Some time after Armor Up: where do we stand?

[BigErkyKid]

Ranking toughness with Rogue Blitz Build as 1.00. A build with 2.00 can go twice as long without needing a heal. These results do not include self healing.

1.00 Rogue Blitz MaxDex
1.07 Rogue Shadowdancer MaxDex
1.39 Barbarian
1.52 TWF Ranger Medium Armor + Dodge Build
1.67 THF Pally Heavy Armor
1.70 THF Pally Medium Armor
2.38 S&B Pally

I would have a hard time believing that low PRR high dodge builds are somehow equal to medium, heavy, or S&B builds. Non-Dex rogue builds will do even worse.

When you factor in that low PRR means getting one shot or two shot much easier, low PRR high dodge builds really lose out.

These results would need to be modified by each builds ability to self heal for a true measure of defense. Again, the rogue loses out here.

I want to make it clear that I don't think rogue needs a buff (they have stealth), but simply that high dodge builds are not equal to tankier builds, as suggested in the OP.

First of all, great work up there. Clear, for the most part, and useful exercise.

I have two comments:

1) Dodge builds in light armor do end up having very similar defensive abilities as typically tankier builds (paladin and barbarian). According to your own calculations:

1.39 Barbarian
1.52 TWF Ranger LIGHT Armor + Dodge Build
1.67 THF Pally Heavy Armor
1.70 THF Pally Medium Armor

Once you bring this numbers back to hits to kill, the differences are virtually negligible.

2) A point of contention is the rogue build. You give it 165 PRR out of the blue, and that conditions a lot the results.
For the sake of consistency, I used the same feats, PLs, and reasonable enhancements across all builds. I see no reason to make an exception for the rogue. Once one does that, we have a rogue with 195 PRR. Let's do the computation for that rogue:

Now let's have some quick numbers on PRR achievable by rogue, using the same configuration as for the other builds.
TWF rogue 165 PRR (38% taken) / (34% blitz)
30 Larmor
6 Enhancements
38 Celestial armor
25 Insight+quality item
10 Epic DR
20 Scion of earth
36 PLs
(30) Blitz
30+6+38+25+10+20+36=165 (195)
Damage taken: (100/(100+195))=0.3389831

We now compute the chance to be hit, which is the same as yours:

18.92%

And redo the calculations. Expected damage taken is:

(0.1892* 0.3389831)*1000=64.1356

Leading to the following hits to kill:
1000/64.1356=15.59196

Now obviously the question that begs is why does the rogue have 1000Hps and the ranger 1300. The ranger has access to 60 HPS via enhancements and 2 hit points more per level. Thats a total of 60 + 60=120. Meaning that if it is reasonable for a ranger to get to 1300, it is reasonable for a rogue to go 1300-120=1180.

If we plug in this new value into the calculations, we get:

1180/64.1356=18.39852.

So my ranking would be:

1.00 Rogue blitz
1.05 barbarian
1.15 TWF ranger
1.26 Paladin heavy armor
etc.

Even in this scenario the rogue is behind, but now by so much. The reason why the inclusion of AC leaves most builds unaltered (THF/TWF) but the rogue is because the main advantage of the rogue was avoidance via not being hit. The problem of that kind of avoidance is that once you start packing multiple stacking sources of avoidance each individual roll loses importance.

Example:


Build 1a: dodge 25 / displacement 50: to be missed (avoid) =.75*.5=0.375
Build 1b: dodge 25 : avoid=.25
Build 2a: dodge 10 / displacement 50: avoid=0.45
Build 2b: dodge 10: avoid=.1

So without displacement, the difference in the chances to be hit between build 1 and 2 are 15%, but adding displacement halves this difference to 7.5%.

Here is my conclusion:

In order to create different play styles, the devs should have allowed only 1 displacement clicky (or given timers, or limit displacement in other ways). By doing so, dodge would create big differences in the way builds play. Once displacement was limited, they should have NOT given so many sources of stacking PRR to be able to keep the differences between heavily armored builds and the rest. This would split the builds cleanly in the avoidance category VS the mitigation, effectively creating different styles.

[nokowi]

First of all, great work up there. Clear, for the most part, and useful exercise.

I have two comments:

1) Dodge builds in light armor do end up having very similar defensive abilities as typically tankier builds (paladin and barbarian). According to your own calculations:



Once you bring this numbers back to hits to kill, the differences are virtually negligible.

2) A point of contention is the rogue build. You give it 165 PRR out of the blue, and that conditions a lot the results.
For the sake of consistency, I used the same feats, PLs, and reasonable enhancements across all builds. I see no reason to make an exception for the rogue. Once one does that, we have a rogue with 195 PRR. Let's do the computation for that rogue:



We now compute the chance to be hit, which is the same as yours:



And redo the calculations. Expected damage taken is:

(0.1892* 0.3389831)*1000=64.1356

Leading to the following hits to kill:
1000/64.1356=15.59196

Now obviously the question that begs is why does the rogue have 1000Hps and the ranger 1300. The ranger has access to 60 HPS via enhancements and 2 hit points more per level. Thats a total of 60 + 60=120. Meaning that if it is reasonable for a ranger to get to 1300, it is reasonable for a rogue to go 1300-120=1180.

