Status: Unknown

Description: Level-based resistances can not be countered by spell penetration for non-binary spells.

Detailed Information:

Since the introduction of resistance cap in 1.9 (making it impossible to reduce target's resistance below zero) it is no longer possible to counter level-based resistances.

Level-based resistances are not related with spell miss. The term level-based resistance describes the effect when resistance of a target is modified by level difference. This was shown to exist when this bug was first discovered in 1.9 and was posted here on these forums. Further information that supports this can be found on the following pages:

http://forums.subcreation.net/viewtopic.php?t=482
http://forums-en.wow-europe.com/thread.aspx?FN=wow-mage-en&T=132218&P=1
http://www.curse-gaming.com/en/wow/addons-4169-glock-mob-magical-defense-analysis.html

Experiments described there suggest that the resistance score of a target is modified by 8 per level difference. This would make a boss with effective level of 63 to have a resistance of 24 to all schools of magic even if it has no resistances by itself.

This can be further confirmed by using Beast Lore on targets of level 61+, which shows for example that Core Hounds have no arcane resistances, but a caster will still experience partial resists on them when no penetration is used.

The fact that shows that level-based resistances are not affected by spell penetration is the following. Bosses tested had the same level of partial resists before and after 1.9 when no spell penetration was used. However before 1.9 it was possible to get vulnerabilities on them when spell penetration was used while now even when using penetration there is the same rate of partial resists as without it. There is specific parsed data for Onyxia from before and after 1.9 that shows this.

The reason why I believe this is not simply a feature but a bug is because binary spells are not affected by it. When no penetration is used the level of resists corresponds to the product of level-based miss rate and here described level-based resistances. However when penetration is used it is possible to counter the resistance part and achieve hit rate expected from miss rate alone. (it might be the case that binary spells don't get level-based resistances applied in the first place, I lack data to confirm this)

This gives a 6% penalty in damage when fighting bosses for arcane and fire mages as well as warlocks that is not present for frost mages.

EDIT: this is a response to someone asking for more data. They removed the post but I'll leave the answer since others might have similar questions.

If you would check the links above you would see that tests consist of several thousand samples.

I can't give you screenshots but here's some parses accumulated over 1.11 (I reset the collection between patches).

We're not doing Molten Core much anymore but I have some samples for Core Hounds since you're asking. This data is for Arcane Missiles against level 62 Ancient Core Hound with no penetration gear used.

5% to hit, no penetration
159 samples
Full resist: 0.0%
75% resist: 0.0%
50% resist: 2.5%
25% resist: 6.9%
No resist: 90.6%

Average partial resist: 3% (expected 4% from 16 level based resistance)

5% to hit, Curse of Shadow (75 penetration)
475 samples
Full resist: 1.7%
75% resist: 0.0%
50% resist: 3.2%
25% resist: 8.8%
No resist: 86.3%

Average partial resist: 3.9% (expected 0% from 16 level based resistance and countered by penetration)

I fully understand the above data is not sufficient for statistical significance, but it still shows you that partial resists exist even when you would expect them to be negated by penetration.

To show you on an example with more samples that the trend is evident in other places here is data for Instructor Razuvious.

6% to hit, Curse of Shadow (75 penetration)
2052 samples
Full resist: 1.0%
75% resist: 0.8%
50% resist: 4.8%
25% resist: 11.9%
No resist: 81.5%

Average partial resist: 6.0% (6% from 24 level based resistance, not countered by penetration)

If you would like to see more data please check the above links.

I just found a way to retrieve the old forums. Here is the original bug report from 2/21/2006.

Current Status: Blizzard is not aware of the bug.

Bug Description: Non-binary spells get partially resisted by mobs of level higher than 60 even when resistances on the mob are brought down to 0 by -resist mechanics (Curse of Shadows, Curse of Elements and items with "Decreases the magical resistances of your spell targets by X" mechanics). This has been confirmed for Arcane Missiles, Fireball and Shadowbolt. This problem is not existent against level 60 and lower targets (either PVE or PVP).

Experiments:

Before 1.9 Onyxia had vulnerabilities when casting Arcane Missiles and COS was applied. This means that she had less than 75 arcane resistance. With patch 1.9 missiles are getting partial resists at the same rate as in 1.8 without COS. The same holds for most MC trash mobs and bosses.

