Effectiveness of Attack Debuffs versus Slow

[HuacaWuaca]

I remember reading something a few years ago about the effectiveness of Attack Debuffs versus Slows. I'm recently coming back to the game, so I wanted to look up the info and see if anyone had done any parses against current content. I asked about it in a serverwide channel and really didn't get any substantiated answers. The answers were mainly "Slows are better" or "they stack, use them both".

So, I'm wondering if anyone had done any parses to compare the effectiveness or if I'll be starting this on my own.

I saw some debate about it in: viewtopic.php?f=6&t=6746&start=0&st=0&sk=t&sd=a&hilit=attack+debuff

[Kumudil]

On the sonyboard there are some parses somewhere on the druids sections. Those parses where done with tosk-ogres.
If I remeber well, the result has been that on these mobs the best 2 attack debuffs would reduce the dps similar to our slow.

The main effect of attack debuffs, as I see it, is not the dps-reduction but the reduction in spikes. The dps is not reduced evenly like with slow, but by reducing the likehood of max-hits.

The ability to heal a tank against a mobs dps is highly dependent on the density of "lucky rounds" on the mobs side. If the mob can dish a couple of max-hits-rounds back to back, the likehood to get a dead tank is very hight. Its mainly the damage-spikes which kill the tank, not the averange dps.

[Kumudil]

Here we are:

Turgurs
A Rallosian champion on 21.02.2010 in 468sec
Beabia --- DMG to PC: 928344 @1507dps
Produced by GamParse v1.0.3
/G Tanking summary for: Beabia --- Total damage: 928344 --- Avg hit: 1555 --- Swings: 954


Blanched Frost
A Rallosian champion on 21.02.2010 in 585sec
Beabia --- DMG to PC: 986581 @1636dps
Produced by GamParse v1.0.3
/G Tanking summary for: Beabia --- Total damage: 986581 --- Avg hit: 1506 --- Swings: 1110


HoR and Blanched Frost
A Rallosian champion on 21.02.2010 in 700sec
Beabia --- DMG to PC: 1129168 @1551dps
Produced by GamParse v1.0.3
/G Tanking summary for: Beabia --- Total damage: 1129168 --- Avg hit: 1425 --- Swings: 1350




/G Tanking summary for: Beabia --- Total damage: 928344 --- Avg hit: 1555 --- Swings: 954 --- Defended: 179 (18,8%) ---
/G Tanking summary for: Beabia --- Total damage: 986581 --- Avg hit: 1506 --- Swings: 1110 --- Defended: 241 (21,7%) ---
/G Tanking summary for: Beabia --- Total damage: 1129168 --- Avg hit: 1425 --- Swings: 1350 --- Defended: 323 (23,9%) ---




turgurs: 468 sec 954 swings :: 2.04 swings per second
Blanched: 585 sec 1110 swings :: 1.88 swings per second
both : 700 seconds 1350 swings :: 1.92 swings per second

Looking at it now, it seems like attack debuff not only reduce the Avg hit (by cutting max-hit), but also have the effeckt to increase the tanks chance to defend hits completely.
Looking at these numbers, it looks like this effect is so big, that the total amount of real hits per second incoming is less using just 1 attack debuff than using slow.

But could be as well just be caused by the relative short time of the parses and this conclusion is complete off.


Those parsese where done in tosk using a T5 Tank merc to have a situation anyone can set up if he wants to proove whats been parsed on its own.

[Huurgh]

Ok thats in tandem with a dru, which grp wise almost never happens to me.
What about our own debuffs:

Feralize
Speckle of Runedust (I am still unsure if this is worth it to be used. The debuff is one used by NPC and lands always, the rest is .. smoke and mirrors?)

Is SoRD really worth the mana to be used on every mob in UF?

[HuacaWuaca]

Thanks for the response. I'll check out the SOE Forums Druid Section to look for the parses there.

[HuacaWuaca]

If anyone is interested:

http://forums.station.sony.com/eq/posts ... _id=125312

[Fenier]

viewtopic.php?f=51&t=5637

[Kumudil]

Hi Fenier,
I think you are the one who knows more about attack debuffs than anyone else.

/G Tanking summary for: Beabia --- Total damage: 928344 --- Avg hit: 1555 --- Swings: 954 --- Defended: 179 (18,8%) --- (this was slow only no attack debuff)
/G Tanking summary for: Beabia --- Total damage: 986581 --- Avg hit: 1506 --- Swings: 1110 --- Defended: 241 (21,7%) --- (Blanched Frost)
/G Tanking summary for: Beabia --- Total damage: 1129168 --- Avg hit: 1425 --- Swings: 1350 --- Defended: 323 (23,9%) --- (HoR and Blanched Frost)

The parses are relative short, but seem to show, that attack debuff apart from the known effect to shift the likehood of high damage hits versus low damage hits has a noteable influence to the mobs accuracy as well. All 3 parses were done on same mob. The amount of hits completely defended by the tank seems to increase the more attack is debuffed.