If we plug in this new value into the calculations, we get:

1180/64.1356=18.39852.

So my ranking would be:



Even in this scenario the rogue is behind, but now by so much. The reason why the inclusion of AC leaves most builds unaltered (THF/TWF) but the rogue is because the main advantage of the rogue was avoidance via not being hit. The problem of that kind of avoidance is that once you start packing multiple stacking sources of avoidance each individual roll loses importance.



Here is my conclusion:

In order to create different play styles, the devs should have allowed only 1 displacement clicky (or given timers, or limit displacement in other ways). By doing so, dodge would create big differences in the way builds play. Once displacement was limited, they should have NOT given so many sources of stacking PRR to be able to keep the differences between heavily armored builds and the rest. This would split the builds cleanly in the avoidance category VS the mitigation, effectively creating different styles.

Displacement makes no difference in the toughness ratios with my calculations. Here is why:

Calculating the chance to hit we multiply the (1- AC miss chance) by (1-dodge) by (1-displacement) by (1-incorporeal).

All builds get the same multiplier for to hit chance, so the displacement values take 1/2 the expected damage and have twice as many hits to kill. In other words, the math I did displacement doesn't affect the ratios. I don't have time to recheck my work right now, but you can comment if this is incorrect.

Changing rogue PRR from 165 to 195 to bumps them up to 14.01 to 15.59, which is still way behind.

If you are going to build an absolute maximum rogue PRR and AC build, you need to do the same for all other builds. That means bumping up the AC and PRR for all other builds from the "reasonable average" values you suggested.

The result might get worse for rogue, not better as you suggest.

Rouge has no other class sources of AC or PRR, while ranger and pally do. This is a DEX rogue, and all other rogues do much worse.

Should I waste my time maxing out the other builds AC and PRR, or can you concede the point that you are comparing the max rogue build to average other builds?

[nokowi]

The reason why the inclusion of AC leaves most builds unaltered (THF/TWF) but the rogue is because the main advantage of the rogue was avoidance via not being hit. The problem of that kind of avoidance is that once you start packing multiple stacking sources of avoidance each individual roll loses importance.

Paladin S&B had a higher avoidance than rogue. I don't think you can make this conclusion.

Lower PRR and HP is of HUGE importance if you have ever played a rogue. Even if HTK was equal with avoidance (it's not), rogue would be way behind due to one/two shot potential and lack of healing.

[BigErkyKid]

Displacement makes no difference in the toughness ratios with my calculations. Here is why:

Calculating the chance to hit we multiply the (1- AC miss chance) by (1-dodge) by (1-displacement) by (1-incorporeal).

All builds get the same multiplier for to hit chance, so the displacement values take 1/2 the expected damage and have twice as many hits to kill. In other words, the math I did displacement doesn't affect the ratios. I don't have time to recheck my work right now, but you can comment if this is incorrect.

This is correct and incorrect at the same time. It is correct that it simply scales down all the numbers equally, but it is incorrect in the sense that the play style becomes different. Here is why:

Against a boss (no displacement), the chances to be hit of the ranger are 4% better than the paladin. This is halved against trash. In expectation, they take the same ratio of number of hits to die wrt to each other, but when displacement is not active, this creates differences in how it plays out in practice. To give you an idea, the dodge character can have more variance on the distributions of hits to die. Sometimes they are lucky (don't get hit), sometimes unlucky (many hits in a row), and this is less so for the other toon which focuses on mitigation. The more stacking avoidance options they give us, the less that there is a distinction between DODGE based and PRR based toons.

Changing rogue PRR from 165 to 195 to bumps them up to 14.01 to 15.59, which is still way behind.

...

Should I waste my time maxing out the other builds AC and PRR, or can you concede the point that you are comparing the max rogue build to average other builds?

I have not attempted any max PRR, or average. I have given common feat choices to all the builds (this would include the rogue in my calculations). They all play by the same rules, except for the class based enhancements and the armor type. Why? Because the point is to compare builds where the only differences are class an armor based, not feat, destiny, and gear wise. The claim is that ARMOR ends up mattering little once everything is factored in, as builds look very similar anyway. This is what I called homogenization.

Your point was, if I understand correctly, that AC would break this equivalence. I do not see very substantial differences, with the exception of a shield build. Not terrible interested in shield based, as those are not based on the armor, rather on the shield.

Finally, I am not arguing that rogue is better defensively overall. Or ranger, or what not. Here I am strictly focusing on armor types, and my claim is that they have become roughly speaking similar. The differences in playing a TWF LIGHT ARMOR ranger or a THF HEAVY ARMOR paladin, two a prior very different archetypes, are quite minimal. You can check your own numbers. Obviously the rogue is the squishy of the group and no one in their right mind would claim that they are as sturdy as a paladin. This is just simply not the aim of what I was trying do test here.

[BigErkyKid]

Paladin S&B had a higher avoidance than rogue. I don't think you can make this conclusion.

Lower PRR and HP is of HUGE importance if you have ever played a rogue. Even if HTK was equal with avoidance (it's not), rogue would be way behind due to one/two shot potential and lack of healing.

And to be clear, my prior statement in itself was NOT correct. I just wanted to clarify (see my post above this) exactly what I meant regarding homogenization and stacking avoidance eroding that.

