Lucks and Insights do not work the way you think
By
Amarsir,
Posted on: 2006-06-06 11:23 in Player Guides
Arcanaville, you're my hero. Well, you and Captain Marvel are my heroes. But at least you're in good company.
~Gabriel 
This also explains the weird behavior I got stacking insights against a Night Widow.
Thank you for posting this information. So many of my inspiration stacking experiences make sense now.
Elsegame: Champions Online: @BellaStrega ||| Battle.net: Ashleigh#1834 ||| Bioware Social Network: BellaStrega ||| EA Origin: Bella_Strega ||| Steam: BellaStrega ||| The first Guild Wars: Kali Magdalene ||| The Secret World: BelleStarr (Arcadia)
I read it and I have to say I am not shocked. I kept thinking I wasn't getting what I expected from Lucks.
Well anyway thanks for your effort in this.
Get this moved to Guides so it doesn't get deleted!!!
Yeah, I vividly recall this scene:
My 40 something stone tank Stares at the "biggest Clockwork" as the mission describes it. Granite goes away, and he layers on all armors save the fire/cold one. Thinking minerals 3 slotted defense alone won't cut it, he chews 5 purple pills. One taunt later, the awaken saved gets used.....
Yeah......
[ QUOTE ]
Get this moved to Guides so it doesn't get deleted!!!
[/ QUOTE ]
I thought it made more sense as a discussion topic here and would draw more attention. I'll make sure this thread or a reposted version eventually makes its way to the Guides section for permanence.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Great work Arcanaville. We all felt Lucks were overpowered based on the in-game description numbers & expected some kind of nerf to them eventually. Of course even these numbers are likely still too high. And I am surprised about the Luck/Insight imbalance - these 2 inspirs need to be balanced against one another on a one-to-one basis if the Devs continue to expect us to use them as counterweights in an ever-escalating DEF/To Hit Buff arms race, as seems to be their intent. 
Anyway, as always your contributions to bettering our understanding about how this crazy game that we love actually works are well appreciated. 
Liberty
Mister Mass - 50 Inv/SS/NRG Mut Tank [1236]
Doc Willpower - 50 Grav/FF/Psi Mag Controller
Baron Wonder - 50 SS/Elec/Mu Mag Brute
Sound Bight - 50 Son/Son/Mu Tech Corrupter
Could the acc strangeness be due to the change in how to-hit is calculated vs. the old days? In the old days accuracy and to-hit buffs were basically the same (if I recall) but now, to-hit buffs are ADDERS and acc buffs are MULTIPLIERS. If Insps are acting as a boost to your multiplier, that could make it seem rather wonky. Do we know where the +to hit is being added? (Yes, I know it's supposed to be added as a +buff, but I mean where it's *actually* added vs. where it's *supposed* to be added.)
Things like this are why I question the devs so much on crap like, "75% is base to hit and we're sure it's right." Apparently they were "sure" that +acc insps were "right" too, but they're clearly NOT right (at least not according to the values displayed in game for the last 2+ years).
I think they need to stop looking at the SQL database that stores the percentages and values of things, and start looking at the actual CODE to make sure that things are being added and subtracted in the right place.
F
I've added this to my buglist. I'll update the text and make a balance pass on Inspirations in the not so distant future.
[ QUOTE ]
Great work Arcanaville. We all felt Lucks were overpowered based on the in-game description numbers & expected some kind of nerf to them eventually. Of course even these numbers are likely still too high.
[/ QUOTE ]
I'm not sure they haven't been like this for quite awhile. The numbers are consistent with my experiences back through issue 3.
Elsegame: Champions Online: @BellaStrega ||| Battle.net: Ashleigh#1834 ||| Bioware Social Network: BellaStrega ||| EA Origin: Bella_Strega ||| Steam: BellaStrega ||| The first Guild Wars: Kali Magdalene ||| The Secret World: BelleStarr (Arcadia)
Just out of curiousity, that balance pass would move the Inspirations to what values? 
(I.e., I'm assuming the Phosyb numbers given are the correct values, and aren't changing.)
[ QUOTE ]
But in this case, it was the people who complained about seeing oddities that were right, and the people who thought they were working correctly that were wrong. Including me: sorry guys and gals, it took me a while to eventually come around and devote testing time to this.
[/ QUOTE ]
I've said this before, and I will say it again. Statistics are great, and I love them, and I work with them all the time (right now I am running two statistical model analyses on my desktop and its companion machine as part of my job, doing Multivariate Auto-Regressive community models). I am not someone who "does not believe in numbers" or any of that.
