Jump to content


Photo

Bear Form Threat Generation


  • Please log in to reply
17 replies to this topic

#1 saramin

saramin

    King Hippo

  • Members
  • 591 posts

Posted 23 April 2006 - 08:15 AM

Yes, it's another druid thread. I know your immediate reflex is to spam the first page with sparkling witticisms regarding this glorious fact. I ask only that you be strong.

So recently we've been having some rather interesting conversations about warrior vs. druid aggro buildup. Not in terms of bear form mt/offtank viability, but more in regards to what the formulae are like. There are a lot of contradictory numbers out there. As I understand it, warriors generally have a set constant added to their damage when in defensive stance. This pertains to shield block, etc. This doesn't scale with gear as it's not multiplicative and I've seen several warriors whine long and loud that their MS hits generate more threat than any sword and board equivalent as they get better and sexier gear.

According to armchair philosophy, the bear form threat curve is different in that it uses a straight variable. It takes whatever the 1:1 damage to hate ratio is for all classes by default and multiplies it by X. What this amounts to is that while warriors are better in grabbing immediate aggro following the taunt, druids are supposedly better at consistently working their way to the top of the threat list at a faster rate and staying there. In other words, bear form threat buildup scales perfectly with gear.

In pragmatic terms I'm curious what the stat focus should be. I used to be a proponent of agil over str in bear form as white damage is fairly irrelevant but crits generate additional threat. If the above is true though, straight AP granted by strength may be of greater worth. Especially with HotW. Anyone ran some tests regarding this?

#2 diospadre

diospadre

    Hero of the Horde

  • Guild Members
  • 7,465 posts

Posted 23 April 2006 - 08:43 AM

That entire post seems like you just made it up as you were typing.

#3 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,567 posts

Posted 23 April 2006 - 09:02 AM

From one of my old combat mechanics threads (I might do another one of those sometime soon, btw).

------

The 1.45 that Druids and Warriors have is a multiplicative bonus to all threat generated while in the correct stance/form.
Some abilities (Faerie Fire, Sunder Armor) have a flat threat associated even though they do no damage or healing. Sunder Armor generates
C*0.8
threat in Battle Stance, and
C*1.45
threat in Defensive Stance, where C is a constant.

Some abilities (Heroic Strike, Shield Slam) have a constant added to the damage they do. Shield slam generates
(Damage + C)*1.45
threat, and the C is a constant.

Some abilties (Druids only) have an extra multiplicative bonus. For example, Maul generates
(Damage*1.75*1.45)
threat. The 1.75 is a constant associated with Maul.

This is the reason Druids scale so well with attack power. Druid threat generation, overall can be loosely represented as
(DPS*X)*1.45,
where X is somewhere between 1 and 2. Warrior threat generation looks like
(DPS + C)*1.45,
where C is a constant and tends to be much greater than DPS.

As strength/AP levels get higher, Druids simply get better and better in comparison to Warriors.

#4 Graham

Graham

    Soda Popinski

  • Guild Members
  • 3,833 posts

Posted 23 April 2006 - 03:32 PM

I don't believe anyone has done a quantative analysis of where the various proper break points are for bear and maximizing threat. Hamlet has already done a good job with the basics. Beyond that, as to crit vs AP vs +Damage, any attempt to analize it is problematic because without block, parry, or significant defense you need to maximize armor first and foremost in order to be a viable tank.

I guess if there were an optimal set of numbers, I'd be looking for 14000+AC, 900AP, 25% Crit, 15% Dodge, and 350ish Defense. (Crit is just huge in bear aggro generation, two early crits can seal the mob to you for the whole fight.)

Realistically, I tank Jindo with 15500 AC, 800AP, 15% Crit, 8% Dodge, and about 325 Defense and only 11 points in Feral and have yet to have any aggro problems with BWL+ Geared mages and warlocks in the fight.

#5 Zyla

Zyla

    Ravaging the Art World.