[nokowi]

This is correct and incorrect at the same time. It is correct that it simply scales down all the numbers equally, but it is incorrect in the sense that the play style becomes different. Here is why:

Against a boss (no displacement), the chances to be hit of the ranger are 4% better than the paladin. This is halved against trash. In expectation, they take the same ratio of number of hits to die wrt to each other, but when displacement is not active, this creates differences in how it plays out in practice. To give you an idea, the dodge character can have more variance on the distributions of hits to die. Sometimes they are lucky (don't get hit), sometimes unlucky (many hits in a row), and this is less so for the other toon which focuses on mitigation. The more stacking avoidance options they give us, the less that there is a distinction between DODGE based and PRR based toons.

It is much different to occasionally take large spikes of damage or consistently take moderate amounts of damage. When you have a predictable amount of damage taken (consistent but moderate), you can build past this through toughness or self healing. When damage spikes are larger but less frequent, it is more difficult to stay alive because you would need to build past the large damage spikes. They are not equivalent until dodge toons beat mitigation toons by some <undefined> amount.

I have not attempted any max PRR, or average. I have given common feat choices to all the builds (this would include the rogue in my calculations). They all play by the same rules, except for the class based enhancements and the armor type. Why? Because the point is to compare builds where the only differences are class an armor based, not feat, destiny, and gear wise. The claim is that ARMOR ends up mattering little once everything is factored in, as builds look very similar anyway. This is what I called homogenization.

Your point was, if I understand correctly, that AC would break this equivalence. I do not see very substantial differences, with the exception of a shield build. Not terrible interested in shield based, as those are not based on the armor, rather on the shield.

I can easily add more AC to most of these non rogue builds, but not to rogue. I only looked in one class tree for non rogues, while I included everything that rogues can have from all 3 trees. I believe the 165 rogue PRR example is representative of what we will get when I start adding more AC to all other builds (barbarian may be the exception), so I don't see the value of you adding more rogue PRR unless you want me to do more work and add in all sources of AC for the non-rogues, just to get back to similar ratios as I originally posted.

Finally, I am not arguing that rogue is better defensively overall. Or ranger, or what not. Here I am strictly focusing on armor types, and my claim is that they have become roughly speaking similar. The differences in playing a TWF LIGHT ARMOR ranger or a THF HEAVY ARMOR paladin, two a prior very different archetypes, are quite minimal. You can check your own numbers. Obviously the rogue is the squishy of the group and no one in their right mind would claim that they are as sturdy as a paladin. This is just simply not the aim of what I was trying do test here.

This is consistent with D&D (AC with dex bonus is almost identical for all builds), and it makes sense for dev's to give players more build options by allowing high dodge + light armor builds and low dodge + heavy armor builds. This is good design (more player options) and I'm not sure why you would complain about it. In DDO, defense is not equivalent across all builds (I would agree with TWF ranger it is), so in some cases you are correct that players have a reasonable choice of light or heavy armor. Good luck making a rogue heavy armor build, or rogue with shield build. You could do it but they will pale in comparison to a TWF rogue build.

Tankier toons benefit from high MRR regardless of their armor choice, and thus beat out a non-evasion toon. Evasion toons take less magical damage, but suffer from larger damage spikes.

If something like Dex-to-Damage gives the TWF ranger evasion + similar defense, then that is a result of allowing Dex to damage in the first place. That would be the problem, and not armor values.

[nokowi]

To be fair, I went and looked at found 6 more AC for ranger, 10 AC for S&B pally, and 6 for THF pally in additional trees. Barbarian PRR was also fixed in this post. ***

ORIGINAL TOUGHNESS VALUES
2.38 S&B Pally
1.31 Barbarian
1.00 Rogue Blitz
1.07 Rogue Shadow
1.67 THF Heavy Pally
1.70 THF Medium Pally
1.52 TWF Ranger

MODIFIED RELATIVE TOUGHNESS VALUES *** with extra 30 PRR for Rogue and extra AC for ranger (+6 AC), pally S&B (+10 AC) and pally THF (+6 AC).
2.23 S&B Pally
1.17 Barbarian
1.00 Rogue Blitz
1.09 Rogue Shadow
1.55 THF Heavy Pally
1.58 THF Medium Pally
1.45 TWF Ranger

*** Relative Toughness value doesn't change with concealment (blur, displacement, true seeing)

I am happy with using the modified values for discussion.

If I subtract 30 PRR from all choices, the non-rogues gain about 0.02 TOUGHNESS. This tells me that it wasn't particularly important to have all builds choose (20 PRR at level 30 and the 10 PRR epic feat). (It was OK for BigErky to do this)

A TOUGHNESS value of 1.50 indicates that the build can survive 1.5x longer than a build with a TOUGHNESS of 1.00 before they need healing.

TOUGHNESS does not include the effect of one/two shots, so it is not the appropriate measure of defense for fighting a mob capable of killing in one or two shots. It does speak to the overall toughness of the build without considering self healing or getting one/two shot.

The TWF Ranger Light Armor Dodge build relies on dodge, and has less HP than pally. They will be one shot more often, and have to deal with larger damage spikes when their dodge chance fails them. I personally would consider this build less defensive than the Pally builds, but we can all interpret the numbers to our own liking.

