Berserker - Damage Variables

Started 18 Oct 2020
by skipari
in Midgard
Hi,

i was looking around a bit to get me some better insight in the berserker and the damage output the class can provide, considering weapon speed, race and so on. I thought first about putting this all in a forum post, but it got a little bit more extended then i assumed... Also since its a draft, maintaining it on two different sites probably would lead to an outdated version soon enough, so for while it is not finished only external.

So first draft is here: https://daoc-skjold.org/berserker.html


There is still some stuff missing like twohand and different growth rates and its effects for example, and at least one chart has to be graphed again since it doesn't fit the style of the others, but its like halfway done and maybe you guys have some stuff you want to know or see some errors i have somewhere.

For any kind of feedback feel free to write here, "Fintaru" in game or yrgrasil#3576 in discord.
Sun 18 Oct 2020 7:17 PM by DinoTriz
"Whatever tickles your dipstick" was funny, I'll have to use that.

What I learned:

1. Dehaste actually helps probably way more than people realize.

2. A slow off-hand is better in prime situations where Celerity is up all the time.

3. Eye patches are superior.
Mon 19 Oct 2020 4:04 PM by skipari
Did some update regarding the offhand left axe formula, basically this

Phoenix actually uses an Offhand Modifier of 52.50 + 0.34 *
left_axe_spec instead 52.50 + 0.52 * left_axe_spec as above links let assume.
I could not find a reason for it, but in average this translates to a ~10-12%
dps decrease for the offhand. Since most dps come from the mainhand anyway this
is just a ~3-4% overall decrease compared to the original LA NF formula.

see also:
https://forum.playphoenix.online/realm/midgard/4339-left-axe-damage-current-status?page=3
https://forum.playphoenix.online/get-involved/ask-the-team/6140-celtic-dual-dual-wield-and-left-axe-buffs-no-forum-announcement?page=2

example + damage over time charts should be updated. The barcharts aren't updated yet since i want to overhaul them anyway somewhen.

--

Also added the twohand modifier in the math section as preparation for the next graph (soon)
Thu 22 Oct 2020 7:32 AM by trawetsnivek
Interesting bit about the valkyn out damaging a troll in the 20 second window. 20 seconds is rather arbitrary though. If you had extended the horizontal axis to 21 seconds the troll would out damage the valkyn. A better way to compare races is to find out what their average damage per hit is using an anytime style and then divide that number by your swing speed to find damage per second.
Thu 22 Oct 2020 9:35 AM by skipari
trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
Interesting bit about the valkyn out damaging a troll in the 20 second window. 20 seconds is rather arbitrary though. If you had extended the horizontal axis to 21 seconds the troll would out damage the valkyn.

i expected some kind of this, therefore i added beforehand on the left of the graph the values after 20s

troll_nocele.csv
18.60s, 2684
20.46s, 2928
22.32s, 3172
24.18s, 3416

valkyn_nocele.csv
19.80s, 2834
21.45s, 3052
23.10s, 3270
24.75s, 3488

Of course when the troll hits again he does more damage again for whooping 0.99s (which decreases again over time since the valk hits faster). And regarding your logic every race outdamages the other one if he just styled.

trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
A better way to compare races is to find out what their average damage per hit is using an anytime style and then divide that number by your swing speed to find damage per second.

first one is troll, second is valkyn

>>> 244/1.86
131.18279569892474
>>> 218/1.65
132.12121212121212
Thu 22 Oct 2020 11:49 AM by inoeth
once again proven: speed is king in this game
Thu 22 Oct 2020 11:58 AM by Sepplord
more like, the amount of times this difference will matter is neglectable
Thu 22 Oct 2020 12:41 PM by skipari
Sepplord wrote:
Thu 22 Oct 2020 11:58 AM
more like, the amount of times this difference will matter is neglectable

Yeh, its just something interesting, but has barely a impact
Thu 22 Oct 2020 12:41 PM by inoeth
Sepplord wrote:
Thu 22 Oct 2020 11:58 AM
more like, the amount of times this difference will matter is neglectable

for a berserk yes maybe, but this also applies to shadowblades and there it does matter if you are able to pull off one more hit or not
Thu 22 Oct 2020 2:54 PM by Sepplord
inoeth wrote:
Thu 22 Oct 2020 12:41 PM
for a berserk yes maybe, but this also applies to shadowblades and there it does matter if you are able to pull off one more hit or not
the comparison shows that the slower attacking troll still pulls ahead momentarily in between though, so it still could go either way
It would be intresting to see average effective hitpoints of different classes/targets and who actually kills faster, though.