  • Allied Members
  • 15,552 posts

Posted 24 April 2006 - 05:32 PM

Yes, it's another druid thread. I know your immediate reflex is to spam the first page with sparkling witticisms regarding this glorious fact. I ask only that you be strong.

So recently we've been having some rather interesting conversations about warrior vs. druid aggro buildup. Not in terms of bear form mt/offtank viability, but more in regards to what the formulae are like. There are a lot of contradictory numbers out there. As I understand it, warriors generally have a set constant added to their damage when in defensive stance. This pertains to shield block, etc. This doesn't scale with gear as it's not multiplicative and I've seen several warriors whine long and loud that their MS hits generate more threat than any sword and board equivalent as they get better and sexier gear.

According to armchair philosophy, the bear form threat curve is different in that it uses a straight variable. It takes whatever the 1:1 damage to hate ratio is for all classes by default and multiplies it by X. What this amounts to is that while warriors are better in grabbing immediate aggro following the taunt, druids are supposedly better at consistently working their way to the top of the threat list at a faster rate and staying there. In other words, bear form threat buildup scales perfectly with gear.

In pragmatic terms I'm curious what the stat focus should be. I used to be a proponent of agil over str in bear form as white damage is fairly irrelevant but crits generate additional threat. If the above is true though, straight AP granted by strength may be of greater worth. Especially with HotW. Anyone ran some tests regarding this?

undeniably, when i look towards stats for tanking, on a pure aggro basis, I want as much as crit as is humanly possible. AGI owns. You get dodge AND crit, two things that bears love.

arawethion pretty much nailed it...

Bear threat is multiplicative, and the more you can abuse those modifiers, the better.

While tanking I run about 30% crit with proper consumables... I can lead off on vael with this crit % and our dagger rogues can go balls to the wall while i do it. They report more confidence in me then in a warrior tanking that portion.

Mongoose pots are godly for this purpose.

Zyla, International Man of a Certain Standard.

That's right, I met my future wife through Zyla. :shudder:


#6 Husyor

Husyor

    Von Kaiser

  • Members
  • 57 posts

Posted 26 April 2006 - 12:06 AM

Realistically, I tank Jindo with 15500 AC, 800AP, 15% Crit, 8% Dodge, and about 325 Defense and only 11 points in Feral and have yet to have any aggro problems with BWL+ Geared mages and warlocks in the fight.



What gear do you use to get this.

#7 dreadnor

dreadnor

    Von Kaiser

  • Members
  • 68 posts

Posted 27 April 2006 - 02:18 AM

I was under the impression crits dont generate additional aggro.

#8 Graham

Graham

    Soda Popinski

  • Guild Members
  • 3,833 posts

Posted 27 April 2006 - 02:26 AM

I was under the impression crits dont generate additional aggro.

Um. What?

Crits generate double aggro.

#9 dreadnor

dreadnor

    Von Kaiser

  • Members
  • 68 posts

Posted 27 April 2006 - 02:46 AM

All I meant was:

If a hit for 500 generates 500 aggro, then a crit for 1000 generates 1000 aggro. I am by no means an authority on the issue, however Kenco tested this (with a mage's frostbolt) and found it to be true.

#10 Graham

Graham

    Soda Popinski

  • Guild Members
  • 3,833 posts

Posted 27 April 2006 - 03:10 AM

All I meant was:

If a hit for 500 generates 500 aggro, then a crit for 1000 generates 1000 aggro. I am by no means an authority on the issue, however Kenco tested this (with a mage's frostbolt) and found it to be true.

I would love to see the test methodology then.

Sorry, I'm in a snippy mood.

That testing is wrong. Crits generate a substantial aggro bonus, roughtly 1.3x to 1.7x the total damage done.

#11 dreadnor

dreadnor

    Von Kaiser

  • Members
  • 68 posts

Posted 27 April 2006 - 03:23 AM

Here is the post he made in an older thread:

"I just tested this then, and didn't find any bonus threat for crits. The setup was

1) Warrior body pulls mob. 0 threat
2) Mage casts level 5 frost bolt. Crits for 1119 damage. Mage has x threat.
3) Warrior taunts, and keeps aggro. Warrior now has x threat.
4) Mage does autoattack until he gains aggro. It is known he will draw aggro when he exceeds 1.1 x threat. He draws aggro after 115 damage, not 109 damage.