THF Pally is ~10% tougher than TWF Ranger (even without considering damage spikes or one-shots), which falls into the range most people will notice (a 5-10% difference is required to be noticed by most people).


VALUES USED

THF Paladin heavy armor:
1500 HP / 244 PRR / 174 AC / 11 Dodge / 50% displacement / 10% incorporeal

THF Paladin med armor:
1500 HP / 229 PRR / 169 AC / 19 Dodge / 50% displacement / 10% incorporeal

TWF ranger in Light Armor
1300 HP / 228 PRR / 160 AC / 23 Dodge / 50% displacement / 15% incorporeal

THF barbarian
1700 HP / 209 PRR / 131 AC / 10 Dodge / 50% displacement / 10% incorporeal

S&B Paladin build:
1500 HP / 289 PRR / 234 AC / 6 Dodge / 50% displacement / 10% incorporeal

Rogue MaxDex build Blitz:
1000 HP / 195 PRR / 150 AC / 32 Dodge / 50% displacement / 10% incorporeal

Rogue MaxDex build Shadowdancer:
1000 HP / 165 PRR / 156 AC / 32 Dodge / 50% displacement / 25% incorporeal


*** EDITED to fix barbarian PRR to be 209 (OP on page 1 has a typo of 228 PRR)
*** EDITED to fix TWF Ranger to be in light armor.

[Baktiotha]

<snip>

It took me a moment to find the post where the addition error occurred. It took me a bit longer to see if there is anything of particular importance to the analysis being done resulting from the error.

I noticed that the only post referenced has all of the math correct. And I noticed in the post with the addition error that the conclusion will remain the same concerning the two builds being compared.

Nokowi offered a criticism of an algorithm I was using and suggested that I had not correctly accounted for the effect of mitigation on damage reduction. And, the post that did NOT contain the math error demonstrates conclusively that I did in fact account for the effect of mitigation on damage reduction and the criticism, if it were employed, would create an error.

To me that is somewhat different in scope than an addition error (which was then carried forward through the algorithm showing correct process even if the answer is incorrect). It is even less significant when anyone repeating the math can catch the error, correct it, and still see that the difference between the final result and the value derived in the comparison case remain very significantly different from one another.

At issue is not the performance of the maths but the construction of the algorithms. The algorithm suggested by nokowi failed. So, I'll continue to use the algorithm and process that I am using and I will take the rather weak criticism of an inconsequential addition error in one post as a reminder to double check my math for the remainder of the thread.

[Baktiotha]

"They can be regarded as functionally equivalent if their EHP is within 5% of eachother" is not the same thing as "they can be regarded as functionally equivalent if they take 1000 damage". If you select D in increments you will at best approximate the information you get from EHP.

This is a very interesting assertion and I'd like to address it specifically.

Axeyu's complaint is that the bands are arbitrary. But that complaint is completely without merit because the 5% value is arbitrary as well. There is no reason to choose 5%. We could choose 1% or 9% or 300%. We can set a % value where we want and doing so is just as arbitrary as defining a band as being 1000 wide or 500 wide or 10 wide.

So both Axeyu and I are suggesting arbitrary values. The difference is that the arbitrary value I use creates bands that are of equal size.

Using an arbitrary % creates bands of unequal sizes. The 5 samples Axeyu provided would create bands of 300, 426.4, 516, 559.3, and 583.5. And, as we move higher in EHP values more builds will be considered functionally equivalent to one another while, when we move lower in EHP values, fewer builds will be identified as functionally equivalent.

There is nothing wrong with using EHP as a center point for a band. There may be some utility in that. But because the bands are, by definition, arbitrary we should build them in a way so that all the bands are the same size. We can do that by choosing a fixed value.

I effectively do that with the 1000 point band I used. A 1000 point band can be viewed as setting the equivalency as EHP plus/minus 500. We might judge that too high, too low -- the choice depending on the size of the sample we have to consider.

Axeyu provided a sample of 5 builds producing EHP values of 5835, 5593, 5160, 4264, 3000. Using +/- 500 causes interesting results.

The first sample is equivalent to the second. The second is equivalent to the first and the third. The third is equivalent to the second. The fourth and fifth samples have no equivalents.

So, what kinds of conclusions does that let us draw about functional equivalency for the 5 samples? It depends on what argument I want to make doesn't it.

If I want to argue for homogenization I focus on 5593 and how 60% of the samples are functionally equivalent. If I want to discredit homogenization I focus on 5835 and note how only 40% of the samples are functionally equivalent (I might even argue that a band that is a third the size of the smallest EHP is too large and provides too much opportunity to include builds that should not be thought of as equivalent ).

Regardless, choosing an arbitrary band that retains a consistent size provides a picture that is easy to understand and to explain. I think that of the two arbitrary methods choosing a fixed value is better than choosing a %.

[Baktiotha]

Umm, when I went to school a quarter of 40 was 10 not 7.7!

It was but you are misreading the math. The 25% does not apply to the 40% miss chance. You need to struggle through the original post and follow all of the math being represented as it isn't necessarily sequential and the data points and process doesn't necessarily flow intuitively.

