Edit: Debunking that SH is less useful with higher EV


If it doesn't fit anywhere else, it belongs here. Also, come here if you just need to get hammered.

Spider Stomper

Posts: 212

Joined: Monday, 3rd April 2017, 11:44

Post Tuesday, 31st July 2018, 16:52

Re: Edit: Debunking that SH is less useful with higher EV

No sieghurt, I understand his point that when defences are high, it might be better to train offence; there is no disagreement there as we all agree on this point, but that's just him moving the goalpost.

Unfortunately, that wasn't what he was arguing for in the first place:

Spoiler: show
VeryAngryFelid wrote:
Plantissue wrote:Why would higher EV result in SH become less useful?


I am lazy to run fsim so here is an easy example.
Let's compare 2 characters. One has 30% chance to dodge attacks, another has 70% chance to dodge attacks.
Now let's add a shield which blocks 20% attacks.
First character blocks 20% attacks and dodges 0.8*0.3=24% attacks so it blocks/dodges 44% attacks which is roughly 46% (44/30*100%-100%) improvement.
Second character blocks 20% attacks and dodges 0.8*0.7=56% attacks so it blocks/dodges 76% attacks which is roughly 9% (76/70*100%-100%) improvement.


He only argued what you have written after he argued to use absolute numbers and was proved wrong.
Then he ran some fsim and was proved wrong.
Then he changed the metric to defence to 1hp condition, to being invincible, to turns and all sorts of funny metrics.

Then he argued the above, which was not what he was arguing for in the first place, otherwise he wouldn't had written his nonsensical maths.
And now he is quibbling over the word "training", even though it matters not a single jot whether the EV or SH is obtained through training or equipment.
The results are the same. If you ever need to decide to improve defences, the usefulness of SH, is entirely independent of your EV value.

Your EV value doesn't matter, just as much as if you ever want to improve offence, it doesn't innately matter what your fighting skill is to the usefulness of your weapon skill, or fighting skill to the weapon choice.


(except the difference in exp needed for gains in skill lvl but lets ignore that or opportunity cost of equipment swappage but lets ignore that.)

It would also be irrelevant that you have a high offence and may choose to raise defence instead, to the usefulness of fighting skill if weapon skill was high or not.
Last edited by Plantissue on Tuesday, 31st July 2018, 17:05, edited 2 times in total.

Ziggurat Zagger

Posts: 4341

Joined: Friday, 8th May 2015, 17:51

Post Tuesday, 31st July 2018, 16:56

Re: Edit: Debunking that SH is less useful with higher EV

Plantissue

You keep claiming I talked about training. I said nothing about it initially. I guess English is your native language and you are trying to fix what is not broken (i.e. what I meant with my posts, "putting words in my mouth").

Edit. Oh, I have read another sentence from your post. "was proved wrong" lol. You really miss my points like I was missing Berder's points initially but at least I realize it and confessed that I am lazy/tired. Maybe I should read your post till the end, it is so funny ;)
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Tuesday, 31st July 2018, 18:33

Re: Edit: Debunking that SH is less useful with higher EV

Plantissue wrote:Btw, when dodge probability reaches 100%, time to die is infinity, and so a shield will increase the time dying to infinity x 1.25, which is still infinity, but a different type of infinity. Infinity is funny like that.


It's not; it is infinite in exactly the same way.
Ascension reports with too many words since 2016.

For this message the author damerell has received thanks: 2
duvessa, Sprucery

Spider Stomper

Posts: 212

Joined: Monday, 3rd April 2017, 11:44

Post Tuesday, 31st July 2018, 20:20

Re: Edit: Debunking that SH is less useful with higher EV

You VeryAngryFelid have shifted the goalpost several times and each time you have been shown wrong. Did you mention train? I can't be bothered to check right now, but hey, I'll be the better person than you and admit I may be wrong. Not that it matters, for whether the SH or EV is gained through training or equipment or god ability or potions or lvling up or whatever else that can affect SH and EV. The result is the same. The usefulness of SH is independent to the value of EV.

In set theory there are diferent types of infintity. not that it matters as you can never achieve 100% dodge rate. The usefulness of SH is still independent to the value of EV.

Will you two give it a rest already. For what should be something that could be explained in simple words, that the effectiveness of SH is independent of the value of EV, for the sake of that argument you have both made up totally bogus maths and just cannot admit it.