This gives a range for x of 1090 to 1150. As a multiplier, the highest value for critical threat bonus is +2.8%. The value of +0% is about dead in the middle of the error range.

It seems more likely to me that people associate crits with pulling aggro, because you are more likely to pull aggro on a crit than on a normal attack, so they assume there is an extra value for crits. It's also possible that the effect doesn't exist for frost bolts and does for other spells or abilities, but that seems unlikely to me. "

(Posted by Kenco)

I would love to have seen a follow up post and some further testing, however this was the only post he made on that thread.

#12 Hamlet

Hamlet

    Mike Tyson

  • • Guide Author
  • 11,567 posts

Posted 27 April 2006 - 04:45 AM

I would love to see the test methodology then.

Sorry, I'm in a snippy mood.

That testing is wrong. Crits generate a substantial aggro bonus, roughtly 1.3x to 1.7x the total damage done.

Yeah, I've seen the "1.5 aggro" claim from you guys disputed, both by Kenco, and by someone on the Druid forums researching Maul threat (maybe Zyla remembers).

#13 Lord BEEF

Lord BEEF

    Soda Popinski

  • Guild Members
  • 3,826 posts

Posted 27 April 2006 - 08:32 AM

All I meant was:

If a hit for 500 generates 500 aggro, then a crit for 1000 generates 1000 aggro. I am by no means an authority on the issue, however Kenco tested this (with a mage's frostbolt) and found it to be true.

I would love to see the test methodology then.

Sorry, I'm in a snippy mood.

That testing is wrong. Crits generate a substantial aggro bonus, roughtly 1.3x to 1.7x the total damage done.

Not for bear form.

It's likely it's changed for all classes/forms/whatever, it seems all crits have no extra bonus threat
Check out my friend's bitchin' Lord of the Rings Art

#14 Graham

Graham

    Soda Popinski

  • Guild Members
  • 3,833 posts

Posted 27 April 2006 - 04:00 PM

Well then, good thing I rushed in to look like a moran.

(When did that change >:( )

#15 Praetorian

Praetorian

    Mike Tyson

  • ♦ Administrators
  • 27,801 posts

Posted 27 April 2006 - 04:15 PM

Unless we were completely asleep at the wheel, back in the early BWL days when we did significant aggro testing on different abilities and such, two casts for 500 each and one crit for 1000 required quite different amounts of autoattack damage to outaggro. It certainly seemed like there was a built-in modifier at the time, but I'd be lying if I said I'd tested it empirically at any time in the last 8 months.

#16 Oaken

Oaken

    Don Flamenco

  • Members
  • 416 posts

Posted 28 April 2006 - 12:30 AM

This is the post I saved from the Druid Forum where other methodical testing was done. Testing by Chibi.

---------------

Kenco on the warrior forum posted some nice numbers for warrior threat. So naturally this causes me to wonder how druids compare. I tried sifting through tons of posts of mostly anecdotal evidence, but I was unsatisfied and decided to do the testing myself. So here the data I gathered on Maul. Swipe will be incoming, but this was boring so it might be a while. :)

The tests were preformed with me (Chibi, a warrior) using Recap, a Warlock (Impudence), and a Druid with "defiance" (Alys).

The druid initialized combat with a single maul in bear form. Enrage was not used, since I felt it could cause threat. Then the warlock would punch the mob (four damage per swing generally) until he pulled aggro. This resulted in the following data.

