Quote:
Is this your admission that you're wrong? Thank you.|
Originally Posted by Obitus
Yeah, great. Meanwhile, if you're taking 1000 damage per second and your only source of mitigation is 5% DEF, then you're screwed anyway without huge outside intervention.
Again, the point isn't that your numbers are wrong in theory. The point is that your numbers are practically useless. |
I can give you all sorts of numbers spanning any kind of possibility.
Quote:
|
Originally Posted by Obitus
No, my claim was that the opponent's damage is unknowable, and that therefore it's better to concentrate on the practically plausible bounds of your survivability.
|
Quote:
|
Originally Posted by Obitus
It's not clear why you'd prefer a metric for making build decisions that relies on knowing the precise damage output of the opponent(s), rather than a metric that relies on simply knowing your own regeneration/healing rate and resistance values.
|
Quote:
Apparently you can't. You failed to come up with a relevant example where 5 hp/sec is better than 5% DEF on a character who already has 30% DEF and 20 hp/sec regeneration. All examples you did provide showed that the character was infinitely survivable regardless.|
Originally Posted by BunnyAnomaly
Is this your admission that you're wrong? Thank you.I can give you all sorts of numbers spanning any kind of possibility.
|
Your best rebuttal to my method for evaluating the next example (5% DEF versus 50 HP/sec regen on a character with 20 HP/sec initially and no other mitigation) was an example where the character will die in an eyeblink without sufficient buff/heal support that his 5% DEF would be trivialized.
Quote:
| I am abrupt and confrontational because you are lying about what you said and wish to rewrite history. |
Quote:
|
Originally Posted by Iggy_Kamakaze
Nice build
 |
Quote:
Ah, I see. You think the following two quotes (of mine) are so self-contradictory that they constitute intentional dishonesty:|
Originally Posted by BunnyAnomaly
I am abrupt and confrontational because you are lying about what you said and wish to rewrite history.
|
"No, my claim was that the opponent's damage is unknowable, and that therefore it's better to concentrate on the practically plausible bounds of your survivability."And
"It's not clear why you'd prefer a metric for making build decisions that relies on knowing the precise damage output of the opponent(s), rather than a metric that relies on simply knowing your own regeneration/healing rate and resistance values."I see no glaring self-contradiction there. Certainly no evidence of a lie. For all of your bluster about mathematics, it appears you have very little experience reading for context. My metric (and Werner's, and a whole host of other people's) does only depend on knowing the character's regeneration/healing/RES/DEF/HP. That's what makes it practically useful
Notice that I never said that the absolute value of DEF in the terms you choose to describe it (in this case, virtual regeneration) wouldn't change with the opponent's incoming damage. I never even said that the terms you choose to use are incorrect on paper. I do, however, maintain that your terms are impractical for making build decisions.
Tell me, exactly what incoming DPS figure were you looking at when you designed your DEF-capped WP Tanker? How did you know when exactly it would be worthwhile to switch from adding more regen to adding more DEF?
It couldn't be because you instinctively understand that DEF multiplies regen. Nah.
Quote:
|
Originally Posted by Iggy_Kamakaze
Nice build
|
I didn't check this thread for a few days and was wondering why it got to three pages even though the question was answered in the second post.
Gotta agree with Obitus, BunnyAnomaly ; you really are doing the worst kind of theorycrafting here. Neglecting debuffs (particulary defense debuffs) is a fatal mistake when arguing the importance of adding defense relative to the amount of defense you already have, and pulling numbers out of thin air that don't have any application in actual gameplay doesn't do anything to strengthen your point (unless that point is "I want to win a forum argument about something completely unrelated to CoH).
following Bunny's math/argument then the answer to the OP's question is:
The Defense is better as long as the incoming damage is more than 402/0.16= 2512.5 hp per minute, if it is less the Regen will mitigate more damage.
with of course two added caveats -- as earlier posters mentioned -- that the damage has a smashing and or lethal component, if it doesnt the regen wins.
and that you do not care about additional secondary effects you will suffer using regen instead of defense mitigation. If you do care, the defense wins.
Quote:
Indeed. This was noted in my very first post that there are other advantages to defence.
|
Originally Posted by Hai Jinx
and that you do not care about additional secondary effects you will suffer using regen instead of defense mitigation. If you do care, the defense wins.
|
For my toon, it seems to have been the right trade. Feels that way anyway.