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Tuesday, 31st July 2018, 22:49

Re: Edit: Debunking that SH is less useful with higher EV

Plantissue wrote:In set theory there are diferent types of infintity.

There are, but they aren't involved here. While the use of the term "infinity" is usually quite dubious, if you are going to do it, the infinity you get by multiplying infinity by 1.25 is equal to the one you started with in every respect.
not that it matters as you can never achieve 100% dodge rate. The usefulness of SH is still independent to the value of EV.

As said elsethread, just because you can't doesn't mean you can't consider that hypothetical case, and your method for assessing the usefulness of SH says that it is just as useful to a character who dodges 100% of attacks. This is a clear deficiency in your method, because evidently SH is of no use whatsoever to such a character.

(I fear it is necessary to anticipate the quibble about reflection, or something; we're only discussing attacks where you don't care if you block or dodge).
you have both made up totally bogus maths and just cannot admit it.

All you have to do to make me admit it is produce one simple counterexample. Why can't you do it?
Ascension reports with too many words since 2016.

Ziggurat Zagger

Posts: 6181

Joined: Tuesday, 30th October 2012, 19:06

Post Wednesday, 1st August 2018, 00:53

Re: Edit: Debunking that SH is less useful with higher EV

damerell wrote:
not that it matters as you can never achieve 100% dodge rate. The usefulness of SH is still independent to the value of EV.

As said elsethread, just because you can't doesn't mean you can't consider that hypothetical case, and your method for assessing the usefulness of SH says that it is just as useful to a character who dodges 100% of attacks. This is a clear deficiency in your method, because evidently SH is of no use whatsoever to such a character.


As I said above, you're comparing two different things, what they are saying that 20% blocking from SH is always a 20% defensive bonus, if you have infinite evasion (hence 100% dodge rate) a 20% bonus on infinity is, yes, infinity, (Or to put it another way, if 0 damage is left after evasion, then a 20% bonus to 0 is still 0) that doesn't change the amount of bonus that a shield provides, it's the same, and a percentage bonus on 0 is yes, worthless in a practical sense.

However by the same practical sense, it's not possible to achieve infinite evasion, you can't compare a "practical case" to a purely theoretical one.

*In the realm of possible evasion scores* you will not evade every attack, and a given shield score will block a percentage of the ones that remain after your evasion applies, that percentage doesn't change no matter what your evasion actually is.

So the two positions here are "You get to a point where the amount of things not-evaded will not produce a meaningful result if you then block a subset of those" and "The percentage of things blocked after they are evaded produces the same net result as if you either evaded that percentage more things, or if you had that percentage more hps, or if you had that percentage more AC" which *are not mutually contradictory* but provide a different point of view of the same set of facts. It comes down to what you call a 'meaningful result'
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 02:25

Re: Edit: Debunking that SH is less useful with higher EV

Siegurt wrote:
damerell wrote:
not that it matters as you can never achieve 100% dodge rate. The usefulness of SH is still independent to the value of EV.

As said elsethread, just because you can't doesn't mean you can't consider that hypothetical case, and your method for assessing the usefulness of SH says that it is just as useful to a character who dodges 100% of attacks. This is a clear deficiency in your method, because evidently SH is of no use whatsoever to such a character.

As I said above, you're comparing two different things, what they are saying that 20% blocking from SH is always a 20% defensive bonus, if you have infinite evasion (hence 100% dodge rate) a 20% bonus on infinity is, yes, infinity, (Or to put it another way, if 0 damage is left after evasion, then a 20% bonus to 0 is still 0) that doesn't change the amount of bonus that a shield provides, it's the same, and a percentage bonus on 0 is yes, worthless in a practical sense.

Indeed. This was exactly my point in the previous thread; VAF and Plantissue both correctly measure an increase in defense, but one of them - VAF - measures it in a way which is more meaningful. It is precisely because the metric used here says that a character who can dodge every attack gets the same benefit from a shield as any other that we can see this is not a useful means for assessing the benefit.