And going into the assassin territory, you don't only need to look at how fast you kill the enemy, also how fast they kill you, and how that matches against each other. Technically AugCon1 could be the difference between being dead, or surving until another 2handed 600dmg blow comes in 3seconds later letting you attack twice before.
Woar, two whole attack rounds onto the enemy for only a single RA point?

How many hits does an assassin fight take usually? And how many swings are needed to reach that many hits? And has the fast race already pulled ahead in that short timeframe? etc...


Also consider (though this applies more to a berserk than an assasin as assassin have other DPS-RA options and in 1vs1 defense is also important) at high RR a slow race will still be able to cap attackspeed without sacrificing other dmg RAs because there just aren't that many. While the fast race would just be able to put more points into defensives
Thu 22 Oct 2020 3:39 PM by skipari
Sepplord wrote:
Thu 22 Oct 2020 2:54 PM
the comparison shows that the slower attacking troll still pulls ahead momentarily in between though,

i hate this specific argument since it can be used from all races, since every race has momentarily in between where it hits harder then any other, either through damage or speed, only exception is dwarf/norse. It is technically not wrong, but can be used by both sides and makes it therefore somewhat useless.

Sepplord wrote:
Thu 22 Oct 2020 2:54 PM
so it still could go either way
It would be intresting to see average effective hitpoints of different classes/targets and who actually kills faster, though.

And going into the assassin territory, you don't only need to look at how fast you kill the enemy, also how fast they kill you, and how that matches against each other. Technically AugCon1 could be the difference between being dead, or surving until another 2handed 600dmg blow comes in 3seconds later letting you attack twice before.
Woar, two whole attack rounds onto the enemy for only a single RA point?

How many hits does an assassin fight take usually? And how many swings are needed to reach that many hits? And has the fast race already pulled ahead in that short timeframe? etc...


Also consider (though this applies more to a berserk than an assasin as assassin have other DPS-RA options and in 1vs1 defense is also important) at high RR a slow race will still be able to cap attackspeed without sacrificing other dmg RAs because there just aren't that many. While the fast race would just be able to put more points into defensives

The damage over time graphs are done with a chain profile and avg random + full resists, so its one of the more worse cases. I can easily create graphs/data tables with a different profile if requested, also ra's etc are not really a problem.

Don't forget this all is just a clean laboratory consideration, if we consider hit/procrates and all the other rng stuff then it would go really crazy.
Thu 22 Oct 2020 5:46 PM by trawetsnivek
skipari wrote:
Thu 22 Oct 2020 9:35 AM
trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
Interesting bit about the valkyn out damaging a troll in the 20 second window. 20 seconds is rather arbitrary though. If you had extended the horizontal axis to 21 seconds the troll would out damage the valkyn.

i expected some kind of this, therefore i added beforehand on the left of the graph the values after 20s

troll_nocele.csv
18.60s, 2684
20.46s, 2928
22.32s, 3172
24.18s, 3416

valkyn_nocele.csv
19.80s, 2834
21.45s, 3052
23.10s, 3270
24.75s, 3488

Of course when the troll hits again he does more damage again for whooping 0.99s (which decreases again over time since the valk hits faster). And regarding your logic every race outdamages the other one if he just styled.

trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
A better way to compare races is to find out what their average damage per hit is using an anytime style and then divide that number by your swing speed to find damage per second.

first one is troll, second is valkyn

>>> 244/1.86
131.18279569892474
>>> 218/1.65
132.12121212121212

It's not my logic. It's the way you arbitrarily chose 20 seconds. You can choose any amount of time to prove that one race does more damage than the other. I'm just pointing out the flaw in your analysis of the graph.

Also, you have to factor in the alpha hit since you are basically getting a free hit without swing speed being a factor.

If we take your numbers above the troll has a 141.27 damage per second after 24.18 seconds. This number will converge towards the 131.18 DPS you calculated if you run the damage over a long enough period. The valkyn has a 140.92 DPS in 24.75 seconds and will converge towards 132.12.

