Arcanaville

[ QUOTE ]
"If you look over to the left, in the Forest of Data, you'll see a lovely discontinuity. . . meanwhile, in the Land of Overlapping Slopes, you can look up at the majestic face of Regen. . ."

[/ QUOTE ]

Here in the Arcanaville archives, you can see footage of:

Invuln, post-I5: Tough, but a bit depressing.

Regeneration, as it once was in I3: ridiculously powerful, but still occasionally getting smacked by strong bosses.

And Super Reflexes, as envisioned conceptually by the devs.

[ QUOTE ]
While 10k samples might be nice if you're trying to determine the value of Lucks to 12.5%, a thousand should be more than enough to show that the value is "far" from 25%.

[/ QUOTE ]

In this case, since the actual value for two lucks (50% mitigation, 50% incoming damage) is a factor of five from the expected value of two lucks (90% mitigation, 10% incoming damage), it took even less than that.

But you really need to factor everything else out, and just see a bunch of guaranteed to be identical swings with two, and then with four, and then the results become blatant. Four drops all attackers to the floor, so you can see what "the floor" looks like. Then two *should* also drop all the attackers to the floor, but it so very obviously doesn't. Three, if you pay careful attention, looks like something in the middle.

And if you use, say, an unstoppable scrapper to test with, you can use the health bar as a nice proxy for average incoming damage. Its crude, and I wouldn't be able to tell how strong lucks actually were, but I could tell that they were vastly lower than +25% defense, but also couldn't be much lower than about 12% defense. And the test is reproducible as many times as you like.

Its so blatant, that I didn't even think it was worth computing the actual scores from the demorecords, although I was prepared to do that if necessary. The signal is very strong, and most importantly, I could describe exactly how to reproduce the experiment if I needed to (as it turned out, I didn't need to).


Now, this is not the same thing as someone saying "very often, I seem to miss a lot, like I only hit the target two or three times out of ten, when I must be at 90% tohit or better" because *that* sort of thing is bound to happen occasionally, and is vulnerable to selection bias. Without a lot of additional details, even if someone says "I've been playing since beta, and I know this is wrong" I'm not likely to pursue that very far myself.

But "sometimes, it seems like lucks don't work until I pop one or two more than I think I need" was rattling around in my head when I happened to be doing inspiration-related tests. And that isn't a "1000 swing" issue.

[ QUOTE ]
Posts like this always make me feel so horrible! In my 11 thousand + posts I haven't done anything productive.

[/ QUOTE ]

Being productive has its disadvantages. While I was off testing things, Pilcrow used 7,241 CoH eval keys to steal the "most helpful poster" title from me, I just know he did.

[ QUOTE ]
Statesman, Jonyu and I all worked on translating the numbers in game to the Prima Guide. There are likely some typos and a transposed digit or two, but most of the numbers are accurate.

I think the pool powers values were all based on a fictional AT with all mods of 1.0 -- so if you apply your AT mod to the values, that should get you pretty close. Oh, and also, all the values are calculated at level 50. So, in many cases you'll get lower values at lower levels.

[/ QUOTE ]

We need one of those cardboard wheelie circular sliderule-like things, where you dial in the power, and the AT, and the level, and the power value shows up in a little cutout box on one of the wheels.

I'm not sure if I should be flattered, or disturbed.

The odd thing is after reviewing them carefully, this technically doesn't appear to violate any of the forum's rules, as it is an actual guide, of sorts.

Whether this is an exploit, or just working as intended, is I suppose the question.

[ QUOTE ]
Never got this either. I prob spent more time on this game than some of the people who try and out shout me put together. I know when things are off, I know when things aren't working right, and I don't need to do a 1000 hit test to prove it to myself. You think I'm wrong then "YOU" show proof that I'm wrong.

Look at the AV changes as a classic example of what's right and what's wrong. I counted at the most 3 people on a 1112 replied thread that said they had no problems. Out of those people they had a Rad or Dark on there team. Then when we get a response it's "Get 8 people, put it on Heroic, and that's your balance."

More examples are the Master Mind bug that is devestating an entire AT set. Yet they put this hunk of crap they called I7 in anyways.

What's more is way back in the early days before these dickless wonders started prancing around in there bikini bottoms saying "The World is wonderful." we were stating "Dude there's something not right here." and we knew it. We know [censored] is wrong and then the the Dev's say "Go get a 1000 hits and then come back."