Well, I'm way behind and may never catch up, but I guess I'll start plugging away.
Quote:
No, the first method provides AN answer. So does a magic 8 ball. Providing an answer, even a CORRECT answer, doesn't mean that the method used was correct. Also, it only provided an answer because we plugged in a random value, an assumed amount of enemy damage output. This assumed amount was fixed regardless of the current level of damage mitigation and regeneration, which has nothing to do with the way people actually play, which is looking for a challenge but not suicide. Therefore, in actual play, someone with no mitigation is going to be facing much less incoming damage than someone with 95% mitigation.|
Originally Posted by BunnyAnomaly
You've displayed exactly why the 2nd and 3rd methods are poor and why they cannot be used to make decisions between two mutually exclusive options. You cannot answer the question posed in this thread by them. Note that the first method does in fact answer the question.
|
Quote:
|
Originally Posted by BunnyAnomaly
But the problem goes further than that. Not only can you not answer the question, you also make a case for false conclusions.
If someone read this, they might believe that the answer depends upon what existing defence you have. However that is false. |
I think that most people, when asking questions of survivability, care about what the build can survive. I honestly think that when someone asks a question like, "which is better to add, this mitigation or this regeneration", what they care about is what each would allow them to DO. What each would allow them to SURVIVE. THAT answer depends on what existing mitigation and regeneration you have.
What I think most people DON'T care about when asking a question about survivability is how well they mitigate and regenerate one boss, or any one specific fixed level of enemy damage. In table #1, they already have sufficient survivability at 0% defense. In our overly-simplistic terms that ignore discrete effects, their green bar never moves. It doesn't matter if they have 0% defense or 45% defense, because either way, the boss can't kill them. So table #1 tells us nothing useful as is, because it doesn't tell us how much we can SURVIVE. To know what we can survive, we need to take the extra step of calculating the maximum possible incoming damage based on our mitigation and regeneration. That takes us to tables #2 and #3.
Do you not agree that tables #2 and #3 are reasonable representations of what these levels of mitigation and regeneration allow us to SURVIVE?
Do you not see how most people might care more about what they can ACTUALLY SURVIVE than about what table #1 tells them?
Do you really not care about what your mitigation and resistance allows you to actually survive? Is that really not a concern to you when talking about survivability?
"That's because Werner can't do maths." - BunnyAnomaly
"Four hours in, and I was no longer making mistakes, no longer detoggling. I was a machine." - Werner
Videos of Other Stupid Scrapper Tricks
Quote:
But what happens if we change the incoming damage?|
Originally Posted by BunnyAnomaly
Dead wrong.
I will write a lengthy reply to explain why. Compare 10 regen/second to 5% defence, assuming 100 dps incoming before defence factored. You would say it depends on how much defence you have already. That is why you do NOT use method 2 or 3. Here's why: 0% defence to 5% defence. You take 45 hp/s of damage (base 100 - 55). 0%, but with an additional 10 hp/second of regen. You take 40 hp/s of damage (base 100 -50 - 10) Answer: Take the regen (5 hp/s better) Next up: 40% defence to 45% defence. You take 5 hp/s of damage (base 100-95) 40% still, but with 10hp/s regen. You take 0 damage (Base 100 - 40 - 10) Answer: Take the regen (5hp/s better) You are 5 hp/s ahead regardless of your initial defence. It doesn't matter about defence UNLESS you are going to exceed a cap. Must I provide more additional proof? Just because a system is used for a long time doesn't make it any less stupid. |
Compare 10 regen/second to 5% defence, assuming 500 dps incoming before defence factored.
0% defence to 5% defence.
You take 225 hp/s of damage (base 500 * 45%).
0%, but with an additional 10 hp/second of regen.
You take 240 hp/s of damage (base 500 * 50% - 10)
Answer: Take the defense (15 hp/s better)
Next up:
40% defence to 45% defence.
You take 25 hp/s of damage (base 500 * 5%)
40% still, but with 10hp/s regen.
You take 40 damage (Base 500 * 10% - 10)
Answer: Take the defence (15 hp/s better)
You are 15 hp/s ahead regardless of your initial defence.