The rest of your post seems to me to be reiterating my original point that these are two different metrics, which indeed they are, and in the previous thread I have already explained why I think VAF's metric is more useful (eg in the discussion with Berder), beyond the point reiterated here. I don't see much value in this thread since most of it consists of reiteration of already-accepted facts (if you dodge D and block B you will be hit by (1-D)(1-B), which no-one disputes), furious assertions that what I'm going to call the Plantissue metric is correctly calculated (which it is, it's just not as useful as it could be), and casting of aspersions on my head by Plantissue in spite of the fact that they are completely incapable of producing the one simple counterexample that would prove me wrong.
Ascension reports with too many words since 2016.

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 02:46

Re: Edit: Debunking that SH is less useful with higher EV

explained why I think VAF's metric is more useful

Useful for what? Can you give an example of a decision in-game that you think would be made differently in one metric than the other? If all such in-game decisions are exactly the same, then neither is more useful. Except that one might be simpler (the one in which the benefit of one stat remains constant regardless of the other stat) but leave that aside for now. Give an example of an in-game decision that you think would be made differently under your metric.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 03:05

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:
explained why I think VAF's metric is more useful

Useful for what? Can you give an example of a decision in-game that you think would be made differently in one metric than the other?

Frankly, I think this is a bit of a definitions attack. When I say VAF's metric is more useful, I mean that it more closely represents the actual utility of a shield in-game than a metric that thinks that a character who could dodge every attack would find a shield just as useful as one in treeform. I hope we already understand that translating this sort of mechanics arithmetic into actual gameplay decisions is tricky to do - I certainly expect you understand it, inasmuch as you must know that those neat graphs of relative benefit of AC and EV need a bit of careful interpretation (what kind of opponents do I face, how much do I care about variance in damage taken, etc) and yet I imagine you would suppose, as I do, that they are useful.

That said, specifically I think VAF's metric is more useful in that it represents the idea that blocking X% of incoming damage is more useful the more incoming damage there is, an idea we've discussed elsethread. As such, it might [1] well suggest taking a shield on a character with a poor EV who would (by VAF's metric which I think is more useful) receive more benefit, but not doing so and perhaps hence skilling elsewhere on a character with good EV.

Conversely the Plantissue metric would suggest someone who gets hit all the time can do no better out of a shield than someone who - if we exclude all hypotheticals - dodges as many attacks as is reasonably possible. That doesn't sound to me like so useful a basis for decisions in-game; the first character is surely crying out for the extra defence a shield can provide, where the second has plenty of consumables to use in the rare case something lands a bad hit.

[1] Frankly, except inasmuch as VAF's central point that blocking a fixed percentage of attacks is worth more if you evade less, or have worse AC, already appeared obvious to me.
Ascension reports with too many words since 2016.

For this message the author damerell has received thanks:
VeryAngryFelid

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 03:11

Re: Edit: Debunking that SH is less useful with higher EV

damerell wrote:Frankly, I think this is a bit of a definitions attack. When I say VAF's metric is more useful, I mean that it more closely represents the actual utility of a shield in-game than a metric that thinks that a character who could dodge every attack would find a shield just as useful as one in treeform. I hope we already understand that translating this sort of mechanics arithmetic into actual gameplay decisions is tricky to do - I certainly expect you understand it, inasmuch as you must know that those neat graphs of relative benefit of AC and EV need a bit of careful interpretation (what kind of opponents do I face, how much do I care about variance in damage taken, etc) and yet I imagine you would suppose, as I do, that they are useful.

That said, specifically I think VAF's metric is more useful in that it represents the idea that blocking X% of incoming damage is more useful the more incoming damage there is, an idea we've discussed elsethread. As such, it might [1] well suggest taking a shield on a character with a poor EV who would (by VAF's metric which I think is more useful) receive more benefit, but not doing so and perhaps hence skilling elsewhere on a character with good EV.

Conversely the Plantissue metric would suggest someone who gets hit all the time can do no better out of a shield than someone who - if we exclude all hypotheticals - dodges as many attacks as is reasonably possible. That doesn't sound to me like so useful a basis for decisions in-game; the first character is surely crying out for the extra defence a shield can provide, where the second has plenty of consumables to use in the rare case something lands a bad hit.

[1] Frankly, except inasmuch as VAF's central point that blocking a fixed percentage of attacks is worth more if you evade less, or have worse AC, already appeared obvious to me.

So you can't give an example of a decision in-game that you think would be made differently in one metric than the other. To me, "useful" is something that helps me make good decisions to increase my chances of winning. If your metric is not useful in that way, I would not call it useful.