I7 is wrong and there still not listening. When the numbers start crashing, when there customer base gives them the finger (just like AC 2 and SWG customers did), you won't be able to get them back. MMORPG customers are finiky and once they leave there gone. They invest major amounts of time and money in a Hobby we expect others to care for.

I7 is bringing everything to a head. I call [censored] as I see em and right now this pile is pretty deep. I hope you guys can save this game but you are headed in the same direction as AC 2 and I called that hunk of garbage the first month it went live. You can't get customers back once they leave. First goal is to keep them from leaving in the first place and right now they might as well have poured gasoline on the fire that is sinking the ship.

[/ QUOTE ]


Hmm. Well:


1. I, more than anyone, have suggested that for certain types of problems with accuracy any anecdote based on less than thousands of swings is questionable. In actual fact, I'm unaware of any dev specifically asking anyone to conduct any such test. If anyone out there thinks the devs are forcing people to do such things, I've seen no evidence of it. *I* do it because I know, given all the reports of accuracy not working - most of them unfounded - the burden of proof on all things accuracy is pretty high. And in fact, even for me, my own burden of proof is pretty high when other players report things to me. I don't dismiss reports out of hand, but neither do I track down every issue that's mentioned to me, either.

But the 1000 swing thing is probably my fault as much as it is anyones. Not every problem requires such a test to uncover. But frankly, many do. And even knowing which is which is, itself, a complicated problem. Everyone thinks they understand statistics. Statistically, most have to be wrong.


2. There is a difference between pointing out that something isn't working the way it ought to mechanically and pointing out that something isn't working they way you expect it to conceptually. And in fact, as I mentioned previously, I think its important to keep the two separate, because one is provable unambiguously, and the other one is highly subjective.

Inspirations are broken now, unambiguously, because they do not do what they are explicitly labelled as doing. There is no argument there at all.

If I think the actual values are not the right ones for game balance, thats a different issue entirely. I know I'm always right about such things, but not everyone has gotten the memo yet: I still have an obligation to state my case, and not everyone will agree, as horribly wrong as they obviously must be.


3. The burden of proof is on the person that wants something changed. Except for me, of course.


Random worthless anecdote: got asked to join a team to take down Numina, while playing my EM/DA brute. There was already a DM/DA brute on that team, so I wasn't the only one with psi resistances when I joined. They were apparently having a ton of problems with her before I joined, and after I joined, we creamed her.

Thing is, I have no idea what I did to help: I did not have continuous aggro, and my damage is good but not that good, and the team's composition was pretty random otherwise. I think it was just one of those things: sometimes AVs (and Heroes) just seem to have your number, and sometimes a small change in the team, or a change in tactics, resolves the situation. AVs and Heros seem harder now than before, but not inordinately to me. No idea whats responsible for the wide range of opinions there, but that sort of thing might be part of it.

[ QUOTE ]
So, whens Arcana getting the Qa job.....right now shes doin it for free...but im sure you can throw in a free subscription and some airline penuts to get her to be your Qa

[/ QUOTE ]


QA?

Heck no. Then I would have to start being nice to all of you people.

Give me commit privileges, though, and you might see toxic defense before the next millenium.


You know, I don't test things just for the sake of testing things. Invariably, I'm thinking about a balance issue, and that requires testing things, and that requires testing other things, and that requires getting good numbers on other things, and before you know it, I'm trying to see if combat jumping is 1.875% defense, or 1.900% defense.


In this specific case, I started out testing SR vs Invuln with and without inspirations, against a really tough I7 target. And then the testing took a left turn, and I decided to pursue another angle. And here we are.


If I charged Cryptic what I normally charge to do the type of testing I do in CoH, they would have to comp my subscription until Issue 937: The Osteoporosis Wars. Statesman's travel power will be a power scooter, Swan will still be in the same costume, and no one will want to train in Brickstown anymore.

[ QUOTE ]
We all know that Arcana doesn't swing trout

[/ QUOTE ]

Damn straight.



Err...




...this is a compliment, right?




And before this becomes too much of a "thank god they listen to Arcana" thread, I think a little perspective is in order.


I think there are red names that would listen to anyone who:

* is respectful, even if angry

* is patient, even if anxious

* knows what the heck they are talking about, and is willing to learn to speak their language when they talk about it

* Knows and respects the difference between reporting something is broken, and complaining about something that you don't approve of.

