IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
> Best Dps At High Stats?, SPOILER: It's not poets
Rokugo
post Sep 25 2008, 02:19 PM
Post #1


Ee san
Group Icon

Group: Nobles
Posts: 401
Joined: 11-July 06
Member No.: 327
Characters: Rokugo, Pythagoreus, PaulWight, ????



After reading a lot of the balance discussions, i was curious as to which path can really put out the best DPS, given all are at similiar experiences sold (NOTE: experiences sold, not stats), so I played with a few numbers. It might be of general interest, so I'll post it here. If you don't care about the methodology, feel free to skip to the results at the bottom.

Hokay, so the idea is each of these people is being cycled beyond perfection, for the purposes of simplification. This means that everyone's DPS is bumped up, but I'd say that it's almost by the same amount for each path, maybe warriors more than rogues and mages, since it's pretty much impossible to aether a buff warrior's zerks, while the biggest simplifying assumpion for rogues is that they get two full DAs, instead of the one that they get in reality.

So I used the following formula to calculate DPS (Not in any useful units, but all I care about is relative to other paths, so I will normalize at the end)
Vita DPS = Vita * sum[over attacks](vita multiplier/aethers)
Mana DPS = Mana * sum[over attacks](mana multiplier/aethers)

Obviously, it's pretty much impossible to perfectly aether many of these attacks (zerk for (buff) warrior, da/ls for rogues, hellfire for mages, and even ret for poets has 20s aethers according to NA, lower than the 23s aethers for invoke)

I used what I think of as the "high stats approximation" to get ratios for each path that equate to similiar experiences sold. I checked these with stataddict for decently high stats (~5m total stats sold), and they worked well. The approximation is as such: At very high stats, total experience sold is directly proportional to stat^2. I started with a rogue with a 3/1 ratio. So an entirely vita warrior would be sqrt(3^2+(2*1)^2) ~ 3.6. Obviously warriors have some mana, so this is a little bit of the built in error in this calculation. I have chose to ignore it because it doesn't add much to the damage of the warrior or to his total experience sold. Again, the purpose is to get a general idea, so if you don't like these simplifying assumptions, feel free to do a more rigorous analysis. I also chose a 1/1 ratio for the poets and mages.

Anyway, here are the results:
CODE

    warrior    rogue    mage    poet
vita    3.60    3.00    1.60    1.60
mana    0.00    1.00    1.60    1.60
vita DPS    0.74    0.50    0.00    0.00
mana DPS    0.00    0.20    0.39    0.03
total    0.74    0.70    0.39    0.03
normalized    1.00    0.94    0.53    0.04


So in this simple model rogues and warriors do approximately the same DPS. Mages do about half the DPS of either of them, and poets come in at about 1/25 of the damage of the two fighter paths.


--------------------
Hi, I'm Rokugo. You probably don't know me.
There is a 95% chance the above post is sarcasm and is not to be taken seriously.
I play Real Life to help escape NexusTK.
"Baby, for you I'd divide by zero" <--- My # 1 pickup line
Go to the top of the page
 
+Quote Post
darkmaverick
post Sep 25 2008, 11:41 PM
Post #2


Oh san
Group Icon

Group: Nobles
Posts: 4,627
Joined: 4-July 06
Member No.: 90
Characters: DarkMaverick



This isn't considering things like overflow or rage, is it?


--------------------
I host The RPG Fanatic Youtube Show. Home of fanatical reviews and commentary about RPGs.
And my top picks for Sa san warrior shield were......

...I totally want that sword too.
Go to the top of the page
 
+Quote Post
Rokugo
post Sep 26 2008, 06:36 AM
Post #3


Ee san
Group Icon

Group: Nobles
Posts: 401
Joined: 11-July 06
Member No.: 327
Characters: Rokugo, Pythagoreus, PaulWight, ????



Overflow is more or less irrelevant. The multiplier is something like 1.05 now, so the DPS it adds by using it is small enough that I can ignore it for simplicity. There are much larger simplifications thatn that in this model that I illustrated.

Rage and cunning are ignored. it would be quite difficult to stick those in here in a sane way. I am thinking of doing an optimal poet/warrior duo exp/rage calculation, but that will be a bit more complicated.