HOWEVER, I also recognize that games, like sports, are often played "by feel." In a sense, what happens is, when you get really good at a sport or a game, you internalize a lot of the numbers. A good tennis player doesn't have to look up the fact that hitting a cross-court return into the wind against this opponent is a 'high-percentage' shot, and he's certainly not computing averages during the match. What he's doing is instinctively, from his experience playing tennis, returning the ball where he just "knows" it's the best place to return it. He knows this because he has had hundreds or thousands of hours of practice and match-play to tell him, without needing to resort to calculators or lookup tables, what the "right play" is here. At the end of a match, a player probably can't tell you he had "10 double faults" or that his normal average is 6 per match... but he will definitely (after dumping 10 second serves into the net) say something like, "Wow my serve was off today." He knows how it feels to play "right", and not so right.
By the same token, a good gamer, who has long played a video game and gotten good at it, has a good "feel" for the game. He knows what normally happens, or what he might say is "supposed to happen" -- that is, what experience tells him is common -- in a variety of circumstances. A good veteran COH player has faced a 3-yellow-con spawn of Vahziloks hundreds, probably thousands of times, on a whole range of characters. He knows pretty much what to expect, and he can play the game against that spawn pretty much without thinking, by reflex, by instinct, the same way the tennis player can do so. He has the "feel" of it down pat.
As a result, when something bizzare happens, like getting one-shotted by something that's never done half his hitpoints on his squishiest character before... or whatever odd thing might happen... when something goes on that simply "feels" wrong, and goes against his instinctive expectations of the game, the player will notice. When an experienced, veteran player, therefore, says, "Wow, I am missing a lot today," unlike most other stats-oriented people (as I say, I am one), I do not discount it just because he "has no numbers to back it up." Rather, just as with the tennis player who said (without any stats to back it up) "wow my serve was off today", I assume that a good, veteran player knows what he is talking about, and it's at least somewhat possible that something strange is "up." Maybe the RNG got "stuck" today... or maybe he's fighting something with +def and doesn't realize it (but that's unlikely if it's a vet)... or maybe (insert one of a thousand possible reasons here)... Whatever the reason may be, I think just waving one's hand at a vet player who knows the game backwards and forwards and saying, "It's all in your mind," or "you just notice bad luck streaks" is not only contumelious, but completely mistaken. It'd be as foolish as waving your hand at a tennis player after he said, "My serve is off today," just because he doesn't have a video recording of all his double-faults.
I hope this little nugget will be a lesson to the Doubting Thomases who arise to shout down anyone without a 5 MB file of 10,000 hits and misses to back him up. The people who were correct in this case were, not the statistical gurus (sorry Arcana, but then I am in this group most of the time too), and not even the devs themselves, but rather, the people who KNOW the game inside out from many hours of playing it, and could tell just by feel, that something odd was going on with inspirations. So maybe from now on people will be a little more open-minded when the more intuitive, but experienced, gamers have something to say about these things.
I know, I know... "dare to dream." But hey, I can hope can't I?
F
[ QUOTE ]
Could the acc strangeness be due to the change in how to-hit is calculated vs. the old days? In the old days accuracy and to-hit buffs were basically the same (if I recall) but now, to-hit buffs are ADDERS and acc buffs are MULTIPLIERS.
[/ QUOTE ]
I think the difference between the old days and today is that people *thought* it was one way back then, and *think* (correctly) its this way now. I don't think the mechanics have actually been changed since release, or even significantly prior to that.
However, even red names have occasionally misspoke on the subject of accuracy, which can further confuse the issue. pohsyb, one of my sources for technical details, admitted to me that one of his accuracy examples that was being used as the "definitive" proof that accuracy was additive was in error, and explained to me that in actual fact its multiplicative. It wasn't changed, his example was just off in that one detail. It happens when you are trying to illustrate one thing, and overlook a detail somewhere else: it happens to me occasionally, which is why I specifically proofread for that type of error in everything I post.
Its also why the devs are sometimes gunshy about posting numerical examples: sometimes a typo can loom large as people extrapolate an entire game engine around it.
This doesn't excuse the Mr. Magoo mistake on the inspirations' text descriptions, though. Which is, for me, the meta problem with all things accuracy related. If they had screwed up the sturdies, it would take all of five seconds to figure that out and bug it, and a lot of people could generate unambiguous proof of the error.
How many people can actually test accuracy problems and know how to draw proper attention to them, in all of CoH? A couple dozen? A couple?