When casting Arcane Missiles at Druid with 48 arcane resists almost every second missile was partially resisted with no -resist gear. With -50 resist gear there were no partial resists.

Data collected on Ebonroc. In both cases there was Curse of Shadows on him. The tables show in percentages how much damage hit the target (0 = full resist, 0.25 means 75% damage was resisted and so on)

1. case:

-0 resist
7% to hit
658 samples

0.00 : 0.0%
0.25 : 1.0%
0.50 : 3.3%
0.75 : 12.5%
1.00 : 83.1%

2. case:

-40 resist
5% to hit
849 samples

0.00 : 2.0%
0.25 : 0.6%
0.50 : 4.2%
0.75 : 10.2%
1.00 : 82.8%

Comparing both cases first one can see that +to hit gear is performing as expected (+7% to hit and arcane focus results in no misses, +5% to hit results in 2% of white resists).

The partial resists in both cases however are the same, worth around 30 arcane resist. Distribution of partial resists is the same up to the small difference due to randomness. -40 resist gear should result in substantial drop in partial resists if not eliminate them completely.

Additional information:

http://forums-en.wow-europe.com/thread.aspx?FN=wow-mage-en&T=132218&P=1

http://forums.subcreation.net/viewtopic.php?t=482

EDIT:

Note that the test were done before 1.11. Post 1.11 Arcane Focus no longer guarantess 100% hit.

Further experiments detailed on the above link have confirmed that indeed the resistances that we are encountering here are level based, 8 resist per level difference, equaling 2% damage mitigation.

Based on the information understood about resist calculations it would seem that the change in code that prevents resistances to go below zero was placed in wrong place, before the level based resistances are added. I'm concluding this is a bug, because binary spells which use a separate calculation do not exhibit this problem (i.e. penetration can counter level based resistances for binary spells).

Bumpage

I am a frost mage, while I do not have to deal with partial resists...I do get yellow full resists.
In a raid setting before the vulnerability change I never saw yellow resists.. only white resists. nNow I am seeing too many yellow resist considering the -100 resistance on bosses n my current gear.

---

Q u o t e:
I am a frost mage, while I do not have to deal with partial resists...I do get yellow full resists.
In a raid setting before the vulnerability change I never saw yellow resists.. only white resists. nNow I am seeing too many yellow resist considering the -100 resistance on bosses n my current gear.

---

Blizzard changed the color of resist messages, now they are all yellow. There is no way for frostbolt to tell wheter the resist is a miss or result of resistances, you can only infer it by comparing expected rate of resists based on your +to hit and level of target.

Very interesting.

Are you using a special mod to collect this data? Last time I used GLOCK it didn't save data for different gear configurations (different penetration, different hit %). CritRate? I can't find a download anywhere.

Thanks, and keep up the good work.

Yes I'm using a custom made mod for this. It tracks resist distribution for each combination of +to hit, -resist (including buffs and debuffs) for each mob.

You can't find the download because I never published it. Main reason is that I don't have plans to maintain it beyond making it work for me, plus it has several dependencies. But if there is interest I can post a link for download if someone wants to give it a try.

/bump

Bump. There are more of us that are interested in getting a Blizz response to this, especially with all the hard work by the community to figure out the details of resistances.

In general it seems a little odd that today not a single boss encounter requires penetration gear. Do such bosses appear in TBC? (I'm not familiar with all the boss fights today, we're not that deep into Naxx yet, so if somebody is familiar with a current encounter that benefits from penetration gear plz correct me.)

I polished the mod I'm using for tracking resistances and posted it on curse. You can get it here:

http://www.curse-gaming.com/en/wow/addons-5685-1-resisttracker.html

A simple intended/bug reply would work. An explanation of why there is this discrepancy between binary and non-binary spells would be even better. I just can't believe that we couldn't get a simple reply in as long as 8 months.

This issue still exists in bc ^^

The problem is still here and it looks like it's here to stay. If anyone is interested the latest round of testing shows that level-based partial resists are not 8 per level difference, but 2% per level difference. So at level 70 bosses will act as if they have 28 resist.

What in the gods' names caused you to dig up this old thread? More to the point, what makes any of you think that level-based resistances are not working as intended? Spell penetration is mainly a PvP stat; it's worthless in PvE except for very specific encounters.