of all level 10-15 players or all 45-50 players, or all teamed players).

However, when those two metrics are balanced within a reasonable margin for error other factors are entirely appropriate to account for, including how well the set responds to being buffed by inventions. You can't break the above two requirements to go after high end performance, but you can and should go after high end performance in general to ensure its reasonable when compared to peer powersets.

No, that's not possible.

First of all, to the best of my knowledge Mitos in the LGTF are only weakened by being a lower rank than the ones in the zones. They otherwise have exactly precisely the same powers (the nucleus is a different story). That lower rank (they are basically Lts in the mission, but a special mito class in the zones) does alter the strength of powers like attacks, but the attacks themselves are the same power. So whatever the design intent of the "weakened Hamidon" was, it was implemented as a redesigned nucleus and a scaled down set of the same mitos.

This means the implementation intent of all the mez protection powers must be the same for both the standard and the weakened mitos, because they are the same thing. That's probably the real source of the problem: if the design intent was for the weakened hamidon to require a much lower level of mez magnitude than the trial version, they should have made two entirely different mitos. As it is, they are currently linked together.

My guess is that when the encounter was designed, the Healing Mito was originally set to have mag 30 protection to mez (it still does for things like sleep and stun). And then they decided at some point to increase the hold protection from 30 to 50. At that point a sign error was introduced.

Something worth mentioning is that protection powers are actually *negative*. When you run a protection power, it actually in a sense *debuffs* your mez attributes. If you have mag 10 hold protection, the power is actually *reducing* your hold by 10, or applying a -10 to hold. The meaning of the mez attributes is that if they are greater than zero, you're in that mez state. So negative is good, positive is bad. My guess is that at some point someone accidentally adjusted the hold protection by typing a "50" in there when they should have typed "-50."

This cannot possibly be by design, because it almost never makes any sense for something to hold itself unless something else is supposed to come along and free it eventually. Even if you wanted something to hold itself as a weird containment buff for controllers, there's no reason to apply a mag fifty hold to yourself, because there's no difference between applying mag 10 and mag 50 to yourself, again unless you're trying to overcome someone else's protection buff (and nothing I'm aware of can come along and buff Hamidon's mez protection). And if I was worried about that, I would apply mag 9999, not 50.

There is just no way that was done deliberately in this case. They'd be doing it to both versions of Hamidon, and there's no good reason to do it at all.

He said he set it to match the original design, not the original design intent. The original design apparently mandated 50 points of protection. Its almost certainly the case that such a high level of protection doesn't match the intended team requirements of the TF.


Vererne:

Actually, I was thinking almost the reverse. I think it might make more sense for the greens to have negative hold resistance and still high (but maybe not that high) hold protection. Negative hold resistance would mean holds would last longer, so they could stack higher. Having high hold resistance means holds will last less long, and that means it might be impossible to perma-hold the greens without enough hold casters. And if the hold flickers off for even a single combat tick, the mito will be freed to use its heals.

Negative hold resistance means a smaller number of mezzers could eventually overlap enough holds to overcome protection, which means holding the green would take more effort but that accumulated effort could come from less actual hold casters. It would open the door for three mezzers to be able to hold the mito fast, two to take longer, but even one could do it eventually, just with more effort.

Not sure off the top of my head if there are any indesirable side effects to self-debuffing hold resistance like that, but I can't think of any off the top of my head at the moment.

Anything is possible, but this is unlikely. The problem is that every body scaler is designed to adjust the geometry of its associated part in a specific way. The parts get bigger or smaller, longer or shorter. The different hair styles have different aspects of them that you'd want a slider to change, and that would mean the slider would have to do completely different things for different hair styles. I don't think that is possible without completely changing how the sliders work.

For example, you might want the slider to increase the size of an afro hair style. That essentially is equivalent to scaling up the hair geometry's external radius without changing its internal radius, so it remains fixed to the head but gets otherwise larger in all directions. Fair enough. But you might want a medium shoulder length hair style to get longer, which involves lengthening the back of the hair geometry without a corresponding increase in the parts on the head. The slider has to change what it does as you change hair styles, or else it will apply the wrong transform.

The only way to do this practically is to have multiple sliders for overall hair volume, hair length, etc, which map to all possible changes to all possible hair styles (like bangs, back length, pig/pony-tails, etc). And somehow I doubt the devs want to spend the time making eight separate sliders just for hair, at least not at the moment.

I always suspected, but didn't want to make unsubstantiated accusations.