So when making this comparison, the incoming damage has a dramatic effect on whether you should take the defense or the regeneration. If I can just randomly pick a value to get either answer, surely this is no better than a magic 8 ball at answering the question. What we NEED to make this method more reliable is a WAY of choosing an amount of incoming damage that this player would actually see IN THE GAME.
Hey, wait, I just though of a way we could do that! We could assume that the player likes a challenge, but doesn't like to face plant constantly. So we could assume that the player will be facing an amount of incoming damage that would provide that challenge without killing them. It's a little over simplistic, but we could use an immortality line calculation to figure out how much DPS the player can survive, and plug THAT in as the enemy damage output instead of picking at random!
Hmmm, that's strange. When we do THAT, it gives us a different amount of damage to plug in depending on what the player current has for mitigation and regeneration. But then that means that the answer to this question would depend not on some randomly-chosen value for damage, but instead on the current level of mitigation and regeneration. But that's different at different levels of defense!
Ah, well. It must be wrong then, because we already know that additional survivability can have nothing whatsoever to do with our current level of defense. Pity, it seemed like such a promising way to avoid a random variable in the survivability calculation. Guess we'll keep picking numbers at random, keep shaking that magic 8 ball.
Must I provide more proof? Just because you can plug in a random number that supports your answer to a question doesn't make it any less stupid to do so.
"That's because Werner can't do maths." - BunnyAnomaly
"Four hours in, and I was no longer making mistakes, no longer detoggling. I was a machine." - Werner
Videos of Other Stupid Scrapper Tricks
Quote:
The best part is that the immortality line standard already has the effect of additional DEF on debuffs implicitly baked in. Each additional point of DEF is more valuable than the last in both cases.|
Originally Posted by Werner
Hey, wait, I just though of a way we could do that! We could assume that the player likes a challenge, but doesn't like to face plant constantly. So we could assume that the player will be facing an amount of incoming damage that would provide that challenge without killing them. It's a little over simplistic, but we could use an immortality line calculation to figure out how much DPS the player can survive, and plug THAT in as the enemy damage output instead of picking at random! Hmmm, that's strange. When we do THAT, it gives us a different amount of damage to plug in depending on what the player current has for mitigation and regeneration. But then that means that the answer to this question would depend not on some randomly-chosen value for damage, but instead on the current level of mitigation and regeneration. But that's different at different levels of defense! |
Bunny's model works in theory. It's just not particularly useful in practice. Any model that predicts that a given amount of regen is more useful than a given amount of DEF -- but only when the character in question will survive indefinitely anyway -- isn't worth the paper it's printed on.
Likewise, if you have to raise incoming damage so high that no one could possibly survive -- without so much buff support that his build decisions are irrelevant, anyway -- to prove that a given amount of DEF is more valuable than a given value of regeneration, then you've made a pointless observation.
Bunny started quite reasonably in this thread. After a couple of posts, though, he apparently decided that it was more fun to denigrate his opposition. I guess he was bored.
Quote:
|
Originally Posted by Iggy_Kamakaze
Nice build
|
Werner presumably didn't read anything I wrote else he would know that the decision between regen and defence is going to change dependant on how much damage you expect to face. I only repeated that for I don't know how long while Obitus said otherwise. Obitus was wrong. If Werner had have bothered to read he wouldn't've wasted his time writing that.
Same reason why he doesn't understand IRR.
Or why you can't decide between "+100% survivability through defence" or 5 hp/s regen, because writing that can't decide anything.
I have provided a metric for determining it if these are your considerations, and it was used before to provide the correct answer to this question by someone other than me, using this very method.
Werner says we need a magic 8 ball for my solution. That's because Werner can't do maths.
Obitus thinks the solution is only useful for when invincible survivability is reached. That's because Obitus can't do maths too.
Answer was provided using my method already. Your method describes nothing useful. This is it here, because apparently you don't read and just write bad math the whole way through:
Quote:
Solution provided.|
Originally Posted by Hai Jinx
following Bunny's math/argument then the answer to the OP's question is:
The Defense is better as long as the incoming damage is more than 402/0.16= 2512.5 hp per minute, if it is less the Regen will mitigate more damage. with of course two added caveats -- as earlier posters mentioned -- that the damage has a smashing and or lethal component, if it doesnt the regen wins. and that you do not care about additional secondary effects you will suffer using regen instead of defense mitigation. If you do care, the defense wins. |
This one I particularly like. Because it came from Werner's own workings yet he doesn't understand the importance.