But, the 7.7% value is correct once you find all the maths and work the implied calculations.

[Baktiotha]

Hits to Kill Boss = 1500/128.55 = 11.67
Hits to Kill Blur = 1500/96.41 = 15.56
Hits to Kill Disp = 1500/64.28 = 23.34

The problem with this analysis is that you are not talking about hits to kill when you are discussing blur or displacement. Those things (along with AC and other causes of misses) alter the number of attacks that resolve as hits. They are in fact misses. They have zero impact on damage.

It is this imprecision in the discussion that is disturbing and misleading. It is the root cause of the criticism leveled against some of my earlier postings.

If we want to be clear in our conversation we cannot refer to attacks as hits. Attacks create an opportunity to hit. Whether that opportunity resolves as a hit or a miss is dependent on the factors causing avoidance.

Those attacks that are avoided are resolved as a miss. Those attacks that are not avoided are resolved as a hit. Each hit does damage and that damage is mitigated (in the case of physical attacks) by PRR and DR.

Hits to kill is a simple relationship between hit points and damage and is expressed by HTK=HP/((D-DR)*M).

Attacks to hit is a simple relationship among all of the avoidance factors. Attacks to kill would be a relationship between attacks to hit and hits to kill.

This may be what nokowi is representing with the erroneously labeled Hits to Kill Blur and Hits to Kill Disp(lacement). I don't know because I haven't worked the math to see.

At this moment the math isn't important -- the definition of terms is.

The idea presented that a character can sustain more hits displaced than blurred is false. A character can only sustain the number of hits that HP allow and the strength of a hit is not altered by blur or displacement.

[Baktiotha]

Ranking toughness with Rogue Blitz Build as 1.00. A build with 2.00 can go twice as long without needing a heal. These results do not include self healing.

1.00 Rogue Blitz MaxDex
1.07 Rogue Shadowdancer MaxDex
1.39 Barbarian
1.52 TWF Ranger Medium Armor + Dodge Build
1.67 THF Pally Heavy Armor
1.70 THF Pally Medium Armor
2.38 S&B Pally

It is amusing to see an analysis that is where we were already at when I had to attend to RL.

Once people understand that the "toughness" of a build is not a measure of hits to kill but instead is a measure of how frequently a build is hit they will grasp that this is exactly where I left off in the step-by-step character analysis. All that is missing is assigning a length of time to a 1.00 toughness so we know how much time a character has available in order to obtain the needed heal.

And, I find it even more amusing that after stressing the need to model with self heals the wording in the above post is "without needing a heal." I think it is time to recognize that it isn't important where the heal comes from, just that it is needed.

[Axeyu]

This is a very interesting assertion and I'd like to address it specifically.

Axeyu's complaint is that the bands are arbitrary. But that complaint is completely without merit because the 5% value is arbitrary as well. There is no reason to choose 5%. We could choose 1% or 9% or 300%. We can set a % value where we want and doing so is just as arbitrary as defining a band as being 1000 wide or 500 wide or 10 wide.

So both Axeyu and I are suggesting arbitrary values. The difference is that the arbitrary value I use creates bands that are of equal size.

Using an arbitrary % creates bands of unequal sizes. The 5 samples Axeyu provided would create bands of 300, 426.4, 516, 559.3, and 583.5. And, as we move higher in EHP values more builds will be considered functionally equivalent to one another while, when we move lower in EHP values, fewer builds will be identified as functionally equivalent.

You would do yourself a favour if you just forgot about your misguided notion of "bands", it really is not not a helpful approach. There is nothing complicated what-so-ever with "if two values are similar they will have similar impact".

I effectively do that with the 1000 point band I used. A 1000 point band can be viewed as setting the equivalency as EHP plus/minus 500. We might judge that too high, too low -- the choice depending on the size of the sample we have to consider.

You are forgetting that every EHP has some effect on survivability. Your method would only make sense if mobs only did damage in multiples of 500, whatever specific damage number you want to assing. In all cases the results will never do more than approximate a straight up EHP comparison.

Axeyu provided a sample of 5 builds producing EHP values of 5835, 5593, 5160, 4264, 3000. Using +/- 500 causes interesting results.

The first sample is equivalent to the second. The second is equivalent to the first and the third. The third is equivalent to the second. The fourth and fifth samples have no equivalents.

So, what kinds of conclusions does that let us draw about functional equivalency for the 5 samples? It depends on what argument I want to make doesn't it.

If I want to argue for homogenization I focus on 5593 and how 60% of the samples are functionally equivalent. If I want to discredit homogenization I focus on 5835 and note how only 40% of the samples are functionally equivalent (I might even argue that a band that is a third the size of the smallest EHP is too large and provides too much opportunity to include builds that should not be thought of as equivalent ).

Regardless, choosing an arbitrary band that retains a consistent size provides a picture that is easy to understand and to explain. I think that of the two arbitrary methods choosing a fixed value is better than choosing a %.