--------------------
Hi, I'm Rokugo. You probably don't know me.
There is a 95% chance the above post is sarcasm and is not to be taken seriously.
I play Real Life to help escape NexusTK.
"Baby, for you I'd divide by zero" <--- My # 1 pickup line
Go to the top of the page
 
+Quote Post
LeXington
post Sep 29 2008, 03:28 PM
Post #4


Ee san
Group Icon

Group: Nobles
Posts: 453
Joined: 10-September 07
From: Edmond, OK
Member No.: 3,989
Characters: Sanguis, LeXington, Pagan, Hwarang



Not to knock you, I appreciate your efforts in your calculations but it's rather a moot point to do so. It is impossible to get a max DPS calculation that is accurate due to spawn numbers and times in between, Cunning/Rage cast timing, equipment variations, time it takes poet to put down his cigarette and heal lol.

It's not possible.


--------------------
Abyssus abyssum invocat~!
Go to the top of the page
 
+Quote Post
AceoStar
post Sep 29 2008, 03:48 PM
Post #5


Oh san
Group Icon

Group: Retired Staff
Posts: 3,120
Joined: 3-July 06
Member No.: 2
Characters: AceoStar



Thats why you use constants to assume the same spawn, cunning etc. He isnt saying every Rogue earns this DPS, just that a rogue has the potential to earn this DPS in the mentioned setting


--------------------
AceoStar, Imperial Guard of Buya.

Put www.StatAddict.com in your profile to track your stats today!


Learn to keep up


Errr, NF is filtered at work these days, I'll be checking the forums a lot less often.
Go to the top of the page
 
+Quote Post
Rokugo
post Sep 29 2008, 03:55 PM
Post #6


Ee san
Group Icon

Group: Nobles
Posts: 401
Joined: 11-July 06
Member No.: 327
Characters: Rokugo, Pythagoreus, PaulWight, ????



QUOTE(LeXington @ Sep 29 2008, 02:28 PM) [snapback]63175[/snapback]

It is impossible to get a max DPS calculation that is accurate due to spawn numbers and times in between, Cunning/Rage cast timing, equipment variations, time it takes poet to put down his cigarette and heal lol.


Many of those things don't matter for a -max- DPS calculation, especially one like this that wasn't intended to include rage or cunning.

Equipment has negligible effect on stats at the very high end, so it doesn't affect this all that much. Cunning/Rage wasn't considered at all, the purpose of this was to compare vita attacks. Impossible expectations for poets were covered in my discussion of the methodology, but the "time that it takes the poet to put his cigarette down and heal" certainly has no effect on -max- DPS. Similarly for spawn numbers, it's not something important to me here, because I really only care about the theoretical output of their damages, nowhere did I claim that this applied as anything else.

But, yes, there are many factors that complicate accuracy of the calculation, and that's why I covered many of these sources of error in my first post. This should definitely not be used to conclude, for example, that warriors are 6% better than rogues, or really any sort of direct hunting comparison. But from this I would say that claims that rogues have the best multipliers on spells have been somewhat exaggerated.

Edit: Also, what AceoStar said. He said it much better than I could tongue.gif


--------------------
Hi, I'm Rokugo. You probably don't know me.
There is a 95% chance the above post is sarcasm and is not to be taken seriously.
I play Real Life to help escape NexusTK.
"Baby, for you I'd divide by zero" <--- My # 1 pickup line
Go to the top of the page
 
+Quote Post
LeXington
post Sep 29 2008, 04:03 PM
Post #7


Ee san
Group Icon

Group: Nobles
Posts: 453
Joined: 10-September 07
From: Edmond, OK
Member No.: 3,989
Characters: Sanguis, LeXington, Pagan, Hwarang



Ah, sorry I misunderstood.

I'm curious, could you please post max potential DPS for a mage individually? I've always wondered.

PS- I loved the "Hokay" reference.


PPS- Alaska can come to!


--------------------
Abyssus abyssum invocat~!
Go to the top of the page
 
+Quote Post
Rokugo
post Sep 29 2008, 04:12 PM
Post #8


Ee san
Group Icon

Group: Nobles
Posts: 401
Joined: 11-July 06
Member No.: 327
Characters: Rokugo, Pythagoreus, PaulWight, ????



Sure, you mean a basic invoke sam hf invoke reg hf, invoke ferno cycle? To remove the "perfect spiring poet" assumption, all I have to do is change the aethers to sam hf, regular hf, and ferno to 70s each. That gives (relative to the warrior == 1) 0.36 for the uninspired mage. Which isn't too surprising, since the real bread and butter of the mage's DPS is Inferno (and sam HF), and this effectively takes out two regular hellfires per 70s.