* Checks first, reports second; not the other way around.


Of course, none of them will have the Bragging Rights that I do, but that's another story.

[ QUOTE ]
[ QUOTE ]
They are tohit buffs.

[/ QUOTE ]

Arcanaville, have you experimentally verified this?
In one of the earlier issues my testing indicated that Burn was being affected by Insights.
The same testing also suggested that Burn in PVP had a base accuracy of 75% and that lucks were 12.5%. (I think it was I5.)

[/ QUOTE ]

I have suggestive tests that would seem to confirm they are tohit buffs. I cannot claim to have definitive tests that prove they are.

What I know is this:

1. I've been told they are tohit buffs.

2. I've been told that if I wanted to make something like them accuracy buffs, there isn't an obvious way to do that given how accuracy buffs work internally.


I consider the statement "there's no supported way to make a self or ally accuracy buff in the current game engine" (paraphrased) quite strong. It implies not just that insights aren't accuracy buffs, but more importantly *nothing* that you think might be an accuracy buff can be one, except for accuracy enhancements, and inherent attack bonuses.

Nothing is absolute: I might have misunderstood, the person (people) that told me that might have been wrong, or there might be yet another thing going on that simulates accuracy buffs without being one literally.

But when I asked really specific questions about accuracy buffs, my understanding was that it was basically impossible for a power of any kind to confer one, given the current limitations of the game engine. That would seem to be fairly absolute on the question of whether or not insights could possibly be accuracy buffs or not.

I think people thought they were accuracy buffs instead of tohit buffs because they seemed to be much weaker than a tohit buff of their stated value could possibly be, but in fact, it was simply that they were, in fact, weaker than their stated value.

[ QUOTE ]
As for something that matter to me though, I'd like to ask if there is any ETA for costume fixes

[/ QUOTE ]

True story: the day I tested Rikti Mesmerists against my MA/SR and discovered SR had no defense against them, I noticed for the first time her skirt didn't have those patchy clipping problems that the short skirts have.

The weird thing is I wasn't sure if that was a net plus overall or not.

[ QUOTE ]
[ QUOTE ]
So, Castle, drop whatever you're doing and get CoH/V Korea out the door so we can clone you

[/ QUOTE ]

Clone Pohsybs instead, they pack into boxes better and don't require "vintage" rock music

[/ QUOTE ]

I was going to suggest genetically crossing pohsyb and Castle and making a Developer Queen that can go anywhere and do anything, but I've already been accused of being too obscure several times this month.

[ QUOTE ]
I've been looking through the Prima Guide and notice that it lists the values for Aim and Build Up differently from how we've been assuming they all worked. We all had believed that Build-Up was +50% To-Hit, +100% Damage, and Aim was +100% To-Hit, +50% Damage.

According to the downloadable Prima guide, however, Build Up gives +20% To-Hit, +80% Damage, and Aim gives +42.5% To-Hit, +42.5% Damage.

What's up with that?

[/ QUOTE ]

I'm still going through the manual carefully, but it seems there is the potential for scale errors in the numbers. For example, power pool defenses have actual quoted numbers, but they vary in strength from AT to AT, and so *cannot* be the values quoted for everyone.

As to BU and Aim, still thinking about that.

[ QUOTE ]
I just created a sample Super Reflexes Scrapper using the City of Heroes Character builder by Sherk Silver. I took every single SR power and 3 slotted them for defense (at least the ones that can be slotted for defense), even level SO's. Now it lists the defense totals for Melee, Ranged (both at 27.3%) and AoE (46.8%) attacks against this character. Elude adds (at least I'm assuming it stacks) another 70.2% defense to Melee, Range and AoE.

Now after all the long windedness, here is my question. Is it possible that the defense numbers without Elude (27.3% and 46.8%) are based on Arcanaville's theory here? Could it be possible that SR defense numbers are based on the To-Hit numbers and not actually making the attacks against SR miss 27.3% of the time, but actually only 13.7% (rounded up) of the time? And if this is the case, where does that leave the often cited "1 defense is more or less equal to 2 resistance"?

[/ QUOTE ]

SR isn't *that* bad. Its defense numbers check out fine.

Except SherkSilver, and probably every builder out there, has wrong numbers for SR.

SRs toggles are 13.875%, not 12.5% like everyone thought they were. SR's passives are 5.625%, not 5% like everyone thought. That makes 3-slot defense 30.4%, not 27.3%.