No. Your specific damage numbers are not just arbitrary, they are meaningless. In your post you even show how you must tailor them to get an interesting situation. Yet you claim that your method is better? Thats ridicoulous.
When will it sink in that using specific damage numbers does not do anything but approximates the EHP differences?
Every time you have proclaimed victory so far you have been completely wrong about everything, so is it not time for you to just accept that this topic is beyond you? No offence, ofcourse.
I asked you to do some examples, you clearly have not. So in your inevitable reply please do some examples with groups of varying EHP levels and evaluate the method of assigning a specific damage band versus the method of "similar values give similar results".

[Baktiotha]

This is correct and incorrect at the same time. It is correct that it simply scales down all the numbers equally, but it is incorrect in the sense that the play style becomes different.

This discussion is very much off base.

AC, dodge, concealment and incorporeal misses are all forms of avoidance. Every character class benefits from having the highest possible values in each of these areas.

Armor type is the biggest factor in determining AC. Armor type is also the biggest factor in limiting dodge. Concealment and incorporeal have exactly the same impact on every build presuming they are the same. In the case of incorporeal that is likely the case. In the case of concealment it is inaccurate to assume sameness. This is because some builds have inherent concealment and others must rely on equipment. It is misleading to discuss miss chances without making it explicitly clear where the concealment comes from -- because then you are not talking about "every" character but rather about "a" character.

When talking "a" character there can be similarities because of the unique qualities of that character. When talking about "every" character those similarities may be reduced or disappear.

When you limit the discussion only to AC and dodge -- the only 2 elements that will certainly be the same for all classes -- there is a great deal of difference between armor types and the avoidance that is provided.

And, because the armor types also provide PRR the impact of hits varies.

A rogue with 32% dodge and 150 AC from light armor is dependent on avoidance to a huge degree. A paladin with 6% dodge and 234 AC from heavy armor and shield is dependent on mitigation to a huge degree. So they do play "differently" in the sense that their response to a threat depends on different factors (or, the same factors but to different degrees).

Where we see two different characters being played the same isn't a consequence of the builds being too similar -- it is the result of players not carrying and not choosing to use different methods. Brute force wins most encounters in DDO. There is no incentive to choose to play differently. So, play style isn't dictated by the builds. Play style is dictated by players taking the path of least resistance.

I'd blame that on Turbine except the players effectively demanded that from the quests. Everything has to be soloable on the highest difficulty level without the necessity for thought or preparation. It must be accomplished by the least competent player so as to be beatable by all players.

This takes me back to LE TS. The complaints are without merit if it can be beat by using group strategies that involve willing and cooperative heals from the divine classes. That players worry about being killed in 2 hits only speaks to their inability to work together to solve a problem.

If homogenization exists it is in the players and their approach to questing.

[Baktiotha]

No. Your specific damage numbers are not just arbitrary, they are meaningless. In your post you even show how you must tailor them to get an interesting situation. Yet you claim that your method is better? Thats ridicoulous.
When will it sink in that using specific damage numbers does not do anything but approximates the EHP differences?
Every time you have proclaimed victory so far you have been completely wrong about everything, so is it not time for you to just accept that this topic is beyond you? No offence, ofcourse.

You seem to confuse all sorts of things.

You fail to understand how a specific damage number provides a metric that allows players to measure their effectiveness. If I work with a specific damage number I know if a build can sustain that much damage. That is information. It isn't necessary that the damage number represent a real case. It is important as a measurement.

If the damage number is a reasonable approximation of the maximum damage in a known encounter then it provides meaningful feedback on how the character will behave in that encounter. So working with 4500 damage from Sorjek in LE TS gives an indication on how a build will do.

But, damage numbers are not the same thing as defining bands for the purpose of comparing builds. We can use the very specific value that EHP provides. That doesn't tell me how the build does in LE TS. It does tell me if other builds with similar EHP values should be considered as functional equivalents.

To do that we don't use a +/- some %. Instead we use a +/- some value. This creates consistency. If we choose a % then 5% of 1000 is 50 and 950 is functionally equivalent to 1000 just as 1050 is functionally equivalent to 1000. Now, the two values 1000 and 950 ought to always be functionally equivalent to one another. That is, after all, what functionally equivalent suggests.

However, 5% of 950 is only 42.5 and if I look at a build with 950 EHP I'll never note that it is functionally equivalent to one with 1000 EHP. And that is because 950+42.5=992.5.

If I use a fixed value then that problem does not exist. If I want the distance to be 50 from a known EHP then 1000 includes 1050 and 950. At the same time 950 includes 1000. So now the 950 EHP and the 1000 EHP are *always* functionally equivalent.

So, the banding is completely unrelated in purpose to the damage numbers. The two things provide different types of information. EHP can provide me with information about damage. But it is not the damage information that I am comparing -- in other words, it is immaterial if EHP provides damage information or not -- when I am looking at functional equivalency.

Random damage values let me assess a build against a mock opponent (if HTK>1 and "time of avoidance">"time to heal" then viability = yes). EHP informs me when HTK=1 (if EHP=D-DR then HTK=1). EHP and/or HTK informs me when builds are similar (if EHPa - N <= EHPb <= EHPa + N then EHPa ~= EHPb OR if HTKa - N <= HTKb <= HTKa + N then HTKa ~= HTKb).

These are all different things and we can use either EHP or HTK to get to all of them.