Edit: For those interested, the total row can be interpreted as the mils of raw damage per second for a fighter with the vita and mana given above in millions. So a 3.6 mil warrior's theoretical max DPS is about 740k damage per second. Also, the value for the uninspired mage was 0.27, so about 270k raw damage per second for a 1.6/1.6 mage.


--------------------
Hi, I'm Rokugo. You probably don't know me.
There is a 95% chance the above post is sarcasm and is not to be taken seriously.
I play Real Life to help escape NexusTK.
"Baby, for you I'd divide by zero" <--- My # 1 pickup line
Go to the top of the page
 
+Quote Post
darkmaverick
post Oct 1 2008, 10:54 PM
Post #9


Oh san
Group Icon

Group: Nobles
Posts: 4,627
Joined: 4-July 06
Member No.: 90
Characters: DarkMaverick



QUOTE(Rokugo @ Sep 26 2008, 07:36 AM) [snapback]62626[/snapback]

Overflow is more or less irrelevant. The multiplier is something like 1.05 now, so the DPS it adds by using it is small enough that I can ignore it for simplicity.


Overflow is hardly irrelevant.

Theoretically, it allows a Warrior to cast a modified Whirlwind, Berserk, and Siege on up to 3 targets if used on an enemy at 10 vita who is surrounded by three additional enemies. Even though there is a cut on the damage, it basically allows you to cast a weaker version of your attack on more than one enemy at a time.

I've had Mages that setup overflow sets in such a way, and heal the enemy out of death so I can overflow all three targets. It makes a substantial difference in the number of enemies killed, especially if you can kill more than one at a time.

Example: Say your Siege deals 2mil damage. You cast this on an enemy who is at 20k Vita. So 2mil- 20k = 1,980,000

This overflow damage is reduced down 693,000 damage [ (1 980 000 / 3) x 1.05]

So.. 693,000 damage dealt against 3 enemies = 2,079,000 damage for that attack

You do not believe increasing damage output on multiple targets is significant, especially if you only need the reduced damage to kill 3 enemies? Consider a PVP situation. It is entirely possible to kill multiple enemies in a shorter amount of time by using overflow. It spreads your damage over several targets, rather than consuming the whole damage value on only 1 target.

Looking at the raw damage output is fine and all, but your calculations are not realistic because you are throwing the damage at thin air. That is why Rogues and Warriors come out similar DPS, even though the damage between R6 and Cun5 are very similar, and Rogues and Warriors (on paper) appear to do similar DPS.


--------------------
I host The RPG Fanatic Youtube Show. Home of fanatical reviews and commentary about RPGs.
And my top picks for Sa san warrior shield were......

...I totally want that sword too.
Go to the top of the page
 
+Quote Post
Vortextk
post Oct 1 2008, 11:00 PM
Post #10


Sam san
Group Icon

Group: Nobles
Posts: 808
Joined: 6-February 07
From: Cape Canaveral, Fl
Member No.: 1,757
Characters: Frog



You're making it more complex than Rokugo is attempting to show the numbers..make a new thread?
.
If you fight 1 monster with 100 million vita and just vita the hell out of it, who would kill the fastest? This is what his numbers show. And adding in OF is irrelevant and at best, if you like summon a horse infront of it and overflow for the 5% increase, it won't be constant enough to add into his calculation.



--------------------
IPB Image
Pandemonium Frog My Progress
Warriors; so easy a caveman could do it.
Go to the top of the page
 
+Quote Post
darkmaverick
post Oct 1 2008, 11:12 PM
Post #11


Oh san
Group Icon

Group: Nobles
Posts: 4,627
Joined: 4-July 06
Member No.: 90
Characters: DarkMaverick



QUOTE(Vortextk @ Oct 2 2008, 12:00 AM) [snapback]63491[/snapback]

You're making it more complex than Rokugo is attempting to show the numbers..make a new thread?
.
If you fight 1 monster with 100 million vita and just vita the hell out of it, who would kill the fastest? This is what his numbers show. And adding in OF is irrelevant and at best, if you like summon a horse infront of it and overflow for the 5% increase, it won't be constant enough to add into his calculation.


