Paradox: High Damage Resistance result in Greater Critical Damage

[tdphys]

I did a more formal writeup of my arguments in the previous critical damage thread,  and gave hopefully a more snazzy title to catch DEV's eyes.    Apparently Sensuki thinks there might be difference to how damage is calculated in code vs what's been stated, so this might be a moot argument anyways. However,  I introduce to you:

 

"Balancing the Damage Threshold (DT) Paradox in Pillars of Eternity (POE)" 

poedt.pdf

 

I realize that this isn't really a paradox,  but the word is too cool

 

 

 

 

Edited January 11, 2015 by tdphys

[Sensuki]

Heh I like the "I have written academic journals before" font, page linings and everything"

 

Your ranges for attack resolution are wrong

 

its 

 

01-05: Miss
06-50: Graze
51-95: Hit
96+: Crit

and those values are actually incorrect in the source code at the moment as well (bug reported, and acknowledged by the devs).

Edited January 11, 2015 by Sensuki

[tdphys]

Hmm... is the percentile roll in integer or float?

 

I've modified the pdf a bit...

Edited January 11, 2015 by tdphys

[tdphys]

Heh I like the "I have written academic journals before" font, page linings and everything"

 

 

It's just Latex

[Sensuki]

Check your PM Box, I can even paste you the source code as proof.

[drunetovich]

So basically you propose to proportionally increase armor mitigation on crit, like 2x DT against 2x crit? This may lead to damage being actually worse on crits and that would indeed be a paradox. We could pull it off with DR type armor, but DT armor is not so flexible esp for real time combat.

[Sensuki]

Damage mults just need to be additive rather than multiplicative

[nipsen]

All in for linear damage models, then?

 

I'd just like to take this opportunity to congratulate you all for ensuring that the story-telling and writing will be the exclusive focus of the game once it's reviewed, though. Since the combat system description will feel comprehensive and redundant after about two sentences.

[tdphys]

No.  critically modified damage would always be greater then regular hit damage.   It's just the armor DT would get applied before the crit bonus.   It would make weapon type and stats and DT more meaningful as counters to each other.  Crit would still be important,  especially against unarmored and for light weapons that could have armor piercing (higher min damage).  

 

If you looked at the amount of damage blocked compared to the same damage roll and crit modifier against an unarmored target,  the amount of damage blocked rises with the crit value (hey, like DR) .  But it never decreases the total damage amount done by a crit as compared to a regular hit.

[nipsen]

So wearing armor is a liability in terms of damage in the long run.

 

You're literally proposing to "balance" the reduced damage while wearing armor, by increasing the amount of damage they receive on average over time.

 

Not with speed or agility or number of attacks, depletion of stamina or weaknesses to certain types of damage, while specific damage could bypass the armor completely, etc. But by simply increasing the damage dealt on critical hits, so that a sharpened toothpick can bypass the armor if you survive for enough time.

 

Honestly, I can not see why either Sensuki Reversecompile Inc. or Obsidian Entertainment haven't brought you on as a game-designer yet.

[tdphys]

So wearing armor is a liability in terms of damage in the long run.

 

You're literally proposing to "balance" the reduced damage while wearing armor, by increasing the amount of damage they receive on average over time.

 

 

 

 

All of  speed, agility , number of attacks, depletion of stamina or weakness  to types of damage figure into the calculations the same way.  

 

What I'm trying to do is reduce the chance of getting critical hits as the sole primary path to damage  amplification,  and a way to make armor values relevant in the face of geometric damage progression ( every modifier multiplied together makes a lot of damage). 

 

Math now:

 

damage*mod1*mod2*mod3*mod4  - dt  = loads of damage and irrelevant armor

 

Proposed math:

 

(damage*mod1*mod2 - dt)*mod3*mod4  = less loads of damage and relevant armor.

 

It makes thematical sense that mod1 and mod2 are things like stats and weapon type,  while mod3 and mod4 are situational things like backstab and crit.

 

For people who like light weapons,  you give a higher pierce rate through dt and a greater backstab bonus.  This is also thematic.  

 

 

 

 so that a sharpened toothpick can bypass the armor if you survive for enough time.

 

 

Right now I'm worried that the optimal way to fight Pillars is to run around as a naked rogue doing massive critical damage with a sharpened toothpick.

Edited January 15, 2015 by tdphys

[nipsen]

 

Right now I'm worried that the optimal way to fight Pillars is to run around as a naked rogue doing massive critical damage with a sharpened toothpick.

 

Really, now.

 

So do you think that this could have something to do with how the attack roll is relative to the armor class? In the sense that a lower armor class that is beaten by a significantly stronger attack roll, has a higher chance of receiving a critical hit, compared to an armor class that barely is beaten by the attack roll?

 

So that when damage tresholds are tweaked down, and the threat range and attribute modifiers are tweaked into a class setup with fairly high base default values and no weaknesses, as apparently Obsidian Q&A and people around here have all wanted -- then all characters tend to do critical damage often, have no weaknesses. And this ends fights very suddenly for no reason. Etc.