I was responding to this, although I didn't connect the dots fully:

An "abusive" arc could be an arc which generates reward rates so high its very obviously exploitive in some way. But it could also be an arc which generates otherwise reasonable reward rates in unreasonable ways. For example, an arc which spawns an ally which then goes berserk in the mission killing everything in sight and generating 20% higher than average rewards for the player while they go afk and eat a sandwich would be abusive, even though 20% higher than average reward earning rates are not in and of themselves obviously exploitive.

So you could catch the worst offenders with a boundary checker, but in terms of what the upstream poster was talking about - allowing the AE to include higher than standard rewards - an abuse checker would need to also look at whether higher than normal but not astronomical rewards were not abusive in some way. Or to put it in the way they did, those rewards have to have an appropriate "cost." Which is why I mention that even if you could datamine difficulty and reward rates, neither one of those directly determines cost, the target you want to hit.

The only reason the community is always right, is because the community says everything. ED is good, and its bad. PvP changes are necessary, and unwanted. The AE needs rewards to succeed, and it will destroy the game with them.

In other words, the player community is always right because it is also always wrong. The only reason why this is not always obvious is because most people perceive themselves to either be with the vocal majority, or the silent majority, or the singular voice of reason. So its supposed to be obvious to the devs that they should listen to the vocal majority unless there is a silent majority they are drowning out, except when everyone is obviously wrong and they are supposed to ignore all of them and listen to the singular experts.

Or to put it more bluntly, the devs should always listen to me when everyone else says they agree with me, me when everyone else agrees with me but doesn't say so, and me when everyone else is wrong. How hard is that?

You could do that, but the two entity approach seems safer in terms of making the boulder itself being a separate entity subject to the physics engine, a separate bounding box, and a separate targeting location. You open the door to options like potentially pushing the boulder off the target instead of being forced to destroy it (moving it far enough away from the trapped entity could have the same effect of removing the "trap" buff by moving it out of range).

Of course pushing an actual boulder that is sitting on a trapped civilian would probably turn them into a skid mark, but then again shooting a fireball at it is probably equally unhelpful to the typical trapped person.

Separate from the problem Venture mentions, which is the difficulty in trying to compute the difficulty of an arc by computer algorithm, there is the separate, completely independent problem that even if you could, that doesn't directly tell you how much the mission should be worth, because not even standard PvE missions are designed in that way, and its not clear that's a good idea even if it could be done. You have to figure out how to turn that measure of difficulty into something useful for computing rewards, and that's not straight forward.

Solving Venture's problem gets you to the right longitude, but it does nothing to get you to the right latitude. Orthogonal.


This infinitely simpler process doesn't seem to have a simple way to select and vet player-reviewers, or deal with the potential backlash to players moderating players.


The devs need to get better, certainly. But personally, I think in a game of numbers my own chances would be better than average.

A mistake is a mistake whether its your job or not. In this case, though, the reason I'm kicking myself is because I actually caught a related problem, this one in the AE.

Anyone ever make a custom critter in the AE that used Lightning Rod before I17 went live?

I have only two extremely vague notions about that.

First of all, the last statement recorded an accounting loss of about 2000 MWon for NCI and a revenue of about 3350, which I'm assuming is mostly CoH. This suggests that at least for that quarter NC was accounting NCI as having 5350 MWon of expenses, at least roughly for our purposes. If we assume that about half of those expenses are being accounted for as part of whatever new thing they are cooking up, that means CoH the game itself was at least on-paper slightly profitable: to the tune of about 700 MWon.

That's like aiming for the 50 yard line and claiming to have landed somewhere in the stadium, but its the best I can do. Even that loss number is suspect because that bounces all over the place in their consolidated statements, and that can't be due to costs themselves fluctuating that wildly, I don't think. So a more in-depth analysis of the numbers would need to be done to figure that out. But I suspect that the game itself is still profitable in terms of its real costs to support, separate from the accounting perspective which may be different.


The more critical thing I think is actually not profitability. Its critical mass. I think below a certain critical number of subscribers the difficulty in growing the game crosses a point where, short of a miracle, the trend can't be reversed in practical terms. In other words, the game could last for years but it would be constantly dwindling downward, albeit very slowly. I personally don't think we're there yet, but I personally wouldn't want to flirt with subscriber numbers much lower either.


Personally, I would rather have a slow but sustained increase in subscribers over time than a big spike followed by an inevitable drop, even if the net effect is to end up somewhat higher. In my opinion the game needs long term forward momentum, not a quick burst of activity and cash. Of course, I'm not the one that has to manage the cash.