Also, from I5 to I6 SR had higher AoE defense, but in I7, its AoE defense is now exactly the same as its melee and ranged defense.

[ QUOTE ]
Off the top of my head, I'd say any report of strangeness voiced by Arcana, Circeus, or Fraktal is probably worth paying attention to, especially if there are a large number of people nodding in agreement.

[/ QUOTE ]

Just to reiterate what I said previously: in virtually every case, every red name I've brought a specific problem to has listened and checked it out carefully, so I have no specific complaint there. Of course, I try not to bother them too much unless I have a smoking gun. Often, I try to give them a picture of the bullet actually hitting the target. That makes me, I hope, a good risk/reward candidate.


Of course, my "change it, it sucks" reports go into the same pile everyone elses goes to.

[ QUOTE ]
[ QUOTE ]
[ QUOTE ]
Or perhaps there is a core group of a dozen or so players the Devs should always 'listen to'-- even when those players disagree with each other?

[/ QUOTE ]


When in doubt, they should always listen to me and do whatever I say. Its not always the best thing, but its a good rule of thumb. And not just for the devs, either.

[/ QUOTE ]

let me know if they buy that. I have some Ideas about defiance that I'd like to funnel through you.

[/ QUOTE ]

If the devs obeyed me on matters of defiance, there'd be no defiance for you to have ideas about.

[ QUOTE ]
Or perhaps there is a core group of a dozen or so players the Devs should always 'listen to'-- even when those players disagree with each other?

[/ QUOTE ]


When in doubt, they should always listen to me and do whatever I say. Its not always the best thing, but its a good rule of thumb. And not just for the devs, either.

[ QUOTE ]
Or in this case, Arcana was looking for one type of bug ("inconsistent behavior"), and found something completely different. It happens.

[/ QUOTE ]

Actually, I was specifically looking for scaling behavior, and I found it in one.


[ QUOTE ]
Anyone who thinks that debugging software is easy has never done it.

[/ QUOTE ]

Fixing software can sometimes be a complicated task. Finding bugs like this *should be* easy. I say, *should be* easy because it often isn't, but when it is, *that* is in and of itself a problem.


Sometimes, you have to take a step back, and realize not even your designers and developers have what they need, and give it to them. Historical Precident is not hard to find.


Someone once explained to me the difference between the design of a nuclear submarine, and a jetliner. Jetliners are designed to hide the details of their construction and operation from the passengers, because the passengers are presumed to be, on average, much more likely to cause problems than solve them. Nuclear submarines are designed to show all of their inner workings to every single passenger, unavoidably, because its presumed every single passenger is much more likely to spot and address problems than cause them.

Even if the weird ugly details of the game mechanics are hidden from *us*, the game should be shoving them in the face of every Cryptic playtester and designer constantly, inescapably, obtrusively (at least while testing) - so that if something isn't happening the way they think it ought to, it hits them over the head day after day after day.

If I were in charge, that would be the number one feature I would ask the coders to add for me. And I wouldn't take no for an answer.

(and then I would order them to fix stacking)

[ QUOTE ]
Because it would be interesting:
Time to death against a "reliable" "fast" damage source (PBAoE DoT, ...).

I can't say for sure how easy it'd be to find a damage source that's "just right", but it'd be an interesting approach.

[/ QUOTE ]

I got lucky. There's a level 50 contact that just happens to give out a mission very early on that generates a large number of attackers in a very small area that cycle attacks very quickly that an unstoppable scrapper can survive when using at least three small lucks.

[ QUOTE ]
This also explains the last battle I fought with my rad defender. I was 49 fighting a level 52 malta titan. With 4 lucks and RI running that thing was hitting me like crazy. The combination of debuff and lucks should have had him whistlin dixie. Instead I was getting hammered like a drunken fratboy. The only reason I won was Enverating field, temp invul and a couple large resistance inspires. THe lucks didn't even seem to be there.

[/ QUOTE ]

If this battle was relatively recent, given what we now know about lucks, this definitely would not have been enough to floor that titan boss (with the purple patch acting on RI and the four lucks "only" offering 50% defense against something with 80% base tohit). But on top of that, and my memory is a bit sketchy on them, its possible their energy blasts have -DEF debuffs - which would also be boosted by the 3 level difference. And all Malta seem to be accuracy-boosted on top of that.