 

Because, you see, I've never heard about this before, or ever thought or ranted at length about exactly this particular ****ing point for pages upon pages, of course. And it's not at all sad to see someone - after everyone Obsidian listens to got what they wanted - having to make the point that the mobs are now simply too weak, and that this makes the combat unsatisfying.

 

But at least your paper seems to be based on something valid. The last "analysis" was based on a complete misunderstanding of how maths work. 

[Sensuki]

The last "analysis" was based on a complete misunderstanding of how maths work.

Hahah. Nope.

[nipsen]

 

The last "analysis" was based on a complete misunderstanding of how maths work.

Hahah. Nope.

 

When you think any correct solution of a random equation will always answer the question you have at hand - then yes, you have fundamentally misunderstood how maths work.

 

You've also said you have put me on the ignore list. Which I think is a good idea, since you do not deal well with criticism. So please keep me on the ignore list.

[Sensuki]

The last paper was based on in-game testing and 'engineering'. You wrote it off the moment you saw it, and I'm 100% positive you read almost none of it either. I have however enjoyed how you've been blaming me for all your problems with the game

 

Pretty sad of you to request a refund over it though IMO.

Edited January 15, 2015 by Sensuki

[nipsen]

I have however enjoyed how you've been blaming me for all your problems with the game

 

Pretty sad of you to request a refund over it though IMO.

I suppose it says something about you when you think your opinion is that important. Specially when what I said is that I was extremely disappointed when Obsidian looked for an opportunity to throw out the original design. And literally used isolated ditto-opinion on this forum to justify it as "what most people wanted".

 

That Obsidian could do something like that - in spite of the very obvious problems implementing the changes would cause, even if they ever managed to "balance" it afterwards. And in spite of the amount of work it would result in, that would have been better spent on fixing mechanical problems that apparently still exist. That's what disappointed me. That they didn't have the foresight to ignore the internet-noise, stick with their design - and really thought pleasing some super-fans was more important than designing a consistent ruleset. With the deep role-playing they obviously advertised and promised they would choose, now that they didn't have to sell to a publisher and argue for "common appeal".

 

It's also why they allowed a refund, which they have no obligation to process (and normally refuse). It's also what had certain.. extremely well-placed people close to the design process argue that the focus of the game for them now had become the dialogue and story-telling. While that the combat system and character build system was at least more dynamic than in the IE games (which doesn't take very much). Saying then that therefore the Kickstarter promise was fulfilled, technically.

 

Frankly, I'm also not the only person who rolled their eyes when you and matt came up with a model that had some very obvious flaws. Flaws that you ignored very happily. Such as that you "proved" the PoE attack rolls and damage model were skewed, by assuming a completely linear system. You literally set up constants for a special case, and ignored how the rolls would be variable. Ending up with an analysis that first made the PoE system seem unbalanced. And you then proposed changes to it to make the system statistically balanced.

 

When what you really did was simply ignore how the design was meant to function, forcing the data over your assumptions of how the system worked(a d&d model). While swearing that the math you had set up proved your completely disconnected point, because the math you presented was internally consistent.

 

Maths doesn't work like that. Everyone knows that. Obsidian folks know that. But they chose to not ignore how very loud people, people they assumed were their superfans, had such a ridiculously hard time understanding the original rulset proposal. And how they started to noise about betrayed Kickstarter promises. That was what they acknowledged - that some of you were incapable of grasping how the system worked, for whatever reason.

 

So they thought this was a common problem - that they underestimated their audience like this - is something that helped motivate the changes they made. And that was disappointing. I didn't think Obsidian had that attitude towards developing a game, and I really don't think people who like Obsidian games typically think that either. Basically: We didn't go to the PoE kickstarter to fund a "AAA"/common denominator game that Obsidian could have otherwise gone to a publisher to fund. To essentially get a streamlined and shallow game, that also had bad graphics. That wasn't the point with the Kickstarter project.

Edited January 15, 2015 by nipsen

[Sensuki]

Once again, you're basing all of that off a misconception because you didn't read the paper properly. The difference between a 1% chance of a 100% increase and a 100% chance of a 1% increase was illustrated and accounted for in Matt's formulas.

 

Their changed designs also did not please me and were not a result of that paper. Something else you conveniently keep ignoring for some reason.

 

Just like you, I wanted interrupt to stay in the attribute system.

Edited January 15, 2015 by Sensuki

[nipsen]

Well, yes, I know that. But what they reacted to, and what Obsidian internal testing and apparently Paradox Q&A latched onto, was the idea that the first design was too complicated. As in they didn't make any changes based on your suggestions directly. But were very sensitive to any evidence that what they began with was too difficult to understand. As well as extremely open to the idea that giving fans what they wanted, even if it wasn't a good solution, was a good idea.