Percentage increase in damage dealt or time-to-live is relatively easy to use for equipment and skill decisions.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 03:23

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:So you can't give an example of a decision in-game that you think would be made differently in one metric than the other.

I think I did just suggest an example of such a decision.
Percentage increase in damage dealt or time-to-live is relatively easy to use for equipment and skill decisions.

Being easy to use doesn't make it correct. As discussed, the character with on average 1 turn to live who extends that to 2 turns receives the "same" benefit as the one with 1000 turns to live who extends it to 2000; yet in practice we can see the first character, who buys a vital extra turn to use a consumable, receives a far greater benefit than the second, who just gets a sore finger from leaning on Tab longer.

A still easier to use metric is "the character with the most HP is best". It certainly satisfies your criteria, being "easy to use for equipment and skill decisions" - train nothing but Fighting, and select equipment at random unless Vitality turns up. Furthermore one can easily construct an analysis showing that if other factors are constant, it is better to have more HP.

Yet of course you don't make decisions on that basis. It is not enough that a metric be easy to use; you also demand that it provide useful insights.
Ascension reports with too many words since 2016.

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 03:31

Re: Edit: Debunking that SH is less useful with higher EV

[1] well suggest taking a shield on a character with a poor EV who would (by VAF's metric which I think is more useful) receive more benefit, but not doing so and perhaps hence skilling elsewhere on a character with good EV.

Ah, yes. I missed that you said this. Now give an actual example with fsim results for a realistic character and justifications. Explain what your method tells you to do, and why your method tells you to do that. Also explain why you think the percentage increase method would tell you to do the opposite choice. Then show fsim for the two choices and use the result to support your claim.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 03:42

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:
[1] well suggest taking a shield on a character with a poor EV who would (by VAF's metric which I think is more useful) receive more benefit, but not doing so and perhaps hence skilling elsewhere on a character with good EV.

Ah, yes. I missed that you said this. Now give an actual example with fsim results for a realistic character and justifications. Explain what your method tells you to do, and why your method tells you to do that. Also explain why you think the percentage increase method would tell you to do the opposite choice. Then show fsim for the two choices and use the result to support your claim.


Sha'n't.

I certainly don't need to do that for my original purpose; to prove that VAF, while they calculated the odds of being hit in an overly complex fashion, calculated them correctly, and used a method which produces a correct result under all circumstances, and as such was subject to an entirely unjustified attack from Plantissue who had simply failed to comprehend it.

I also don't need to do that to argue that a metric that shows a character who can dodge every attack benefits from a shield is bogus. It is obvious that such a metric is bogus, because a shield would give no benefit whatsoever then. You might as well demand I fsim to show that more regeneration gives no benefit to a character who never takes any damage, either.

I suppose again I must head off the "oh, just hypothetical" business. We don't have to talk hypothetical characters to see that - as mentioned above - a character who basically cannot dodge and so takes a lot of incoming damage cares more about blocking 20% with a shield than a character who dodges as much as is possible.

The demand is also grossly hypocritical, unless I have missed a thread where every point on your graphs is backed up by fsim showing that the decision to take AC over EV here or vice versa is so justified.
Ascension reports with too many words since 2016.

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 04:01

Re: Edit: Debunking that SH is less useful with higher EV

My ac/ev/sh graphs are another thing entirely, they show the logarithm base 1.1 of the percentage improvements. And yes, when developing those I frequently checked them against fsim for accuracy. But this thread is about the raw percentage improvements.

Here is the truth: your method doesn't actually tell you what to skill. It gives no basis to decide whether to skill fighting, weapon skill, shields, armor, or dodging, because it doesn't measure them on the same scale.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 04:21

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:My ac/ev/sh graphs are another thing entirely, they show the logarithm base 1.1 of the percentage improvements. And yes, when developing those I frequently checked them against fsim for accuracy. But this thread is about the raw percentage improvements.

Indeed. This thread (inasmuch as it's not simply rants) produces numbers, but with little analysis of their meaning. That doesn't mean the numbers are useless, but they do cry out for interpretation.
Here is the truth: your method doesn't actually tell you what to skill. It gives no basis to decide whether to skill fighting, weapon skill, shields, armor, or dodging, because it doesn't measure them on the same scale.