Geko is basically right: accuracy is not broken, and basically never has been. But practically everything related to accuracy has been broken at least once in some way, sometimes for extended periods of time, and often the hurdle to prove it is extremely high. I'm not sure what the best way to fix that is, but *somehow* I think the game needs to tell people, in at least general terms, what's going on with accuracy, when they do things like pop lucks, or use radiation infection, or have six guys in the invincibility field. Not necessarily with a flood of numbers, but something.
My instinct is to say there should be a "/debug accuracy 1" command that when enabled shows the net tohit percentage on every hit and miss in your combat chat. Heck, even if it was locked out for players, I would think this is the sort of thing the devs should have access to, so I could PM a red name and say "check this please" and they could check the true net in-game behavior in like ten seconds.
By the way, does anyone think I'm being selfish in thinking all this accuracy testing deserves a badge? 
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
[ QUOTE ]
I've added this to my buglist. I'll update the text and make a balance pass on Inspirations in the not so distant future.
[/ QUOTE ]
Thanks Castle. This saves me my next PM.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Well, I still hate hitting Aim on my Corrupter which has 2 Single Origin Accuracy Enhancers on all attacks and know that I get a 90% miss rate on the first attack I use - stupid random number generator
[*]Scrapper Secondary Info in Issue 7 (draft) [*]Tanker Primary Info in Issue 7 (draft) [*]The Harsh Reality of PVP
[ QUOTE ]
Geko is basically right: accuracy is not broken, and basically never has been. But practically everything related to accuracy has been broken at least once in some way, sometimes for extended periods of time, and often the hurdle to prove it is extremely high.
[/ QUOTE ]
Even that is not a correct statement. One of the serious problems with the game is that there are THREE things related to to-hit, and those things are confused even by the devs themselves. In this case when Geko said "Accuracy has not been nerfed," what he meant was, "Base to-hit has not been nerfed", but he called base-to-hit, in this case, "accuracy."
There are three numbers here: Base To Hit, To-Hit Buffs, and Accuracy. The problem is that the devs use them interchangeably with WORDS, but they are not interchangeable MATHEMATICALLY. So Insights SAY +25% _Accuracy_ but what they mean is +25% (or not even that, but whatever) _To_Hit_Buff_.
The devs need to learn to, as Lucy said in the Peanuts years ago, "Say what they mean, and mean what they say." The sloppy use of the lingo is 90% of the problem.
F
[ QUOTE ]
By the way, does anyone think I'm being selfish in thinking all this accuracy testing deserves a badge?
[/ QUOTE ]
I hereby award you the 'Who Watches?' Badge. It will be arriving in the mail shortly. Now be quiet and go back to testing stuff!
The guy that updates text only works 3 days a year
[ QUOTE ]
I hope this little nugget will be a lesson to the Doubting Thomases who arise to shout down anyone without a 5 MB file of 10,000 hits and misses to back him up. The people who were correct in this case were, not the statistical gurus (sorry Arcana, but then I am in this group most of the time too), and not even the devs themselves, but rather, the people who KNOW the game inside out from many hours of playing it, and could tell just by feel, that something odd was going on with inspirations. So maybe from now on people will be a little more open-minded when the more intuitive, but experienced, gamers have something to say about these things.
[/ 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).
I've felt the "odd" behavior of lucks before, but I was willing to dismiss it as being just that: a feeling. And feelings are so often so very wrong when it comes to accuracy. But the more I thought about it, and the more I thought about the specific examples being posted, the less easy it became to dismiss it as a perceptual error - as *most* of these things actually are.
You know, in the realm of accuracy-related issues, the devs have made some doozies. But in terms of the total number of real problems relative to the number of complaints about accuacy, the devs are probably batting 0.999 or better. That's an additional problem with this sort of thing. Denying there is a problem automatically is not a good idea, but its so very very right so very very often.
I should point out, that in the main, I've never been treated like that by the devs: they've patiently listened to almost every testing-related problem I bring to them. Its either because I'm always armed with a lot of helpful information useful to them, or because I'm armed with so much information its easier to check into the problem than read my testing information. Too accurate to ignore, or too annoying to ignore. Not sure which I am. Not sure which one its more useful to be, either.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
Great job, but one thing:
[ QUOTE ]
You would think by now we wouldn't trust the text descriptions for anything, and here's yet more proof that you shouldn't believe everything you read. Unless it comes from me, of course.
