Just want to clear up something quickly, has it been confirmed if casters operate on 1 roll system or 2 roll system?

In particular, i am wondering if hit% affect my crit% over all spells

As far as I know nothing has been proven one way or the other.

Some Nihilum Mage got a GM to hint at it being 2 roll system if I'm not completely mistaken.

I'm inactive now, but this should be very easy to clear up if you can get your hands on a mage a 3 levels or so below dr. boom. Just stack lots of +crit, and no +hit and go play.

I really wouldn't trust anything in-game GM's say about advanced game mechanics.

We do know that partial resist is a separate roll than hit/crit/miss, but I'm guessing you're wondering if hit/crit are done in the same roll or not. I always assume they are for modeling but I could be wrong.

I don't think it is possible to conclusively test for one system or another at this point. Tests have been done to show that when crit chance is the same or exceeding hit chance, the caster is still getting non-crit hits. While this seems to indicate a two-roll system, it is not really conclusive. It is possible that something else happens when crit chance approaches hit chance, thus a non-linear pattern is observed.

On second thought, it would probably be very easy to test for the effect of a two-roll system given a high crit chance and a low hit chance. The one-roll theory and two-roll theory would separate into two very different ranges of expected values for the crit chance, and you would just have to do enough tests to see within which expected value range the actual crit chance falls. I assume this test has been done already, does anyone know where I can find it?

You would have to find a target 3 levels above you. Like someone already mentioned (I think) it would be very easy to do with a 65 mage and Dr.Boom.

This seems like such an obvious test that it must have already been done. Can I find the results somewhere around here?

Why limit yourself to +3 levels. Why not +10 or +15? Your miss rate doesn't ceiling at 17%, just boss mobs are at +3.

From the blue post about +hit:
If your target is the same level as you, a spell has a base chance to hit of 96%.

If the target is +1 level compared to you: 95%

+2 levels: 94%

+3 levels: 83% if the target is a mob, 87% if the target is a player.

+4 levels: mob: 72% player: 80%

+5 levels: mob: 61% player: 73%

Etc...

Get a high enough level mob and a high enough crit rating if you connect and it is not a crit than it isn't the one roll system.

It was proven that Rogues use a 2 roll system (in Nax in WoW 1.0). It is very likely that all dps classes use a 2 roll system, since it seems silly to me that only one class would use a different roll system.

Test:
Get a mob +9 levels greater than you.* (preferably with someone to hold its attention)
At +9 levels your hit chance works out to be -5%. I'm guessing there is floor in there so that you always have 1% chance to hit.

Attack the mob with 0 +hit

If at any point you connect with a spell and it is not a crit than we are working on a two roll system.

*Using the blue post and infering a linear relationship, for every level over +3 of yours the mob gains an 11% chance for you to miss. Extrapolate that down the line and that is where I came up with +9 levels. (Anything greater than +9 would work, just +9 is the minimum.)

From the tests I saw in the Naxx type days, spells are on a two roll system.
Roll 1) Hit or Miss (Binary spells resistance tossed into the mix here)
Roll 2) Crit or non-crit
Roll 3) If direct damage spell, roll for partial. (time of impact) 0%/25%/50%/100%

Almost convinced roll 3 happens on the "client" side. If a mob is firing at you and you resist it, the resistance message isn't shown until the spell lands. In other words, I don't think this happens until the decrement health stage of calculations.

Check your math.

The effect of level on crit in PvE outside of the yellow range (+/-2 levels) is undocumented and very well could also scale horribly, eliminating your crit chance as well as your hit chance. Your best bet is either finding an even-level with a -hit debuff (attumen's curse comes to mind) or else going out to gurubashi arena with a dozen casters and a dozen rogues and all popping a rogue simultaneously with CloS up. If you're doing Attument, I recomend using a different spell (eg searing pain for warlocks, ice lance for mages, rank 1 wrath for boomkin) so that the data is more easily parsable. Even attumen is sketchy since it's a boss (at best just +3, outside of linearland).

How do we know that level differences have effects on crit chances? Is it based on data generated from melee white attacks, since we are certain of a single roll system for those?

We have absolutly no data suggesting that casters do not have reduced crit rates against +4 mobs, and so without testing that first, any other tests involving crits on +4 mobs are pretty much worthless.

A nice way to test it would be taking an MT-level warrior, a healer and a mage to Kazzak, and having mage spam scorches until Kazzak enrages (or healer OOMs). Then repeat the process a few times.

Should cost ~100gold in total repairs and will net some cunclusive evidence.

Going at Kazzak a few times isn't "conclusive." A sample size of thousands of casts would be needed, meaning hours of time. After you get into the thousands, THEN you actually have a large enough sample size to compare crit versus hit without a lucky string of crits completely screwing everything.

Plus, do partial resists every make a spell fully resist? That could, once again, screw lots of stuff up, since there's no way of differentiating a level-based resist and a resistance-based full resist.

It seems to me the best place would be the level 73 elite demon in the fly-mount-only area in Nagrand. But again, you'd need a caster (preferably casting a non-binary, 1.5-sec cast), a tank, and a healer, all willing to devote at least several hours to spamming the same abilities over and over.

Attumen's curse, plus Combustion, should be plenty to produce a conclusive answer. Just make sure you don't get decursed.

Proposed test:

Take off all gear (to lower your crit chance - this will make sense in a second.) Combustion. Use fireballs to stack combustion up to the point where your crit, with gear on, is ~55%. Preferably with two or three charges left.

Put on all your gear. Go to Attumen (don't lose combustion!).

Wait til you're cursed. Cast. Record results. If you see a hit while cursed with >55% crit, then it's a safe assumption that it's two-roll, since miss/crit should exclude hit at that point.

(Alternately, if you can manage to have *zero* +hit, get cursed, and have >33% - 17% miss, 50% curse, leaves 33% on the table - crit, that should be enough to exclude hit from the table in a one-roll situation as well. For every % +hit in that circumstance, add 1% +crit to the equation of what's necessary to exclude hit.)

It's already been stated that hits can still occur when crit chance is greater than hit chance.

Thing is, we don't know if something screwy goes on when crit chance supersedes hit chance, or even gets near it. It *appears* it's a 2-roll system, but there's really no way to *prove* it's a 2-roll system.
EDIT: The closest we could do, unless I'm mistaken, would be to take something as already proposed where you're dealing with *realistic* numbers (17% resist rate and high-ish but realistic crit ratings), get thousands of tests, and THEN hope you come up with a statically significant figure. But there's still problems with that as well, as I already said.

I just thought of something.

We can easily test if casters have reduced crit rates against higher level mobs.

That's not the difficult part. The problem, once we find out that casters do indeed have reduced crit rates against higher level mobs, is to find the reason for it. The two possibilities are,

1. Crit rates actually scale down due to level differences.
2. Crit rates do not scale down, but crit chances are reduced due to the effect of the double roll system on mobs with higher miss chances.

It's these two confounding possibilities that are preventing us from finding out if it's indeed a single roll or double roll system. But what if it doesn't matter? What if the two possibilities are actually giving us the same expected results? Meaning, for all modeling intents and purposes, a double roll system is giving us approximately the correct crit chances, despite the first possibility being the correct one?

Has there been tests that actually shows that casters do receive reduced crit chances against higher level mobs?

We have absolutly no evidence, anecdotal or otherwise, that suggests that they do. Or that they don't.

Or that there isn't a 1% chance that a caster will instantly fall over dead when trying to cast a spell on a +9 mob.