NORMAL MAUL

# MAUL PULLED DIDN'T THREAT THREAT BASE   BASE   MAX   MIN   MAX %  MIN %
        DMG    PULL   MAX    MIN    THREAT THREAT DIFF. DIFF. OF DMG OF DMG
                                   MAX    MIN

1  285 796    792    723.6  720.0  499.0  496.6  214   211.6 75.1   74.2
2  292 818    814    743.6  740.0  512.8  510.3  220.8 218.3 75.6   74.8 
3  263 735    731    668.2  664.5  460.8  458.3  197.8 195.3 75.2   74.3
4  270 755    751    686.4  682.7  473.4  470.8  203.4 200.8 75.3   74.4
5  288 806    802    732.7  729.1  505.3  502.8  217.3 214.8 75.5   74.4
6  286 801    797    728.2  724.5  502.2  499.7  216.2 213.7 75.6   74.7
7  277 774    770    703.6  700.0  485.2  482.8  208.2 205.8 75.2   74.3
8  268 749    745    680.9  677.3  469.6  467.1  201.6 199.1 75.2   74.3
9  266 744    740    676.4  672.7  466.5  463.9  200.5 197.9 75.4   74.4
10 260 726    722    660.0  656.4  455.2  452.7  195.2 192.7 75.1   74.1

Threat min and max are determined by dividing damage by 1.1. 10% is the melee range stickiness, that is, the amount needed beyond the current aggro holder's threat in order to pull aggro. This was determined by various sources (including myself) in other tests.

Base threat is determined by dividing threat totals by 1.45. The 1.45 threat modifier in bear form was crudely tested beforehand, but mostly this number comes from faith in thottbot.com and the assumption that bear form works just like defensive stance for warriors.

DIFF. means difference which is determined by subtraction (duh!).

Smallest max difference: 195.2
Largest min difference: 218.3

From this impossible range we can conclude maul does not have a fixed modifier as heroic strike (similar warrior skill), but notice the beautiful final columns.

Smallest max % of damage: 75.1%
Largest min % of damage: 74.8%

I feel fairly confident saying maul provides a 1.75 threat modifier to its damage from this data.

To verify these findings and investigate the possibility of strange crit behaviour we tested a few crits using the same methods.

CRIT MAUL

# MAUL PULLED DIDN'T THREAT THREAT BASE   BASE   MAX   MIN   MAX %  MIN % 
        DMG    PULL   MAX    MIN    THREAT THREAT DIFF. DIFF. OF DMG OF DMG
                                   MAX    MIN

1 594  1659  1651    1508.2 1500.9 1040.1 1035.1 446.1 441.1 75.1   74.3
2 575  1610  1602    1463.6 1456.4 1009.4 1004.4 434.4 429.4 75.5   74.7
3 596  1665  1661    1513.6 1510.0 1043.9 1041.4 447.9 445.4 75.2   74.7
4 587  1640  1636    1490.9 1487.3 1028.2 1025.7 441.2 438.7 75.2   74.7

Crits apparently function the same.

So threat from maul is:

Maul_Damage * 1.75 * 1.3 (or 1.45 with talent)

And

We went back and tested swipe.

The data is broken into two tables so that it won't be too wide.

#  TOTAL WHITE SWIPE PULLED DIDN'T 
    DMG   DMG   DMG          PULL 

1  220   153   67    432    428
2  369   302   67    670    662
3  214   146   68    424    421
4  222   154   68    436    433
5  219   146   73    439    435
6  218   151   67    430    426 
7  205   137   68    412    405 
8  218   151   67    432    424 
9  238   164   74    470    466 
10 211   144   67    418    414 
11 252   180   72    491    487
12  67     0   67    188    184
13  67     0   67    191    187
14 284   145  139    627    623
15 302   158  144    654    650
16 293   147  146    643    639
17 275   140  135    604    600

# THREAT THREAT WHITE SWIPE SWIPE BASE  BASE MAX  MIN  MAX % MIN %
  MAX  MIN  THREAT THREAT THREAT SWIPE SWIPE DIFF. DIFF. OF DMG OF DMG
             MAX  MIN  THREAT THREAT
                    MAX  MIN