So essentially in long fights, Valkyns will deal more damage than trolls. In shorter fights trolls win.
Thu 22 Oct 2020 6:23 PM by skipari
trawetsnivek wrote:
Thu 22 Oct 2020 5:46 PM
skipari wrote:
Thu 22 Oct 2020 9:35 AM
trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
Interesting bit about the valkyn out damaging a troll in the 20 second window. 20 seconds is rather arbitrary though. If you had extended the horizontal axis to 21 seconds the troll would out damage the valkyn.

i expected some kind of this, therefore i added beforehand on the left of the graph the values after 20s

troll_nocele.csv
18.60s, 2684
20.46s, 2928
22.32s, 3172
24.18s, 3416

valkyn_nocele.csv
19.80s, 2834
21.45s, 3052
23.10s, 3270
24.75s, 3488

Of course when the troll hits again he does more damage again for whooping 0.99s (which decreases again over time since the valk hits faster). And regarding your logic every race outdamages the other one if he just styled.

trawetsnivek wrote:
Thu 22 Oct 2020 7:32 AM
A better way to compare races is to find out what their average damage per hit is using an anytime style and then divide that number by your swing speed to find damage per second.

first one is troll, second is valkyn

>>> 244/1.86
131.18279569892474
>>> 218/1.65
132.12121212121212

It's not my logic. It's the way you arbitrarily chose 20 seconds. You can choose any amount of time to prove that one race does more damage than the other. I'm just pointing out the flaw in your analysis of the graph.

Also, you have to factor in the alpha hit since you are basically getting a free hit without swing speed being a factor.

If we take your numbers above the troll has a 141.27 damage per second after 24.18 seconds. This number will converge towards the 131.18 DPS you calculated if you run the damage over a long enough period. The valkyn has a 140.92 DPS in 24.75 seconds and will converge towards 132.12.

So essentially in long fights, Valkyns will deal more damage than trolls. In shorter fights trolls win.

Ah now i understand what you mean, and yes of course the free hit at 0:00 gives a advantage which normalizes over time. I wrote in the page itself that the valk will outdamage a troll on the long without specifing a time, should probably add this and a note that just by the numbers at this time the troll has mathematically still an dps advantage.

Still i don't like the argument itself, too many factors will play in, the hitchance combined with the increased rate alone will get the valk far earlier to the point he overtakes the troll, not even to speak of procs/already dead target etc.


As a explaination for the arbitrary 20s, it is just cele/teddy time and i don't wanted to fiddle around with several different timespans. And i think if you hit a target for >20s then there is some other far more important issues that need to be fixed first.
Thu 22 Oct 2020 8:43 PM by DinoTriz
So play whatever looks cool to you?
Thu 22 Oct 2020 8:58 PM by skipari
DinoTriz wrote:
Thu 22 Oct 2020 8:43 PM
So play whatever looks cool to you?

Well, that is what i would suggest anyway, the race difference is not really huge, or from what i see more a thing for the people who want to squeeze out the last 1-2% out of every class. More important is actually to have the right weapon, as in the already 4.1/2.4 defacto standard. But switching to a fast 2h or better 4.1/4.0 in case of celerity is something everyone should really consider. Since phoenix has the /switch command even old tards like me can singlehanded change the offhand.

The race importance would probably change if the server had used the old enemy armor af variable, and therefore high strength build can more consistent scratch at the damage modifier cap. But for whatever reason the live variable is used and therefore all melee dps reduced by ~20ish % (even before considering that all realms have here spec af)
Thu 29 Oct 2020 8:58 PM by skipari
added now
- additional notes to factor in above discussion
- 2h damage
- +Weapon Effect to style damage
- some ranting about spec af and others

I think i'm somewhat done there actually, there is still some smoothing going on in the background when i've time (especially poor wording/phrasing, english sometimes sucks as non native speaker). And i'll somewhen add html anchors and a table of content for better usability (somewhat my initial idea of a short summary got wrecked...). Maybe in the future when i've too much time i rework the whole thing and extract the math part from the rest since i want to do that for other classes/mechanics soonish.

So if you miss something or think there is something wrong etc. feel free to yell at me.
This topic is locked and you can't reply.

Return to Midgard or the latest topics