Quite correct, although it may give hints. As such, it may well be superior to a method that does tell you what to skill but gets it wrong, like people now who recommend that STR >= ER. I already gave an example of a method what does tell you perfectly what to skill (Fighting) and measures it all on the same scale (HP). A simple method is no use if reality is more complicated.
Ascension reports with too many words since 2016.

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 04:28

Re: Edit: Debunking that SH is less useful with higher EV

The percentage increase method gives good results in real games. It's useful, and the only reason it's not consulted more often is that it takes time to set up a character in wizmode and check fsim to see what the combat power actually is.

Your method gives no advice or predictions so is not useful.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 04:43

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:The percentage increase method gives good results in real games.

Does it? You are an expert player, you use it, and you win a lot; but where is the control player who is just as expert but doesn't use it? Since as you observe the extreme cases don't come up in normal play, a slight deficiency could easily never be observed.
Your method gives no advice or predictions so is not useful.

As discussed above, it does give some, inasmuch as it is "my method" at all, rather than that I once said that a method that says when you dodge every attack you should still have a shield is bogus. It simply doesn't present them as ironclad rules. That is best when, as illustrated, the ironclad rules would be wrong.
Ascension reports with too many words since 2016.

Tartarus Sorceror

Posts: 1774

Joined: Tuesday, 23rd December 2014, 23:39

Post Wednesday, 1st August 2018, 05:13

Re: Edit: Debunking that SH is less useful with higher EV

As discussed above, it does give some, inasmuch as it is "my method" at all, rather than that I once said that a method that says when you dodge every attack you should still have a shield is bogus

The percentage increase in time-to-live when you already dodge every attack, is (new time to live) / (old time to live) - 1 = infinity/infinity - 1 = undefined. The method of percentage increase gives no advice in this case. It's not the case that it says you should still have a shield.
streaks: 5 fifteen rune octopodes. 15 diverse chars. 13 random chars. 24 NaWn^gozag.
251 total wins Berder hyperborean + misc
83/108 recent wins (76%)
guides: safe tactics value of ac/ev/sh forum toxicity

For this message the author Berder has received thanks:
duvessa

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 06:07

Re: Edit: Debunking that SH is less useful with higher EV

Berder wrote:
As discussed above, it does give some, inasmuch as it is "my method" at all, rather than that I once said that a method that says when you dodge every attack you should still have a shield is bogus

The percentage increase in time-to-live when you already dodge every attack, is (new time to live) / (old time to live) - 1 = infinity/infinity - 1 = undefined. The method of percentage increase gives no advice in this case. It's not the case that it says you should still have a shield.

I dealt with this long ago; as the chance to dodge tends to 100%, the supposed value of a shield stays constant. This is a completely trivial objection; even if you don't accept the endpoint (in sharp contrast to Plantissue's casual use of "infinity"), it remains true when one dodges 99%, 99.9%, 99.99%, etc. The method predicts that a character who can expect to be hit once in an entire game gets the same value from adding a shield as one who cannot dodge a sloth.

Furthermore you must recognise that each time you post in this thread it is not to address the entire previous post but to desperately single out some particular point to object to.
Ascension reports with too many words since 2016.

For this message the author damerell has received thanks:
VeryAngryFelid

bel

Dis Charger

Posts: 2085

Joined: Tuesday, 3rd February 2015, 22:05

Post Wednesday, 1st August 2018, 06:28

Re: Edit: Debunking that SH is less useful with higher EV

damerell wrote:Furthermore you must recognise that each time you post in this thread it is not to address the entire previous post but to desperately single out some particular point to object to.

I have no interest in this discussion, but just because someone picks a point to reply to doesn't mean that they're trying to dismiss your argument. You don't simply quote the parts of the post which you agree with, just to keep saying "I agree", do you? That would be pointless, not to mention boring. People reply to those parts of the argument which they disagree with.

At the end of the day, this entire thread is of dubious practical value. The best one can hope for is that one or two points get across.