1 392.7 389.1 221.9 170.8 167.2 117.8 115.3 50.8 48.3 75.8 72.1
2 609.1 601.8 437.9 171.2 163.9 118.1 113.0 51.1 46.0 76.3 68.7
3 385.5 382.7 211.7 173.8 171.0 119.9 117.9 51.9 49.9 76.3 73.4 
4 396.4 393.6 223.3 173.1 170.3 119.4 117.4 51.4 49.4 75.6 72.6 
5 399.1 395.5 211.7 187.4 183.8 129.2 126.8 56.2 53.8 77.0 73.7 
6 390.9 387.3 219.0 171.9 168.3 118.6 116.1 51.6 49.1 77.0 73.3 
7 374.5 368.2 198.7 175.8 169.5 121.2 116.9 53.2 48.9 78.2 71.9 
8 392.7 385.5 219.0 173.7 166.5 119.8 114.8 52.8 47.8 78.8 71.3
9 427.3 423.6 237.8 189.5 185.8 130.7 128.1 56.7 54.1 76.6 73.1 
10 380.0 376.4 208.8 171.2 167.6 118.1 115.6 51.1 48.6 76.3 72.5 
11 446.4 442.7 261.0 185.4 181.7 127.9 125.3 55.9 53.3 77.6 74.0 
12 170.9 167.3  0  170.9 167.3 117.9 115.4 50.9 48.4 76.0 72.2
13 173.6 170.0  0  173.6 170.0 119.7 117.2 52.7 50.2 78.7 74.9 
14 570.0 556.4 211.7 358.3 344.7 247.1 237.7 108.1 98.7 77.8 71.0 
15 594.5 590.9 229.1 365.4 361.8 252.0 249.5 108.0 105.5 75.0 73.3 
16 584.5 580.9 213.2 371.3 367.7 256.1 253.6 110.1 107.6 75.4 73.7
17 549.1 545.1 203.0 346.1 342.1 238.7 235.9 103.7 100.9 76.8 74.7 


Once again the smallest maximum difference is 50.8 and the largest minimum difference is 107.6, which creates an impossible range. So swipe appears to be affected by a threat multiplier like maul.

The range for the swipe multiplier is 1.75 to 1.749. 1.75 might be a good guess :) This is the exact same modifier that maul has, and since swipe and maul are your only damage causing abilities (as far as I know), we could generally say bear damage causing abilities cause damage * 1.75 threat.

So swipe's threat, like maul is calculated thus:

Swipe_Damage * 1.75 * 1.3 (or 1.45 with talent)



We also did some quick fairy fire and demo shout tests. The warlock was wimping out cause he wanted to go to sleep, so there isn't much data. Sorry. Here are the results.

Demo shout was used once. The warlock pulled with 67 damage. 63 damage did not pull.

67 / 1.1 / 1.45 = 42.0
63 / 1.1 / 1.45 = 39.5

Faery fire at one application (pulled_number / 1.1 / 1.45):
1) 180 pulled aggro. 172 did not. 112.9 max base threat. 107.8 min base threat.
2) 174 pulled aggro. 170 did not. 109.1 max base threat. 106.6 min base threat.

Faery fire at two applications (pulled_number / 1.1 / 1.45 / 2):
3) 345 pulled aggro. 341 did not. 108.2 max base threat. 106.9 min base threat.

Faery fire at four applications (pulled_number / 1.1 / 1.45 / 4):
4) 693 pulled aggro. 689 did not. 108.6 max base threat. 108.0 min base threat.


Thus, before bearform modifier:
Fairy Fire: 108.0-108.2
Demo shout: 39.5-42

#17 Oaken

Oaken

    Don Flamenco

  • Members
  • 416 posts

Posted 28 April 2006 - 01:31 AM

As strength/AP levels get higher, Druids simply get better and better in comparison to Warriors.

True. But keep in mind that as mob armor (i.e., physical damage mitigation) increases, Druids suffer a commensurate decrease in aggro as well. Just one more of the differences between Druid and Warrior capabilities.

#18 Xaviar

Xaviar

    Von Kaiser

  • Members
  • 25 posts

Posted 28 April 2006 - 06:55 AM

The thing is mob armour doesn't get higher and higher. A melee ogre is Dire maul probably has the same armour value as Nefarion.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users