Quote:
| You are 15 hp/s ahead regardless of your initial defence. |
Taking this a bit further -
for any given level of defense ...
and any given level of regeneration ...
There will be a incoming damage intersection point. Below which the regeneration will always mitigate more damage, above which the defense will always mitigate more.
So there is a point where 402 hp/m will exceed even the defense soft cap for mitigation. (893.3 Incoming DPM actually).
Quote:
Precisely.
|
Originally Posted by Hai Jinx
Taking this a bit further -
for any given level of defense ... and any given level of regeneration ... There will be a incoming damage intersection point. Below which the regeneration will always mitigate more damage, above which the defense will always mitigate more. So there is a point where 402 hp/m will exceed even the defense soft cap for mitigation. (893.3 Incoming DPM actually). |
Our standard model measures survivability in terms of what a build can survive. Although you've continually screamed that we're bad at maths, and that we're liars, you haven't even attempted to show that this model does NOT accurately measure what a build can survive. Instead, you just keep doing your own maths, as if proving your maths true somehow proves our maths false. Sorry, but that's not the case. Your maths are like step one - understanding mitigation. Our standard model is like step two - understanding how mitigation relates to survivability, to what a build can survive. And what a build can actually survive is the much more interesting and relevant question.
Hopefully most people can follow the argument and move from step one to step two, even if you obviously cannot. Hopefully most people understand that what matters when discussing survivability is what a build can survive.
I'd be happy to do some in-game testing to demonstrate the survivability numbers in tables #2 and #3. Well, not happy, because it's a huge waste of time, but willing. But since you already believe we're liars, I'm sure it wouldn't do any good.
"That's because Werner can't do maths." - BunnyAnomaly
"Four hours in, and I was no longer making mistakes, no longer detoggling. I was a machine." - Werner
Videos of Other Stupid Scrapper Tricks
Quote:
Except that you cannot answer the question posed by the thread, and I can.|
Originally Posted by Werner
Well then, just some last thoughts before you take your final ad hominem shot, and then we can part.
Our standard model measures survivability in terms of what a build can survive. Although you've continually screamed that we're s***house at maths, and that we're liars, you haven't even attempted to show that this model does NOT accurately measure what a build can survive. Instead, you just keep doing your own maths, as if proving your maths true somehow proves our maths false. Sorry, but that's not the case. Your maths are like step one - understanding mitigation. Our standard model is like step two - understanding how mitigation relates to survivability, to what a build can survive. And what a build can actually survive is the much more interesting and relevant question. Hopefully most people can follow the argument and move from step one to step two, even if you obviously cannot. Hopefully most people understand that what matters when discussing survivability is what a build can survive. I'd be happy to do some in-game testing to demonstrate the survivability numbers in tables #2 and #3. Well, not happy, because it's a huge waste of time, but willing. But since you already believe we're liars, I'm sure it wouldn't do any good. |
Because I can describe mitigation, you convert it into a useless statistic unfit for comparison.
That's why you cannot answer which is better.
Except he already answered, second post on the topic. You then proceeded to put your efforts towards making yourself look like an ******, and you've done a fairly good job at that.
Quote:
While I appreciate it, mine wasn't a mathematically-derived answer, simply "very likely". So in that sense Bunny's point stands, at least in the sense that I never did answer the question mathematically (though not that I am incapable of doing so).|
Originally Posted by Nihilii
Except he already answered, second post on the topic.
|
With a mitigation calculation, and with some qualitative instead of quantitative observations about damage types and so on, this can be answered mathematically if you know how much enemy damage output the person expects to face. Since All Hell seemed unlikely to face such a small amount of damage as is required to be at the break even point, my "very likely" answer seemed sufficient. Basically, the question was simple enough that it could be answered based on experience (me), or by simple reference to the basics of damage mitigation (Shred Monkey and BunnyAnomaly).