Have you ever noticed similar? Have you ever had your focus on that aspect when analysing your parses?

[Fenier]

For comparison purposes you look at damage over a equal number of swings. The number of swings should be fairly high to rule out RNG variance, but you can see trending in less. Your 1000 Swing sample size should have ruled out variance fairly well.

Atk debuffs would by shifting the DI value to lower integers. This means you are shifting the hit distribution curve so you get less max hit values and minimum values. Your parses show this. Because this calculation takes into effect the amount of Armor the person tanking is using, you will see larger shifts on people with less armor and lower shifts on Tanks who are AC focused.

Nothing I parsed indicated an Avoidance difference, that could be RNG, or they may have changed how the calculation works since I quit playing in SoF.

I spoke extensively with Rashere prior to leaving about ATK debuffs and the way NPCs are design gives them exceptional power provided the NPC atk power has been increasing in relation to armor values (which they started in SoF) This actually lead to situations where Atk Debuffs would shave off as much if not more damage then slow over a equal number of hits. Slow may still shave off more damage then Atk debuffs over a equal duration.

The only way to test this is by doing several tests based on the exact same amount of attacks and several tests based on the exact same time duration, then compare the data. As I played a Druid in EQ, I didn't have access to slow so that part of my testing was never completed.

I don't play any longer, but I set the post up to notify me if you have any more questions. Feel free to ask whatever and I'll attempt to answer you the best I can.

[Kumudil]

Same amount of time would be possible. Exact same amount of attack maybe not so easy.

1000 swings is around 10 minutes. Howlong of a parse would be needed to proove/disproove the assumption that mobs accuracy is affected as well?

[Fenier]

Most parses on the beta servers are calculated on several hours against a training dummy.

You may not have to go that long, but it should at least be the same NPC so you add as many controls to the parse as possible.

[Kumudil]

In arena you can setup your stuff and go afk.

In tosk that does not work that easy.
I cannot stand to herd a tankmerc beating an Ogre in tosk for ours. Refreshing the acutal debuffs every couple of minutes.

The parses done have been with the same identic mob. At the end of one parse I would root the mob and evacc all to ZI and run back to the mob to assure its always the identic one.

If I could stant to keep each parse running for 30 minutes. How would be the statisctical chance that the results are still within the possibilities of just an evel RNG, if the trend shown by the 10 minutes parses remain the same?

[Fenier]

I didn't major in statistics. I have totally no idea.

Truth be told, I -did- sit for hours at one point, I think I did like 1800 hits on a Noc Darkluch in Riwwi. I was that determined to prove my findings at the time. (Fun Trivia: The reason I did the research in the first place was to help me win a debate versus another poster on a thread on TDG. I not only proved that Attack Debuffs worked, but they worked quite well.)

I still have the parse, oh and it includes the accuracy, this may be of help to you.

Noc Darkluch - Without Debuffs:
Attack Adjustment: 0
Number of Hits: 1800
Number of Misses: 1756
Accuracy: 50.62
Damage Delt: 287,557
Average Hit: 159.75
Mob DPS: 81.09

Debuffed: Hand of Ro, Fixation of Ro, Sun's Corana
Attack Adjustment: -255
Number of Hits: 1800
Number of Misses: 1754
Accuracy: 50.65%
Damage Delt: 198,869
Average Hit: 110.48
Mob DPS: 55.99
Note: DI19 (360) never once occurred in 1800 hits during the debuffed parse

Differences:
Average Hit Dropped: 49.27
Damage Dropped: 88,688
Percentage Difference in Damage: 31%
DPS Dropped: 25.1
Percentage Difference in DPS: 31%

The .03% variance is likely just RNG. Even if it wasn't, it would take subtracting several hundred attack to increase avoidance by a fraction of a percentage. While I have been out of EQ since SoF, I really would be surprised to find out Rytan had the coders rework the combat system to that degree. (Fun trivia: EQ was written in C / C++ and was over a million lines of code in SoF)

It's been so long, iirc the stat which affects NPC accuracy is likely Dex. There are examples of this, and the one that comes to mind are the Anueks in Txevu. They had a High DB low DI so ATK Debuffs didn't do much, but they hit consistently around DI16+ with a horridly high accuracy. So while they didn't hit for much, the fact they seldom missed and hit so fast is what would kill tanks.

I hope that has been helpful to you, and if there is only one thing you take away from this post, let it be that Tharkis is still a pretty kitty.

[Samanna]

/wave Feiner!

[Fenier]

Sup Sam

[Tharkis]

Ack a sabbified woof. /sets Fenier's tail on fire

[Swampfunk]

This was a good read. Need more atk debuffs.

[Fenier]

More like need more combined Attack Debuffs so Druids don't have to cast like 4 to 5 different spells!