To me the more important of these are time of avoidance informing about time to heal and thus viability AND similarities in compared values within some tolerance informing about equivalency. To me these are the more important items because similar builds should have similar performances with similar times of avoidance allowing similar times to heal and creating similar viability.

EHP doesn't tell me anything about LE TS Sorjek. HTK for posited 4500 damage does. That isn't to say that EHP *cannot* tell me something. It does tell me something if I solve for EHP>(D-DR). But what do I have if I do that? EHP=HP/M.

This means solving HP/M>(D-DR), HP>(D-D)*M. That is the same information I already have with HTK=HP/((D-DR)*M). Only I have more information because I have a value that lets me calculate how much of a heal is needed to successfully survive the next hit. As I repeat that calculation iteratively I know how large of a heal I need each and every time.

What I am trying to explain is that you are attacking what I am writing but not understanding how each tool is used. In many ways it is the same issue that I had in understanding your fixation on EHP -- only this time it is you that isn't understanding how these things are used and the information they provide.

Now, you may assert that nobody wants to know if they can survive an encounter with LE TS Sorjek. I am nearly certain that they do. So the specific damage number provides that.

You may assert that nobody cares about HTK or if there is time for healing. I am nearly certain that they do. HTK provides the gateway to that information.

Meanwhile, I've been away since Saturday and still nothing about how EHP and HtD provide any information on healing. Maybe you should stick more to demonstrating the validity of the things you are advocating and less on trying to misrepresent and discredit the processes that I am advocating.

[nokowi]

It is amusing to see an analysis that is where we were already at when I had to attend to RL.

Once people understand that the "toughness" of a build is not a measure of hits to kill but instead is a measure of how frequently a build is hit they will grasp that this is exactly where I left off in the step-by-step character analysis. All that is missing is assigning a length of time to a 1.00 toughness so we know how much time a character has available in order to obtain the needed heal.

And, I find it even more amusing that after stressing the need to model with self heals the wording in the above post is "without needing a heal." I think it is time to recognize that it isn't important where the heal comes from, just that it is needed.

My post addressed the OP. I'm sorry you cant comprehend that addressing all the builds the OP suggested (with a high dodge low PRR build) adds more to the thread than you looking at pally and barb using my calculations. Your contribution was subtracting 4 damage from barbarian and wasting our time.

If you have nothing to contribute yourself, you can at least refrain from complaining about those that do contribute.

[nokowi]

Is there a limit to the number of times someone can be banned from the forums?

As in, if someone openly admits to have been banned several times in the past and is on the forums continually derailing threads, can you report them and have them removed from the forums?

[Axeyu]

You seem to confuse all sorts of things.

You fail to understand how a specific damage number provides a metric that allows players to measure their effectiveness. If I work with a specific damage number I know if a build can sustain that much damage. That is information. It isn't necessary that the damage number represent a real case. It is important as a measurement.

If the damage number is a reasonable approximation of the maximum damage in a known encounter then it provides meaningful feedback on how the character will behave in that encounter. So working with 4500 damage from Sorjek in LE TS gives an indication on how a build will do.

But, damage numbers are not the same thing as defining bands for the purpose of comparing builds. We can use the very specific value that EHP provides. That doesn't tell me how the build does in LE TS. It does tell me if other builds with similar EHP values should be considered as functional equivalents.

To do that we don't use a +/- some %. Instead we use a +/- some value. This creates consistency. If we choose a % then 5% of 1000 is 50 and 950 is functionally equivalent to 1000 just as 1050 is functionally equivalent to 1000. Now, the two values 1000 and 950 ought to always be functionally equivalent to one another. That is, after all, what functionally equivalent suggests.

However, 5% of 950 is only 42.5 and if I look at a build with 950 EHP I'll never note that it is functionally equivalent to one with 1000 EHP. And that is because 950+42.5=992.5.

If I use a fixed value then that problem does not exist. If I want the distance to be 50 from a known EHP then 1000 includes 1050 and 950. At the same time 950 includes 1000. So now the 950 EHP and the 1000 EHP are *always* functionally equivalent.

So, the banding is completely unrelated in purpose to the damage numbers. The two things provide different types of information. EHP can provide me with information about damage. But it is not the damage information that I am comparing -- in other words, it is immaterial if EHP provides damage information or not -- when I am looking at functional equivalency.

Random damage values let me assess a build against a mock opponent (if HTK>1 and "time of avoidance">"time to heal" then viability = yes). EHP informs me when HTK=1 (if EHP=D-DR then HTK=1). EHP and/or HTK informs me when builds are similar (if EHPa - N <= EHPb <= EHPa + N then EHPa ~= EHPb OR if HTKa - N <= HTKb <= HTKa + N then HTKa ~= HTKb).

These are all different things and we can use either EHP or HTK to get to all of them.

To me the more important of these are time of avoidance informing about time to heal and thus viability AND similarities in compared values within some tolerance informing about equivalency. To me these are the more important items because similar builds should have similar performances with similar times of avoidance allowing similar times to heal and creating similar viability.

EHP doesn't tell me anything about LE TS Sorjek. HTK for posited 4500 damage does. That isn't to say that EHP *cannot* tell me something. It does tell me something if I solve for EHP>(D-DR). But what do I have if I do that? EHP=HP/M.