For this message the author bel has received thanks: 2
duvessa, VeryAngryFelid

Ziggurat Zagger

Posts: 4341

Joined: Friday, 8th May 2015, 17:51

Post Wednesday, 1st August 2018, 06:31

Re: Edit: Debunking that SH is less useful with higher EV

Ok, I think I found something everyone can agree with so the thread won't be full of the same arguments mentioned again and again and finally will be closed. If we have a character with EV -1 and it equips a large shield with 0 levels in Shields skill, the shield is more useful to it than to a character with EV 10+ because the latter gets its EV decreased due to shield penalty :)
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Wednesday, 1st August 2018, 06:58

Re: Edit: Debunking that SH is less useful with higher EV

bel wrote:
damerell wrote:Furthermore you must recognise that each time you post in this thread it is not to address the entire previous post but to desperately single out some particular point to object to.

I have no interest in this discussion, but just because someone picks a point to reply to doesn't mean that they're trying to dismiss your argument.


Not in general, no. However, while I would be delighted to discover that Berder agrees with almost everything I have said in this discussion, I don't really get the impression that they do. For example, I don't think that their most recent post means they accept that this percentage-increase metric is obviously flawed and they merely need to consider the situation as the chance to dodge tends to 100% to avoid division by zero. However, since that obvious flaw was the main point of my preceding post, perhaps you are correct and they do accept it.
Ascension reports with too many words since 2016.

For this message the author damerell has received thanks:
VeryAngryFelid

Ziggurat Zagger

Posts: 6181

Joined: Tuesday, 30th October 2012, 19:06

Post Wednesday, 1st August 2018, 14:49

Re: Edit: Debunking that SH is less useful with higher EV

damerell wrote:
bel wrote:
damerell wrote:Furthermore you must recognise that each time you post in this thread it is not to address the entire previous post but to desperately single out some particular point to object to.

I have no interest in this discussion, but just because someone picks a point to reply to doesn't mean that they're trying to dismiss your argument.


Not in general, no. However, while I would be delighted to discover that Berder agrees with almost everything I have said in this discussion, I don't really get the impression that they do. For example, I don't think that their most recent post means they accept that this percentage-increase metric is obviously flawed and they merely need to consider the situation as the chance to dodge tends to 100% to avoid division by zero. However, since that obvious flaw was the main point of my preceding post, perhaps you are correct and they do accept it.

The problem with your theoretical "tends towards zero" situation is that it doesn't apply specifically towards shields, it applies to *all combat power* as you tend towards 100% dodging rate, you win *all combats* with no risk, at that point, not only is there no point in training shields, there's no point in getting more HP, or AC or EV or even offense either (unless you're optimizing for other things than win percentage)

Since Berder et al's point is about the *relative value* of SH when compared to HP, AC, EV and offense (in that a shield that provides a fixed amount of blocking power doesn't change it's bonus *relative to the other possible choices for increasing your combat power*) when the value of everything drops to 0, yes, the value of SH tends towards 0, but so does the value of *everything else* proportionally SH stays just as valuable for increasing your combat power as everything else.

The point about SH always being equally valuable is predicated on the fact that you actually need to improve something, because if you are already perfect, combat-wise, then you have no need to improve, so *nothing* has any value (from a win-percentage perspective, if you win all combats with no risk, then you can't possibly improve no matter what you do)

So the question is "If you want to improve a measure of your combat effectiveness, and reduce your chances of dying and thereby increase your chances of winning, what is most useful" Berder's measure is effective for this, yours isn't (or at least overly complicates the issue for no additional measure of accuracy)

If I evade 20%, 80% or 99% of all attacks, *and still need to reduce my chances of death* a shield that blocks 20% of attacks still improves my chances by 20%. Yes, in the 99% case, it's 20% of a much smaller number, but this entire conversation is predicated on the fact that in that case you wanted to improve your chances of not dying, and a 20% improvement is still a 20% improvement, even if you start with a very small chance of dying in the first place.
Spoiler: show
This high quality signature has been hidden for your protection. To unlock it's secret, send 3 easy payments of $9.99 to me, by way of your nearest theta band or ley line. Complete your transmission by midnight tonight for a special free gift!

For this message the author Siegurt has received thanks:
Berder

Spider Stomper

Posts: 245

Joined: Monday, 12th September 2016, 16:25

Post Thursday, 2nd August 2018, 05:43

Re: Edit: Debunking that SH is less useful with higher EV

Siegurt wrote:
damerell wrote:Not in general, no. However, while I would be delighted to discover that Berder agrees with almost everything I have said in this discussion, I don't really get the impression that they do. For example, I don't think that their most recent post means they accept that this percentage-increase metric is obviously flawed and they merely need to consider the situation as the chance to dodge tends to 100% to avoid division by zero. However, since that obvious flaw was the main point of my preceding post, perhaps you are correct and they do accept it.