[/ QUOTE ]
Are you sure you're female? That kind of self-aggrandizement is more typical of the lamer (male) gender. You've been hanging around us too long.
The City of Heroes Community is a special one and I will always look fondly on my times arguing, discussing and playing with you all. Thanks and thanks to the developers for a special experience.
[ QUOTE ]
So Insights SAY +25% _Accuracy_ but what they mean is +25% (or not even that, but whatever) _To_Hit_Buff_.
[/ QUOTE ]
An interesting theory is that lucks *are* "+25% defense" relative to base 50% tohit. Perhaps in the distant past, all critter types had base 50% tohit (and they do again). Lucks date to that time when a player popped a luck, he was reducing net tohit by 25% for all attackers; +12.5 percentage points DEF = "+25% defense" colloquially. Given other things I've seen and heard, this is not a bad theory.
Insights, though; good luck coming up with a theory there.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
This goes a long way on why I would tell somebody new not to use one small insidght but two if you wanted to see any kind of notacble effect and why I always popped 3-4 Lucks when I face a Boss. This also goes a long way in my validation of using Resist Inspirations in Arena Battles vs the defense ones. I was indeed seeing a better return for the dollar. I can't tell you how much a post like this means when you have been saying for a while something is wrong.
I appreciate your hard work and I think the Devs should reward you for it. Thank you again.
Cuddles
[ QUOTE ]
[ QUOTE ]
Get this moved to Guides so it doesn't get deleted!!!
[/ QUOTE ]
I thought it made more sense as a discussion topic here and would draw more attention. I'll make sure this thread or a reposted version eventually makes its way to the Guides section for permanence.
[/ QUOTE ]
Board Rules
[ QUOTE ]
Thread/Post Content
1. Post in the correct forum and do not cross-post.
[/ QUOTE ]
[ QUOTE ]
That kind of self-aggrandizement is more typical of the lamer (male) gender.
[/ QUOTE ]
I'm not always sure I'm right, I'm just sure everyone who disagrees with me is always wrong.
Better?
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)
A lot of people have reported oddities with inspirations, especially lucks, and occasionally insights. I finally had a chance to test those out. Wasn't easy, let me tell you, and the tests were nowhere near the level of precision I would like to have, but I was able to see pretty conclusively variations in the behavior of the insps over what we all assumed was their behavior, based on their textual descriptions. So I had something of positive proof to take to the devs for enlightenment.
The textual descriptions are wrong: very very wrong.
Here is what the insps actually do, according to pohsyb:
DEF
S - 12.5%
M - 25%
L - 33%
ACC
S - 7.5%
M - 18.75%
L - 37.5%
RES
S - 10%
M - 15%
L - 20%
DMG
S -25%
M - 33%
L - 50%
Well, at least they got the damage and resistance ones right. But lucks are half the strength you think they are, and insights were created with a random number generator.
(Please don't reply that the insights aren't actually random, I know that.)
The good news is that SR isn't a little more than one small luck, its a little less than three. At least I think that's good news: all of you that were wondering why you sometimes needed to chomp four or five of these little guys now know why: it takes five of these guys to floor an even level boss in I6 (four in I7).
The bad news is that lucks and insights aren't actually balanced. In lower level zones where drops are often the smaller variety, lucks are a little stronger than insights. Not too much, but some. In RV, where drops are more often of the larger variety, insights are stronger than lucks.
So as tohit buffs become more common, from the BB to SC to WB to RV, insights also become more powerful than lucks. That's probably not a good thing either.
You would think by now we wouldn't trust the text descriptions for anything, and here's yet more proof that you shouldn't believe everything you read. Unless it comes from me, of course.
And this is worth noting: because lucks are defense, and therefore probabilistic in nature, not everyone will tend to see "average" performance. Some people will see behavior close to what they really do, and extrapolate their performance correctly, and some people will see bursty behavior that is out of line with their true behavior, and extrapolate their behavior incorrectly.
But in this case, it was the people who complained about seeing oddities that were right, and the people who thought they were working correctly that were wrong. Including me: sorry guys and gals, it took me a while to eventually come around and devote testing time to this.
The question, of course, is whether anyone else but me will know this, after this thread gets pushed down into page 42 with all the Issue 7 threads.
[Guide to Defense] [Scrapper Secondaries Comparison] [Archetype Popularity Analysis]
In one little corner of the universe, there's nothing more irritating than a misfile...
(Please support the best webcomic about a cosmic universal realignment by impaired angelic interference resulting in identity crisis angst. Or I release the pigmy water thieves.)