Why would I make a new thread? Is the objective of science not to investigate and discover the way the world works? All I'm doing is pointing out a flaw in the method and how the results can be misleading.

Besides, what is the point of a measuring system that does not accurately measure the very thing it is trying to categorize? In the case of damage, measuring damage per second is quite complex because you can't measure damage that doesn't deal damage to anything. It basically doesn't exist.

(To clarify what I mean, if you have a target that needs 10k damage to die, but you deploy 50k damage, only 10k of that damage matters. The other 40k damage goes poof. It basically doesn't exist, and I dont believe many people instantly realize that because we are conditioned to assume you can't lose energy)

I admit, if we're talking about something like the Forever tree or a boss, sure, the chart works. But NexusTK hunting mechanics relies heavily on the defeat of many common enemies, not high Vita targets. You aren't rewarded for dealing huge damage to one target, but spreading damage out amongst multiple targets.

When people generally talk which paths deal the most damage in Nexus, they actually mean who can defeat the most enemies in the same frame of time.

I'm not pointing out a flaw because I'm trying to knock his work. It is excellent work so far. All I'm doing is I'm trying to help make his calculations more accurate and meaningful.


--------------------
I host The RPG Fanatic Youtube Show. Home of fanatical reviews and commentary about RPGs.
And my top picks for Sa san warrior shield were......

...I totally want that sword too.
Go to the top of the page
 
+Quote Post
Vortextk
post Oct 1 2008, 11:21 PM
Post #12


Sam san
Group Icon

Group: Nobles
Posts: 808
Joined: 6-February 07
From: Cape Canaveral, Fl
Member No.: 1,757
Characters: Frog



Because adding in rage is impossible, adding in overflow is basically meaningless, saying that you are hitting multiple targets and killing them faster doesn't change raw dps, just kills and xp/time.

I said why not make a new thread because you're adding, not fixing, and he seems to be happy with how he presented it.


--------------------
IPB Image
Pandemonium Frog My Progress
Warriors; so easy a caveman could do it.
Go to the top of the page
 
+Quote Post
darkmaverick
post Oct 1 2008, 11:24 PM
Post #13


Oh san
Group Icon

Group: Nobles
Posts: 4,627
Joined: 4-July 06
Member No.: 90
Characters: DarkMaverick



QUOTE(Vortextk @ Oct 2 2008, 12:21 AM) [snapback]63493[/snapback]

Because adding in rage is impossible, adding in overflow is basically meaningless, saying that you are hitting multiple targets and killing them faster doesn't change raw dps, just kills and xp/time.

I said why not make a new thread because you're adding, not fixing, and he seems to be happy with how he presented it.


It isn't impossible, and I think the method he is using to make these calculations can be refined to accurately measure it. Time consuming perhaps because of the sheer number of variables, but entirely possible.

Bear in mind, this chart is extremely old and uses very, very outdated information (aethers are different, melee damage formula is different, etc) and has some errors (some figures are off, and I it didn't count the time per movement action which I would now do), but the method I used is theoretically sound, and with refinement could be used to make accurate projections:

CODE
Standard Ee san subpath Warrior
Spells: Master weapon (x15 on weapon)
Spirit fury (x5 fury)
Weapon:Ee san spike 90m350
Might2 Dam7
Helm:CR helm Might3
Ring: Two Ambrosias Might8 Dam5
Armor: Wind plate Dam1 Might4
Base stats: 130/130/130
400k Vita

Total stats after equipment:155 Might Dam18
Base might damage:31 Damage points
Might damage equals 49 damage points after calculating Dam bonus into base Might.

[Combat data]
(Conditions:Against 0 Ac to the front)
Attack damage with weapon and spells:3,630 to 13,380 damage points.
Attack damage of Slash:11,537
!Combination attack (Slash and Swing Method):15,167 to 24,917 damage points.
Attack damage of Beserk:300k damage
Attack damage of Assault:200k damage on single target; 800k total damage on Four targets.
Attack damage of 10% Whirlwind:630k damage

Battleground: Rat Plague3
Enemy: Brown rat (requires total of 960k damage to destroy)

[i]Note: Kill zone meant the number of damage needed to kill the target decreased to that amount.[/i]