[ QUOTE ]
One things for sure. If you start popping lucks after combat has begun... there's a damned good chance that you're going to get floored or hurt real bad. ALways seemed like the effect wasn't instantaneous.

[/ QUOTE ]

As a super reflexes scrapper, I have often wondered if the reported "delay" in lucks working was based at least partially on a lot of people underestimating the number of attacks with -DEF in them (in combination with what we now know to be our overestimating the insps themselves): players who use lucks a lot might be simply seeing the cascade failure that defense sets are more used to seeing. Just one hit from a defense debuffing attack prior to using lucks looks like its capable of negating a whole luck or more. Which means it might take one more luck on average to get the same performance if you let them shoot first, than if you pop them first (assuming you're only hit once). Until the debuff wears off, and then suddenly the extra lucks suddenly floor the target.

[ QUOTE ]
[ QUOTE ]
Actually, I think the problem in this case was that there were a lot of very experienced people on both sides of this issue who were going by feel, and no one wanted to do controlled testing of something like this, because controlled testing is not easy to do (ten bucks says no one comes close to guessing what sort of testing uncovered this).

[/ QUOTE ]

If I was going to test a transient defense buff like Lucks, I might consider using Rain powers.
Using Ice Storm, it seems like you should easily be able to fit about 150 tics into the duration of one single Luck (more if you use Hasten).

But this seems too easy to be the answer.


[/ QUOTE ]

Yeah, and you know I test in PvE, so thats not a viable option. If you are at all curious, my tests were not with my MA/SR, and they were ultimately suicidal.

[ QUOTE ]
Arcanaville I'm not kidding this sounds like a good power to trade in MoG for. Give us that stance too and it would be fun to take and use. Change revive to a passive that does something small, maybe halve the effect of -regen . And you can call it a day with Regen for now.....

[/ QUOTE ]

I'll suggestify it after the dust settles on I7.

[ QUOTE ]
So that means it takes 6 lucks to equal the defense protection that mog gives minus the psi of course?

[/ QUOTE ]

As of this morning, four (barring tohit buffs).

[ QUOTE ]
Excellent work, thank you.

Do we know if the yellow inspirations work as an accuracy boost (like an ACC SO) or a toHit boost (like Aim)? I read the thread but it wasn't clear.

[/ QUOTE ]

They are tohit buffs.

[ QUOTE ]
[ QUOTE ]
This is astonishing. Truly astonishing. The one set of statistics modification numbers that the devs haven't hidden behind the veil of "Aaah! No 'City of Math'!"...has turned out to be half wrong, and apparently they've been wrong since the beginning of the game two years ago.

Why on earth did this take so long to come out? Why wasn't, say, the fact that the oranges were "so much weaker" than purples when they were introduced, sufficient to get the devs to say, "Oh, and by the way, purples are really not 25/33/50 after all..."? It seems clear that the oranges' numbers were being scaled against the actual values of the other Inspirations rather than their described values. Why didn't anyone correct it at that point?

I mean...wrong for two whole years?

This makes me wonder what other "facts" that we take for granted aren't true.

With this level of QA, Cryptic is really squandering the first-mover advantage in the superhero MMORPG genre. If they don't get their act together, and Marvel or DC come out with something better, they're going to eat Cryptic's lunch but good.

[/ QUOTE ]

like I said robo. those 25/33/50 numbers used to be accurate.

and really they were broken. You shouldn't be able to pop 4 lucks and out tank an ice tanker. I don't mind the "re-balance".

I just don't appreciate the stealth nerf.

[/ QUOTE ]

I'm told the posted numbers have been the same pretty much all the way back to release. To the extent that at least the numeric values of the inspirations haven't been changed, inspirations haven't been nerfed.

And they seem to have always worked the same way mechanically, so I don't think a tohit formula change altered their behavior either.

There are still other possibilities, but I have no real proof their effectiveness has changed over time. They don't *feel* like they did to me, but of course in this case that's probably not worth a lot.

[ QUOTE ]
I figured that. But Reconstruction also gives 15% toxic resistance.

[/ QUOTE ]

Right, but as I corrected in that post, my suggestion was meant to offer toxic resistance, so its already 70% res to smash/lethal/fire/cold/energy/negative/toxic. Recon's toxic resistance is of limited usefulness especially if this version of MoG is 1-slotted with res to bring all resistances to the 75% cap anyway.