With a survivability calculation, we CALCULATE how much enemy damage output the person should be ABLE to face. So instead of assigning an arbitrary value based on gut instinct or something lower in the intestinal tract, we calculate it. To calculate it, we need to know your defense, resistance and regeneration. And since in this case it differs by type, we even need to know this data by type. For that matter, we need a reasonable estimate of how much damage enemies put out by type. Fortunately, someone (Edit: Besserwisser) has already done the heavy lifting to gather that information, and I have that table of data. So this method involves a whole nebula of variables. It's unnecessary for a case this simple, but necessary in the general case. Fortunately, other than the mentioned table, everything you need is in Mids', so this isn't a practical problem when evaluating actual builds you're considering.
I of course don't have all those variables for All Hell's build, so Bunny is correct that I cannot definitively answer the question. But since everyone already agrees on the answer, this is just a demonstration, and we'll make some assumptions.
regeneration and healing = 30 HP/S -> 23.3 HP/SGiven that, here's what my survivability calculations say on the subject, with the usual cautions in regard to immortality-line-based mathematics that I won't go into unless someone wants more background:
toxic/psionic resistance = 0%
other resistance = 20%
smashing/lethal defense = 15.5% -> 33%
positional defense = 0%
fcen defense = 27% -> 33.3%
toxic/psionic defense = 0%
current build survivability = 111So go with the defense.
defense build survivability = 146
Now, I can certainly choose numbers that say otherwise. To take a trivial example, let's say we didn't have inherent regeneration or greens or any default way to recover health, and All Hell was discussing changing his regeneration from 6.7 HP/S to 0 HP/S. Hopefully it is obvious that in this case, the regeneration is the way to go, else you have no damage recovery and will eventually be killed. But we don't actually have to guess and assign this variable arbitrarily as I've done. We can instead just plug in All Hell's real numbers, if we had them.
Hmmm, I intended to post the spreadsheet with the calculations, but I'm having trouble uploading it. Apologies for posting numbers without the calculations. Hopefully I'll get it uploaded later. My spreadsheet actually uses Bunny's definition of mitigation rather than the one in common use on the forums, because like him or her, it made sense to me to calculate from enemy damage output instead of in comparison to a character at 0% defense and 0% resistance. If you want what I'd call forum standard numbers, just divide my given survivability by 2. Since the relationship is that trivial, I never felt the need to change my spreadsheet to match forum consensus on the definition of mitigation.
I should also mention that the spreadsheet isn't overly useful for low levels of survivability running easy content. In cases like that, time exposed to danger is very short, and you're in a kill them before they kill me scenario. The survivability number here is INDEFINITE survivability. What you can just stand there and take for minutes on end. That's what tends to matter for the high end survivability folks, because we're getting ourselves into fights that last a long time. But it isn't what's important for low survivability, low difficulty play. Based on the assumptions I've made above, All Hell's build options are probably on the lower end of what I'd want to give the immortality line treatment. Something like 60 second survivability might be more appropriate, but I've never bothered with such things, because every Scrapper I care about is well into the immortality line realm. And now we're also into a discussion of the weaknesses of the immortality line approach to survival, which I wasn't going to go into, so we'll let that go for now.
Edit: I can upload again, so here is the spreadsheet.
"That's because Werner can't do maths." - BunnyAnomaly
"Four hours in, and I was no longer making mistakes, no longer detoggling. I was a machine." - Werner
Videos of Other Stupid Scrapper Tricks
Good grief, you guys.
Support Guides for all Corruptor secondaries and Fortunatas
The Melee Teaming Guide for Melee Mans
The Melee Teaming Guide for Melee Mans
My only regret is LB and Arachnos. They seem to melt Def and seem to be harder than before.
Quote:
I saw bad math in the 7th post and just had to facepalm. You know it's bad when I'm too ashamed of someone's stupidity to call them out on it.
|
Originally Posted by Silas
Good grief, you guys.
|
I've completed the testing I proposed using my Super Reflexes Scrapper, Sergei. Based on 45 minutes of standing in an easy group that shouldn't trigger passive resistance except on the rarest of unlucky streaks, using HeroStats to track hits, and calculating backwards to enemy damage output, I get this:
Boss = 93 DPSBosses and lieutenants actually put out less damage than this at a higher accuracy modifier, but that's equivalent to just saying they have this damage output for our purposes. I should also note that these bosses were content to shoot from range, and that their true damage output is much higher in melee, but their actual behavior is sufficient for our purposes. We just need a yard stick.