This means solving HP/M>(D-DR), HP>(D-D)*M. That is the same information I already have with HTK=HP/((D-DR)*M). Only I have more information because I have a value that lets me calculate how much of a heal is needed to successfully survive the next hit. As I repeat that calculation iteratively I know how large of a heal I need each and every time.

What I am trying to explain is that you are attacking what I am writing but not understanding how each tool is used. In many ways it is the same issue that I had in understanding your fixation on EHP -- only this time it is you that isn't understanding how these things are used and the information they provide.

Now, you may assert that nobody wants to know if they can survive an encounter with LE TS Sorjek. I am nearly certain that they do. So the specific damage number provides that.

You may assert that nobody cares about HTK or if there is time for healing. I am nearly certain that they do. HTK provides the gateway to that information.

Meanwhile, I've been away since Saturday and still nothing about how EHP and HtD provide any information on healing. Maybe you should stick more to demonstrating the validity of the things you are advocating and less on trying to misrepresent and discredit the processes that I am advocating.

No, if you use +/- some value the result is pointless because it there is no value that is representative of threatening monster damage. You are moving the goal post to compare against a specific encounter, which is something that I am completely uninteresed in. The entire thread has been about comparing builds in general, and in that case more EHP equals more likely to survive incoming damage, and the when you have accepted that fact it becomes glaringly obvious that if you want to compare survivability based on EHP you do it by comparing the EHP, and when you have done so you can decide for yourself if you think they are close enough to be functionally equivalent, no made up band will somehow change that from opinion to fact.

I did show you how EHP can be tied into healing.

[BigErkyKid]

It is much different to occasionally take large spikes of damage or consistently take moderate amounts of damage. When you have a predictable amount of damage taken (consistent but moderate), you can build past this through toughness or self healing. When damage spikes are larger but less frequent, it is more difficult to stay alive because you would need to build past the large damage spikes. They are not equivalent until dodge toons beat mitigation toons by some <undefined> amount.


Exactly. It plays differently. It just needs to be balanced properly. Simply having them get the same expected damage is not necessarily fair for the reasons you state. But we have seen is not that. The number of stacking PRR sources has skyrocketed, as have the dodge sources. Nowadays there isn't a big gulf in mitigation nor avoidance across builds, specially against trash. This is what I called homogenization. We may argue about the relative toughness all day long, but truth is that through feats and enhancement, jointly with item power creep the differences in the base armor are becoming very blurred, as are the defensive styles.










This is consistent with D&D (AC with dex bonus is almost identical for all builds), and it makes sense for dev's to give players more build options by allowing high dodge + light armor builds and low dodge + heavy armor builds. This is good design (more player options) and I'm not sure why you would complain about it. In DDO, defense is not equivalent across all builds (I would agree with TWF ranger it is), so in some cases you are correct that players have a reasonable choice of light or heavy armor. Good luck making a rogue heavy armor build, or rogue with shield build. You could do it but they will pale in comparison to a TWF rogue build.

Tankier toons benefit from high MRR regardless of their armor choice, and thus beat out a non-evasion toon. Evasion toons take less magical damage, but suffer from larger damage spikes.

If something like Dex-to-Damage gives the TWF ranger evasion + similar defense, then that is a result of allowing Dex to damage in the first place. That would be the problem, and not armor values.

In DND heavy armor provides fortification. That's the difference between a heavy armor and a light armor, or it was in early versions at least. The lightly armored toon was more vulnerable to crits, while mitigation based could stand them better. This translation is being broken in DDo.

[nokowi]

In DND heavy armor provides fortification.

I've only played 1st through 3.5 edition (currently 3.5), and I don't remember this to be true. What edition are you talking about?

I can argue how you said you would notice a 7.5% difference in another thread (critmult from barb thread), and that many others complained loudly about fighter having 10% more DPS than barb, but that
1) You dismiss a 8-35% difference in this relative toughness thread
2) spike damage makes the dodge toon need more healing to stay alive
3) lower HP makes the dodge toon get one/two shot more often

Exactly. It plays differently. It just needs to be balanced properly. Simply having them get the same expected damage is not necessarily fair for the reasons you state. But we have seen is not that.

2) and 3) suggest that both builds could receive the same expected damage and still play differently, with the dodge toon being worse. The fact that there is also a 8-35% difference in toughness shows that dodge toons are behind heavy armor toons in physical damage taken, which is exactly what you suggested should be the case.

The fact that you found one build that can be somewhat close (TWF ranger) doesn't make your generalization about dodge toons true because we see other cases where the toughness is much lower. You can argue TWF ranger is not implemented the way you would like (that toughness is too close on a TWF Ranger Light Armor Dodge Build), but can't make any generalizations about all dodge toons being the same because we see rogue builds with distinct disadvantages.

8% = (1.55-1.42)/(1.55) (pally THF heavy vs TWF Ranger Light Armor Dodge Build)
35% = (1.55-1.00)/(1.55) (pally THF heavy vs rogue blitz build)

If you want to investigate more light armor builds, you might be able to make your argument stronger. My judgement so far would be that you have not shown that dodge builds are equivalent in this thread. I am open to revising this if you can show some many light armor builds with toughness equivalent to pally (reaching 1.55 or better).