Somehow I feel the spiritual father of this phenomenal discovery...

Many of you seem to adhere to EP bugging out and giving 3% hit in excess. Granted, it's more plausible than the malarkey with the 5% boss resilience, but if we're registering so low miss rates, and the EP situation is indeed the case, why are we not seeing this 5% anywhere? Last I read about it blizz mentioned this 5% is "impossible to mitigate in any way".

Could this herald the return of the Frostmage? Oh where art thow, long gone days of 31.0.20 raiding spec!

Plankel: fantastic statistical analysis. What you mean to say, (because you lost me somewhere) is that the chance that your theoretical hit rate and observed hit rate match is infact 0.lol%, yes?

The 5% boss resistance only happens with non-binary spells, and only happens with partial resists.

I don't think Blizzard has ever acknowledged it exists, unless you count the spell penetration that keeps popping up on PvE gear.

The extra hit on Frostbolts has already been discovered some time ago but never really looked into further because it wouldn't make it competitive anyway.
The talents that may cause it are Elemental Precision, Winter's Chill or Empowered Frostbolt. That's as far as the discussion went at that time. With the removal of the damage tax it might indeed make a difference.

I've been full frost until I got my T5 set, never went any higher than ~120 hitrating because WWS' never showed any higher than 1.5% resist rate.

That's one of the main things that kept me for a long time not believing in the 5% partials vs bosses, and the fact that I saw far to many conflicting reports.

I have been working the numbers for partials vs bosses and the varying spell penetration numbers, but I put it on hold a few months ago to not enough raiding mages not have more than 50 spell penetration, and 93% of all the data I gathered being on 30 spell penetration.

With myself now having enough spell penetration, I can keep easier track of the numbers, but I fear that it will lead to the same result as with the data I already have: To much variance in the partial % seen.

Would be really nice if Blizzard would give us the exact number for partials, and the same for melee with exact glancing blow % and damage reduction.

I may be mistaken but two months ago i recall either Aeus or Vaneras specifying the 5% partial on bosses to be unmitigatable by any circumstances. Clearly this is not the case for binaries and we all know how notorious blizz can be with the accuracy of their posts, but I was under the impression penetration would have approx. 0 net effect on them.

Hm, I read quite a few mage threads and never heard about it that before now.

I remember that in early beta, Empowered Frostbolt added +5% hit - which was scrapped in favour of 5% crit as mages had high hit anway. Maybe they somehow left that in?

From people who know their gear and log their fights - does the +5% crit on empowered frostbolt work correctly? Or ar there indications that it's broken, and thus might still add 5% hit?

Yea I've seen that information thrown around a bunch, but the problem is that the amount of work they have done to come up with that number is very small. Compared to the information I've gathered up, their numbers are so small. But like I said the numbers I've gathered consist of 93% of them falling on the 30 spell penetration value, the other 7% is numbers of less than 30, or greater than 30, hardly an acceptable gathering to confirm anything. The reason I haven't asked for any public help with this is the chance for tainting the results with the same data multiple times causing the numbers to go bad.

I do wonder what other mage talents have hidden bugs to them if the Elemental Precision is bugged for 6% hit for frost too.

What I really want to know now is whether or not fire has the lower hit cap as well. From a coding standpoint this would almost certainly be the case if EP was bugged.

That brings up another thing...

If the 5% partial resists is actually true, that is only possible with the bosses still having resist left up, and then since spells can get a full spell resist from resist alone, are the full resists fire sees from the resist or the level, harder to tell...

Nearly impossible to test for fire if such would be true.

I don't know enough about actual resist mechanics, so someone can help me calculate what the fire full resist rate should be with this information. If using 121 hit rating and EP is NOT bugged for fire, you'd have first a 4% base miss chance. Now, the additional chance to miss would be based on 24 fire resist on the target. What is the chance for a FULL resist against 24 fire resistance? Tack that on and we have our expected value, and tests can be made to confirm/deny. Unfortunately it's really unusual for fire mages to run without fully hit capping.

To my knowledge, Blizzard has never been so forthcoming as to reveal what the algorithm is that generates the probabilities for the resistance table (i.e. the probabilities of full hit, 75% hit, 50% hit, 25% hit, and full resist due solely to resistance). The average damage reduction is well known, but the actual probabilities of each outcome have eluded me. I once combed their page trying to figure out some pattern (assuming, for kicks, that the expected value of each case might be the same, just on a hunch) but to no avail.

Well let's assume for a second that bosses need 350 resist to mitigate 75% of the damage incoming to them, for the 3 level difference and such. That would mean 467 resist would theoretically provide 100% mitigation. So now if the 24 fire resist is the true average of boss innate resist level, 24/466 = .0514, or 5.14%, this alone would confirm the 5% partial resist rate, but thats not the end of it. That 5% means that 5% of all incoming spell damage gets mitigated, through partials or full resists.

So here comes the problem, if we get 1 resist on a fight, no partial resists, how can we tell if the resist was based off the level or the resist. Is it possible to tell?

Now if I didn't confuse myself just now, didn't I just debunk the 5% partial resists during a fight? Because it's 5% mitigation through full/partial resists, and if it's through a full resist, our thought tells us thats a level based resist. It should be stated that "Fire suffers a 5% damage mitigation through a bosses resist." Now about breaking the 5% damage suffering through spell penetration, still more testing needs to be done there?

