See, Immune filling up the table as a higher priority to Dodge doesn't make much sense to me.
Well specifically having a higher priority than Dodge but NOT Miss.

Why do it that way?
That seems a half and half measure. Either immune trumping the entire hit table (no miss just 100% immune) or merely replacing all damage with immune (still block/parry/dodge - hit/crit/glancing to immune) would make more sense to me.

They were worried about Warrior's overpowering immune mobs and didn't want to give rogues partial energy returns?
I don't buy it.

Well there's always the "realistic" explanation that if you miss, obviously you don't even touch him. But if you attack an immune mob... why would he need to bother dodging?

http://forums.worldofwarcraft.com/th...picId=88680668

They could have left weapon expertise alone in the first place and then they wouldn't have had to do that.

Er, they deleted the thread on accident I guess, but here is the entire post as I still had it open in a browser tab window:

I used some 294 HR gear with Precision/WeExp/Latro's (Screenshot), should add up to something like 24.94% hit if you count Latro's as 3 skill. Doing Prince/Illhoof/Maiden I got 5 misses over Prince + Illhoof, none on Maiden, over a total of roughly 1151 swings (~0.43%) from what a quick scan reveals (Combat Log).

This would mesh with possibly needing 16.3 HR for 1% hit (according to the formula on wowwiki.com) against a level 73 target, assuming hit rating scales up the same way it scales down.

With your scenarios listed in the last paragraph, you are making assumptions about +skill vs +3 mobs which may or may not be correct. Since you cannot assume what +skill does vs a +3 mob, for solid testing of the hit rating/hit vs a 73, one must use a 350 skill until the true value of +skill/hit against a 73 is determined. Remember, when setting up tests, always be careful of your known/unknown variables.

It's true. That test wouldn't prove that it was correct by itself. But coming up with enough hit rating to test that could be hard. And, honestly? If it could be shown that the breakpoint where you can't miss anymore is 308 or 305 w/WEx, that'd be pretty convincing, IMO. But if you can come up with 321 hit rating + precision to test it without WEx, by all means do so. I'd be curious to see the results.

Ald

Well, you could also run some tests assuming +skill gives .2% hit and that would mean that using 360 skill + precision would start you off at 7% hit and only needing 292 hit to cap assuming 25.5 base miss vs a 73 and 15.77 hitrating/hit. I do not think it would take much testing to find out how much hit skill gives vs a +3 given that you can find the hit caps and slide up/down them using +skill/hit. However, I dont have much access to +3's now as noone runs old instances anymore.

...except that's already been disproved by Gogge's numbers. If weapon skill gives .2% hit, with his +13 weapon skill, even if it takes 16.3 hit rating for 1% hit, *and* the base miss chance is 25.5%, you still only need 293 hit rating to cap out. And he had 294, and was still missing. So regardless of what we have or haven't proved, we can be pretty darn sure that +1 skill doesn't give +.2% hit against level 73s.

As for determining the hit number - it takes a *lot* of data to get the sort of accuracy we'd need - we're talking 10k attacks against level 73s only, which is hard to come by. Looking for the breakpoint where hit caps out is a lot more efficient in pinning down the conversions. Hence: if some people have the ability to test with 300+ hit rating, that would be very interesting to see.

Ald

Probably not so interesting anymore but here is more haalum numbers:

0 hit, 0 precision, 350 weapon skill:

73225 immunes, 23251 misses(24,1002% miss)

This was completely naked with 2 white daggers and no talent points spent. