Nostalrius Begins - Quality wow vanilla realm (1.12)

Test Apparatus

Zetox - Level 60 warrior. No shield specialization, 4/8 might, no nagel ring.

Testikill (Armilus) - Level 60 priest. Witch blade and Worn Dagger that we stole from some level 1 rogue in IF. Recap damage meter.

Ellaria - Testing assistant.

Test method

Testikill clears Recap damage meter and starts recording events.

Zetox walks up to a mob (test subject) to get into combat and stands there without attacking until he has enough rage to use the test ability. Zetox then uses the test ability on the test subject once and stops attacking.

Testikill waits for Zetox to use the test ability and then starts attacking with witchblade until close to the expected agro threshold. Testikill then switches to worn dagger and hits for 2-4 damage until pulling agro.

Once Testikill pulls agro he turns his back on the mob to stop attacks immediatly and pauses the damage meter. Test subjects are terminated.

Total damage applied by Testikill until the agro pull is recorded and the damage of the last hit is recorded in order to calculate margin of error.

Method of analysis

Threat is calculated with the following expected formula according to http://www.wowwiki.com/Kenco%27s_research_on_threat

([damage] + [bonus threat of ability]) * [defence stance bonus] * [melee agro threshold]

Known values are entered into this formula and we solve for the unknowns to determine bonus threat value of various warrior abilities.

Assumptions

For these tests we assume a melee agro threshold of 1.1 (melee should not pull agro until they are 10% over the current targets threat). We also assume that the defense stance multiplier is 1.3 without defiance and 1.45 with 5/5 defiance talent.

Notes

Some tests with a margin of error > 2 were discarded.


Test 1 - Heroic strike, defense stance no defianse

Zetox - 1 Heroic strike rank 9 for 216 damgae

Testikill - pulled agro after dealing 562 damage

Expected: Unknown, Kenco's threat research was conducted with Rank 8 Heroic strike.
Result: Rank 9 Heroic strike has +176 bonus threat +/- 1


Test 2 - Auto attack

Zetox - 1 auto attack for 94 damage in defensive stance, no talents

Testikill - After dealing 132 damage did not pull agro. After dealing 135 damage pulled agro.

Expected: Should pull agro at 135 damage
Result: Pulled agro at 134 +/- 1 damage


Test 4 - Sunder Armor

Zetox - 1 sunder armor in defensive stance

should pull agro at 373.23 damage

Testikill - After dealing 370 damage, did not pull agro. After dealing 374 damage, pulled agro.

Expected: Sunder armor has ~261 bonus threat
Result: Sunder armor has 260 +/- 2 bonus threat


Test 6 - Shield Slam

Zetox - 1 shield slam for 334 damage

Testikill - After dealing 834 damage, did not pull agro. After dealing 837 damage, pulled agro.

Expected: Shield slam has ~250 bonus threat
Result: Shield slam has 250 bonus threat +/- 1


Test 7 - Shield bash, def stance no defiance

Zetox - 1 shield bash for 38 damage

Testikill - After dealing 275 damage, did not pull agro. After dealing 324 damage, pulled agro.

Expected: Shield bash has 180 bonus threat.
Result: Shield bash has 188 +/- 25 bonus threat. Large margin of error but within expected result. Didn't bother testing again to reduce margin of error.



Test 8 - Battleshout, def stance no defiance, 3 mobs in combat

Zetox - Buffed self with battleshout, no other party members buffed.

Testikill - After dealing 31 damage did not pull agro. After dealing 33 damage pulled agro.

Expected: Unsure
Result: Battleshout caused 22 +/- 1 threat to all targerts.


Test 9 - Battleshout, def stance no defiance, 1 mob in combat

Zetox - Buffed self and testikill with battleshout, no other party members buffed.

Testikill - After dealing 170 damage did not pull agro. After dealing 173 damage, pulled agro.

Expected: Unsure
Restult: Battleshout affecting 2 targets caused 119.9 +/- 1 threat to a single target.


Test 10 - Battleshout, def stance no defiance, 1 mob in combat

Zetox - Buffed self, Testikill and Ellaria, no other party members buffed.

Testikill - After dealing 239 damage no pull, after 261 pulled agro

Expected: Unsure
Result: Battleshout affecting 3 targets caused 174.8 +/- 11 threat to a single target.


Test 11 - Battleshout, def stance no defiance, 1 mob in combat

Zetox - Buffed self, testikill and Ellaria, no other party members buffed.

Testikill - After dealing 256 damage, did not pull agro. After dealing 262 damage pulled agro.

Expected: Unsure
Result - Battleshout affecting 3 targets caused 181 +/- 3 threat to a single target.


Test 12 - Demo shout, def stance no defiance, 1 target hit

Zetox - Applied demo shout

Testikill - After dealing 61 damage did not pull agro, after dealing 64 damage pulled agro.

Expected: Unsure
Result: Demo shout caused 43 +/- 1 threat


Test 13 - Demo shout, def stance no defiance, 2 targets hit

Zetox - Applied demo shout to 2 targets

Testikill - After dealing 30 damage did not pull agro, after dealing 32 damage pulled agro.

Expected: Unsure
Result: Demo shout caused 21 +/- 1 threat to each target debuffed


Test 14 - 1 Auto attack, def stance with defiance

Zetox - Hit target for 229 damage

Testikill - After dealing 376 damage did not pull agro. After dealing 380 damage pulled agro.

Expected: Should pull agro at 366 damage. 229 damage * 1.45 * 1.1.
Result: Pulled agro at 377-380 damage which matches the following formula: 229 * 1.3 * 1.15 * 1.1
Rather than adding to the def stance bonus, defiance seems to be a 3rd threat multiplier.


Conclusions

Battleshout is causing ~60 threat per target buffed and the threat is divided among all targets in combat. NEED VERIFICATION THAT THIS IS CORRECT!

Demo shout is causing ~42 threat divided among all targets. It is unknown if the threat is divided only among targets hit by the debuff or if it is divided among all targets in combat. Likely the former. NEED VERIFICATION THAT THIS IS CORRECT!

The defiance talent is adding an additional 15% increase to threat applied AFTER defensive stance instead of modifying the defensive stance bonus to be 45% threat increase (up from 30%). The end result is equivalent to a defensive stance threat bonus of 49.5% instead of 45%.

Edit: This indeed seems to be the correct behaviour in 1.12, it was likely changed to be multiplicative at this time http://www.wowwiki.com/Patch_1.12.0

Thunderfury proc test

In this test we modify the threat formula to account for the bug in the defiance talent.

([damage] + [bonus threat of ability]) * [Defense stance modifer] * [Defiance modifier] * [Melee agro threshold]
Note that this is not the correct formula but it seems to be what is used on Nost during the Stress test.

Zetox used a single auto attack on a target and got a TF proc. 157 damage from auto attack, 270 damage from TF proc.

Priest attacking the mob pulled agro at 709 damage, last hit was for 40 damage.

Expected
Reference: http://forums.elitistjerks.com/topic/69 ... eneration/
Thunderfury proc applies 2 debuffs to the primary target. The slow effect should cause 92 threat while the nature debuff that hits multiple targets should cause ~149 threat to each target hit.

Base threat from the attack: 157 damage + 270 proc damage + 92 attack speed debuff + 149 NR debuff = 668 base threat.

After applying threat modifiers threat should be pulled at 1099 damage.

Result
Priest pulled agro at 709 damage with a +/- 20 margin of error.

Analysis
If the bonus threat from the debuffs is ignored we have a base threat of 427 and expected to pull agro at 702 damage.

This fits nicely within the +/- 20 margin of error.

Conclusion

It seems that the Thunderfury debuffs do NOT cause threat.

Edit: Zetox performed an additional test. A secondary target was affected by the NR debuff, agro was then pulled after 4 damage. NR debuff does not seem to be causing threat on secondary targets either.

Additional tests required:

Revenge bonus threat.

Additional Battleshout tests (buff 1 party member with 1 target in combat, buff 4 party members with 1 target, buff 5 party members with 1 target, additional testing of threat being divided among targets)

Demoshout tests (See if threat is divided among targets in combat but NOT hit by the debuff)

Test if rage bonuses from shield spec or 5/8 might bonus cause threat.

Bloodrage threat.

Impressive testing guys

Neat tests and once again thorough work by Armilus, however I must inform you that all addons ingame are watching threat based on their own calculations, they do not reflect the servers' calcs.

Hence why none of it was done with Addons regarding threat values, but based purely off "in-game dmg".

Netherfrost wrote:Neat tests and once again thorough work by Armilus, however I must inform you that all addons ingame are watching threat based on their own calculations, they do not reflect the servers' calcs.

Yes, we did not use any in-game threat addons. The only addon used was Recap because of it's ability to pause data collection manually. The same data could have been collected by adding up damage in the combat log but that is quite tedious. Some tests required 50-60+ hits at 2-4 damage before agro was pulled

If you can think of anything else that may have affected the results please let us know.

Well that is the unfortunate side of private servers, aside from the GMs datalog of threat, you have no way to track it at all, assuming the combat log is correct is trusting unknown calculations, rendering your test incomplete. And as a note, recap is off, same goes for KTM (naturally, even the remade ones). Additionally, keep in mind that the (supposed to) value of overaggro below 10 yards is 110%(above, 130), threat reduction is also added together multiplicatively. Keep it up

Netherfrost wrote:Well that is the unfortunate side of private servers, aside from the GMs datalog of threat, you have no way to track it at all, assuming the combat log is correct is trusting unknown calculations, rendering your test incomplete. And as a note, recap is off, same goes for KTM (naturally, even the remade ones). Additionally, keep in mind that the (supposed to) value of overaggro below 10 yards is 110%(above, 130), threat reduction is also added together multiplicatively. Keep it up

Yup, as listed in the test method, this was taken into account.

Right, I skimmed it and couldnt spot it, searched for the 110% rather than 1.1 which makes even more sense.. ^^. If you're looking for sources, check the githubs for other live private servers, they usually have the work done for you already