It seems we can't fully answer this question till we are able to figure out how to tell the difference between a level base resist and a resist based resist. Fuck.

After doing some digging on thottbot I've noticed something interesting about the mage version of Elemental Precision as compared to similar talent for other classes.

Here's the Rank 3 mage version: Elemental Precision
Reduces the mana cost and chance targets resist your Frost and Fire spells by 3%.
Apply Aura: Add Flat Modifier (Resist/Miss Chance)
Value: 3

Now check out the elemental shaman version: Elemental Precision
Increases your chance to hit with Fire, Frost and Nature spells by 6% and reduces the threat caused by Fire, Frost and Nature spells by 10%.
Apply Aura: Increases Spell % To Hit (Fire, Nature, Frost)
Value: 6

The aura given by those two talents is clearly different yet the perceived bonus is supposedly the same: a +hit% bonus for the specified spell types.

This got me to thinking, perhaps the mage version of elemental precision is both giving a +hit bonus and reducing the resist rate due to resistances. This "double-dipping", while likely unintended or a bug, would account for the ~6% hit instead of 3% for frost.


Eg. Assuming 96% hit before elemental precision against a monster with 5% average damage reduction due to resistances.

Without elemental precision
Chance to hit with frost = 0.96 * (1-0.05) = 0.912

With elemental precision
Chance to hit with frost = (0.96+0.03) * (1-(0.05-0.03)) = 0.9702

Total observed gain in +hit% = 0.9702 - 0.912 = 0.0582 or ~5.8%

People always throw this 5% around, but I've never seen any tests that would actually claim to have come to this conclusion based on controlled data collection. All I've seen indicates it to be 6%. Can someone show me a link to a study supporting the 5% theory?

If a higher level mob is attacking you, you need higher stats to counter his attacks. But if you're attacking a higher level mob, you need higher to break through his stats. But for the higher level mob he does not need higher stats to counter your attacks. So your first argument is not valid as it uses the wrong logic.

And for your second one, thats for melee attacks which follow the Miss/Hit/Partial resist/Crit/Be dodge/parried stuff, it is not the same for casted spells, so that one has no meaning for what we want. That is because mob based spells can either Hit/Resist and in some cases Partial resist.

What I just posted is that bosses have a 5% chance to mitigate incoming spell damage through their resist, which in turn for fire is a partial/full resist, and for frost is a straight up full resist, and for a full resist in either case we cannot tell if its a level based or resist based resist. And since both suffer a 1% innate chance to be resisted through the level, assuming hit cap, Fire would suffer an additional 5% damage mitigation chance.

That is, idealy if all damage done + 1 fireball= 100% damage done, and there are 2 resists, we could assume 1 from level based, 1 from resist based, but the problem occurs at 1 resist for a fight.

No you're misunderstanding there.

Of all damage incoming to a boss, 5.14% is mitigated through partial/full. Of all spells CAST, 1% of them will be a full resist. There is a difference between damage done, and spells cast.

Of course this is going to take a lot longer to elaborate on, but if you can't differ between spells cast, and the damage done being different roles then you need to do some research as I am unable to assist you with that learning because I can't explain it any clearer.

Edit:

I just figured out how to explain this to you.
Ok, so you cast 200 spells, now lets say the average hit is going to be 3000 damage, and your crit rate is 40%, assuming fire spec. Now 2 of those spell are going to resist due to level resist, 6000-9000 damage lost.
So now with the 40% crit/60% hit of those 198 spells, thats 119 hit, 79 crit.
119*3000=357,000
79*4500=355,500
Total of 712,500 damage done. Now assuming of that damage 5.14% mitigated, 18,349.8 damage mitigated, That leaves us at 694,150.2 damage done.

Since the damage mitgated seems to be a nice 2 crit(4500), 3 hit(9000) oriented, getting 7 full resists total, thats 3.5% equated missed chance, not 1% missed, 2.5% resisted like it should be, and thats where the problem is.

I don't have a problem. We're just talking about different things, for you a missed spell doesn't count as damage incoming to a boss, to me it does. But this is beside the point. What I was questioning was the value of 5%. Now that I see that what you really mean is 5.14%, I see that you're refering to the 24 innate resistance that was established pre BC. What I'm questioning is your use of this value at level 70 when tests indicate it to be 28 instead.

We're trying to ascertain the individual effects of resistance and level-based resist chance. It makes no sense to combine the two and look at the damage reduction in this context. We're trying to separate the two.

You're trying to separate the two, but since there is no way to distinguish resist based miss and level based miss in combat log you have to take both into account when doing resist analysis. To calculate level-based resist mitigation you would compute total mitigation and then offset it by known level miss. Either that or you assume all misses are from level difference, but you can't use that with high resistance mobs or for binary spells.

Our goal is not to figure out when a spell is resisted based off level or resists, just the overall chance for the 5.14% mitigation chance to give a full resist. And thats what you're not quite seeing.

And you're still trying to assume that the 5.14% damage mitigated, and the 1% level based resist turn out to be an inflated number, which is not true. If you take the damage done out to an absurd amount you'll see you're wrong.

It's probably 6%. 5% would make sense if mobs resisted as if they had some form of level-based resistance (8 resist/level), while 6% would make sense if it was a flat percentage (2%/level). At level 60, those two numbers would have been equal, so I think it started at 8 resist/level when it was initially looked at. Figuring out what it is would take a huge amount of data though.

Or Blizzard actually telling us how it worked.