The problem with your theoretical "tends towards zero" situation is that it doesn't apply specifically towards shields, it applies to *all combat power* as you tend towards 100% dodging rate, you win *all combats* with no risk, at that point, not only is there no point in training shields, there's no point in getting more HP, or AC or EV or even offense either (unless you're optimizing for other things than win percentage)

Well, yes, it is true that a character with very high defences already receives diminishing value from additional defences. I'm not sure why you think that's a "problem", though.
If I evade 20%, 80% or 99% of all attacks, *and still need to reduce my chances of death* a shield that blocks 20% of attacks still improves my chances by 20%.

Now you're not comparing like with like; "still need to reduce my chances of death" is the giveaway. The point is that that hypothetical character who avoids 99% of all attacks is extremely unlikely to need to reduce its chances of death. A character with a very high but still feasible dodge chance is much less likely to need to do so in any given situation.

(There's a platitude coming about how "ah, but it'll happen sometimes". It will, but a character who it almost never happens to has plenty of consumables to burn for those situations. A character to whom it happens all the time doesn't. The latter will benefit more from a shield stopping them happening.)

The sentence you've written there is a bit like saying "If I am dirt poor, or as rich as Bill Gates, but still desperately need money ..."; the "but still" clause serves to obviate the setup conditions. It's extremely unlikely that Gates will desperately need money.
Ascension reports with too many words since 2016.

Ziggurat Zagger

Posts: 4341

Joined: Friday, 8th May 2015, 17:51

Post Thursday, 2nd August 2018, 06:11

Re: Edit: Debunking that SH is less useful with higher EV

Training SH with extremely high EV is putting all your eggs in one basket. Even ignoring attacks which ignore both ev an sh, you want a defense for situations when you are unlucky (I've miscasted a 1% spell thrice in a row, it is 1 in million chance). So you end up training armour (especially great with gdr), fighting or death's door/cblink, even if math tells you otherwise. Why? Because when you have great defense vs average damage, you change priorities and try to increase your defense vs max damage because you just can't die to average damage unless you are sleeping on keyboard but you still can die to max damage.
Underestimated: cleaving, Deep Elf, Formicid, Vehumet, EV
Overestimated: AC, GDS
Twin account of Sandman25

Spider Stomper

Posts: 212

Joined: Monday, 3rd April 2017, 11:44

Post Friday, 3rd August 2018, 14:29

Re: Edit: Debunking that SH is less useful with higher EV

@Damerall it is correct to say that if the defences are high, you might as well decide to invest in offence, if the offence is low, and indeed we all agree on that. It is also entirely irrelevant, as if that character does decide to improve defences, to decide whether or not to improve SH or EV, the bonus to defence that SH provides is still entrely independent of EV, and vice versa.

@VeryAngryFelid, I beleive in maximising a broad sprectrum of defences are there are many different sort of attacks. Whether it is done by improving AC, or resistances or spell abilities, or even increasing HP, they are all still independent of each other to their effectiveness in terms of defences.



For example, if I want to remove EV and SH altogether from our discussion that you are so hung up on, if you want to improve the chance of not dying to torment (yes I know you can't die to torment alone, so this is a kind of a bad example, but good enough), by reducing the amount of HP lost, then there are two main defences (ignoring god abilties and scrolls and potions and spirit shield), which is resistances and HP.

No matter how much HP you have, or don't have is completely independent of how useful rN+ is. (Actually in theory since you can never die to torment it's a bit more complicated than that for for large numbers of HP it is not irrelevant to the argument at hand and probably not the best example because after rN+++, you gain no benefit). Conversely if you already have rN+ or rN++ or rN+++, increasing your HP is just as useful to survive torments.

It would not be a situation where you can argue that because HP life count is large, getting more rN+ would innately be less useful than if HP life count was smaller.

So it is with HP and rN+, so it is with AC and EV.

No maths, just words, just in case you are lazy/tired/unfun/leaving/another last post/not so lazy/not so tired/ whatever.
Previous

Return to Crazy Yiuf's Corner

Who is online

Users browsing this forum: No registered users and 7 guests

Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group.
Designed by ST Software for PTF.