Sequence: Warrior WWs and zerks a brown rat to destroy it (1 second for ww, 4s for heal, 1 sec to zerk, 3s for heal). Warrior then Assaults 4 rats, bring all 4 down to 760k kill zone (2 seconds to place in front of set and cast). Warrior then requires heal. (2 seconds) After heal Warrior waits 3 seconds to zerk 1 rat, taking it to 460k kill zone (2 seconds) Warrior then requires heal (3 seconds). After heal Warrior WWs different rat in previous Assault set, bringing it to 130k kill zone (1 second). Warrior uses Slash and Swing method, killing it in 7 to 4 strikes. (aprox 12 seconds) Warrior then zerks rat that was taken to 460k kill zone, taking it to 160k kill zone (2 seconds). Using Slash and swing method, warrior destroys rat in 9 to 5 strikes (aprox 15 seconds). Warrior waits 4 seconds then repeats previous WW+slash and zerk+slash scenarios to remaining two rats; defeating them.

Time taken: 80 seconds approximately to defeat 5 enemies.


I assume Rokugo has written a script, and that would be superior way to do the calculations than I've used (which is by hand).


--------------------
I host The RPG Fanatic Youtube Show. Home of fanatical reviews and commentary about RPGs.
And my top picks for Sa san warrior shield were......

...I totally want that sword too.
Go to the top of the page
 
+Quote Post
MeduLLA
post Oct 6 2008, 05:40 AM
Post #14


Ee san
Group Icon

Group: Nobles
Posts: 394
Joined: 23-August 06
Member No.: 698



I'd honestly like too see that scenario played out with numbers that are more up to date and accurate.

I'd also like to see it done with a rogue.

I'm curious about the whole arguement if Warrior's are so much better than Rogues. If the damage comes out almost equivalent, then what's the problem in game? Lazy poets? Lazy people? It would be interesting to see...and would provide some basis for possible rogue improvement techniques. Maybe we could come up with a collective way to improve Rogues vs Warriors. I'm sure if the data is analised in the correct manner we are bound to find a dead spot somewhere in a rogues cycle.


--------------------
Go to the top of the page
 
+Quote Post
Horsham
post Oct 6 2008, 03:48 PM
Post #15


Peasant


Group: Peasants
Posts: 4
Joined: 13-February 08
Member No.: 7,355



Does anybody still have the calculations sheets Yari et al did a while back, they were like MoreFighterAttacks.xls?

Within there he had information about various damages dealt for different stat ranges. Il/Ee/Sam/Sa/Oh, along with poet figures.

I ran some of my own and the general conclusion is that a 4/1 rogue versus a 10/1 warrior the rogue will kill 11 things as fast as the warrior kills 16 however the rogues kills are 80% stronger than the warriors (assuming every attack 1 hits, the rogue out damages the warrior 57/35). If we assume perfect overflow instead of only 1 hit kills, the warriors kills jump up to the incredible number of 33. However total damage done by the warrior is only 28% (73/57) more than the rogues damage total, and the warrior must have every attack be a perfect overflow for this advantage. Basically the rogue should be hunting somewhere much stronger than the warrior. The largest problem is that end game fighters are all trying to hunt in the same location. A 3m vita rogue shouldn't be in hunter3 anymore. They end up wasting most of their attack damage. The warrior would require 180% more healing per attack cycle than the rogue.

As true hunting happens somewhere between 100% kill and 100% overflow this means that the vita attacks are balanced, it is the hunting locations that are the problem.

The problem with balance is where swing damage far exceeds vita attack potential. The power of early cunnings and late rages are too high for the caves which they are in. Early rages need to be tweaked up, late rages tweaked down. I'd like to see the cost of rage changed with more of an emphasis on vita than mana, keep the approx. total stats the same but make it so that a warrior has to have like 400/80 for rage6 instead of going 160/140. This will make newly 99 warrior vita attacks more effective.

The final problem that needs to be addressed is keeping poets and mages in similar caves as equal stated rogues and warriors. Most of this problem is evident in vortex, where magicians are able to get into groups in their highest possible cave, while I can't foresee a good way to prevent this, by elevating the entry levels of vortex3 caves to closer to the stats that would be required to hunt there it may reduce the problem some. Increase the separation between caves, make Assn weak like Rabbit, but make Water strong like Dragon.
Go to the top of the page
 
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 

RSS Lo-Fi Version Time is now: 22nd September 2017 - 10:11 AM
Nexus Forums is part of: Nexus Atlas © all rights reserved.