Lieutenant = 34 DPS
Minion = 17 DPS
Now what will happen as we slowly reduce Sergei's defense from the soft cap by toggling various powers and deleting key IOs?
The survivability calculation gives us the following table for what he can expect to survive with his 23.45 HP/S regeneration rate, and assuming 45% (Edit: Ah, no, I used 40%) for the passive resists (which for these enemies seems about right to me from standing there and watching my hit points blink red with a very high passive resistance for the past two hours):
- 45% defense = 782 DPS = 4 bosses, 7 lieutenants and 10 minions
- 39.61% defense = 376 DPS = 2 bosses, 3 lieutenants and 5 minions
- 33.53% defense = 237 DPS = 1 boss, 2 lieutenants and 4 minions
- 21.04% defense = 135 DPS = 1 boss and 1 lieutenant
- 11.67% defense = 102 DPS = 1 boss
And here are my actual results with the length of time I survived the indicated spawn before stopping the test, in each case I think enough to call it "indefinite survivability" even though obviously I didn't stand there forever:
- 45% defense = 4 bosses, 5 lieutenants and 7 minions (mostly green for 13 minutes)
- 39.61% defense = 2 bosses, 3 lieutenants and 6 minions (often blinking red for 15 minutes)
- 33.53% defense = 1 boss, 2 lieutenants and 5 minions (often blinking red for 10 minutes)
- 21.04% defense = 1 boss and 1 lieutenant (often blinking red for 10 minutes)
- 11.67% defense = 1 boss (mostly yellowish green for 14 minutes, but adding a lieutenant killed me)
The blinking red runs are obviously very near the immortality line, about to cross over. A lucky series of hits could and would eventually kill me. These runs were done at or just above the calculated survivability number. The mostly green run at 45% defense indicates that I could survive significantly worse, which is in agreement with the calculation as well. The mostly yellowish green run for 14 minutes indicates I probably could have done slightly better, but the calculation shows adding a minion would take us over the line, and certainly adding a lieutenant did. I did a handful of runs over the line as well, resulting in deaths. Just one more indication that the blinking red ones are very close to the limit.
So we have a very high agreement between theory and observation in regards to how defense affects how big and how nasty of a spawn you can actually survive in the actual game. And what does this tell us about the 5% defense = 5% defense point of view?
- Adding 9.37% defense when we already had 11.67% allowed us to add a lieutenant to our group, and the fight became more dangerous.
- Adding 5.39% defense when we already had 39.61% allowed us to add 2 bosses, 2 lieutenants and a minion to our group, and survive it much more easily than before.
Now, if you consider what you can survive in the game to be a "useless statistic", then by all means ignore these results and the underlying mathematics.
But if you actually care about what you can survive in the game, then you'll want to pay attention to what we might call the standard model of survivability.
Edit: I seem to be able to upload files again. So here's a graph of theory vs. observation for the more visual minded, including the best fit linear trend line, which is obviously a poor model for our data. The only point that doesn't match theory is the last one, and that was due to difficulty getting a more difficult group. I was one away from the aggro cap there. I could have easily survived more enemies, meaning the point should really be somewhat higher, but I cannot say how much higher merely from observation, so it sits where it sits.
"That's because Werner can't do maths." - BunnyAnomaly
"Four hours in, and I was no longer making mistakes, no longer detoggling. I was a machine." - Werner
Videos of Other Stupid Scrapper Tricks
Quote:
Actually the 4th Post (Shred Monkey's) and the 7th Post are the same equation. Later posts Bunny made use other comparisons. Maybe you meant the 17th post. |
Originally Posted by Umbral
I saw bad math in the 7th post and just had to facepalm. You know it's bad when I'm too ashamed of someone's stupidity to call them out on it.
|
-------
Werner's data suggests however that the linear relationship they described isn't what actually happens. (although its pretty close for 16%)
That curve says there is more happening than damage/defense, otherwise it would be the green line more or less.
----
As to Bunny's personal attack's they really don't predict anything in particular, I am unsure why they were included in those posts.
But then, that's been clear from the outset, and if you were debating in good faith, you would have made at least a token effort at addressing that point with a straight face. Did you answer my question in the latest post? No.
You just like belittling people who disagree with you. Enjoy your 45% DEF-capped Willpower Tanker. We all know that that last 5% is exactly equivalent with the first 5%, right?