 

Which, as I said at the time as well, was what that paper was really arguing. That "even" the hardcore fans couldn't understand what Josh had dreamt up. That the design was simply too different from the IE games, and therefore not possible to sell. It's not the case, of course, when you actually take a look at it (or given that the presentation of the system as given to the player had been better constructed). But that seems to have been the narrative - that some of you seem to have shared in a sense with Obsidian internal testing.

[IndiraLightfoot]

Nipsen: Interrupt is back, and the entire attribute system has been overhauled. Hopefully, you'll buy the game before release and enjoy it.

[Sensuki]

I have a feeling they will change it though as I don't think the upcoming implementation is going to cut it. It is nice to see Interrupt back though.

[illathid]

 

 

Proposed math:

 

(damage*mod1*mod2 - dt)*mod3*mod4  = less loads of damage and relevant armor.

 

It makes thematical sense that mod1 and mod2 are things like stats and weapon type,  while mod3 and mod4 are situational things like backstab and crit.

 

I don't think this is a good idea. I don't think any damage modifiers should be multiplied together. If you have a 50% damage modifier and and 25% damage modifier with multiplication you have an effective damage bonus of 87.5% not 75% as you'd expect. Plus by putting some modifiers before DT and some after you've effectively made 2 different types of damage modifiers. All this does is make damage calculation more complicated and inpenterable to your average player. I'd be fine with either of the following:

[Damage*(mod1+mod2+mod3+mod4)] - DT = Total Damage

(Damage - DT)*(mod1+mod2+mod3+mod4) = Total Damage

Edited January 16, 2015 by illathid

[tdphys]

 

 

 

Proposed math:

 

(damage*mod1*mod2 - dt)*mod3*mod4  = less loads of damage and relevant armor.

 

It makes thematical sense that mod1 and mod2 are things like stats and weapon type,  while mod3 and mod4 are situational things like backstab and crit.

 

I don't think this is a good idea. I don't think any damage modifiers should be multiplied together. If you have a 50% damage modifier and and 25% damage modifier with multiplication you have an effective damage bonus of 87.5% not 75% as you'd expect. Plus by putting some modifiers before DT and some after you've effectively made 2 different types of damage modifiers. All this does is make damage calculation more complicated and inpenterable to your average player. I'd be fine with either of the following:

[Damage*(mod1+mod2+mod3+mod4)] - DT = Total Damage

(Damage - DT)*(mod1+mod2+mod3+mod4) = Total Damage

 

What about 

 

[Damage*(mod1+mod2) - DT]*(mod1+mod2)    

 

Okay, I'm hiding a multiply in there...   I'll admit, your calculations are easier to describe.  I just like pre and post DT modifiers

[illathid]

What about 

 

[Damage*(mod1+mod2) - DT]*(mod1+mod2)  

 

Okay, I'm hiding a multiply in there...   I'll admit, your calculations are easier to describe.  I just like pre and post DT modifiers

 

 

Yeah, and I still have no idea why. I think you may perhaps be someone who worships at the altar of simulationism, which as I've said many times before needs to be killed with fire (simulationism that, you're fine ).

 

And you run into the same calculation problems if you have pre and post DT modifiers. I see absolutely no beenfit from doing it that way at all.

Edited January 16, 2015 by illathid

[tdphys]

 

What about 

 

[Damage*(mod1+mod2) - DT]*(mod1+mod2)  

 

Okay, I'm hiding a multiply in there...   I'll admit, your calculations are easier to describe.  I just like pre and post DT modifiers

 

 

Yeah, and I still have no idea why. I think you may perhaps be someone who worships at the altar of simulationism, which as I've said many times before needs to be killed with fire (simulationism that, you're fine ).

 

And you run into the same calculation problems if you have pre and post DT modifiers. I see absolutely no beenfit from doing it that way at all.

 

 

I do like simulationism,  it's true,  but what I like more is complexity that is meaningful and not redundant.

 

1.  when per and might both were the same modifier,  this seemed redundant.

 

2.  with all these modifiers being multiplicative, damage amplification seem to make DT redundant.

 

So I'd probably be a big fan of your [damage-dt]*(additive mods) because it makes armor less redundant.  I just can't help myself from saying, 

hey,  there's a cool spot to add a mod (after damage)  that's unique, and then look for a place to make it thematic (might seems best).

 

So, maybe it's the reverse.  I'm a big fan of interesting models ( I want to expand the tactical/strategic depth of the game )  , the fun part is making it

thematic.    The challenge is to make it more complex (then say DnD), but accessible and still able to be balanced.  

[illathid]

Game mechanics should serve a purpose, and I can't for the life of me figure out what purpose a pre DT damage modifer would serve.

 

The effects of such a modifier are:
1) reduce the the benefit of oponents DT
2) make damage calculation more complex

3) be thematic (?)

 

1) is already achieved with penetration/pass through or whatever it's called in game, making this benefit redundant. For 2) this is objectively a bad thing. And 3) could potentially justify the redundancy of 1) or the problems with 2); but you just keep saying it is thematic, not explaining how it's thematic. In my opinion at least.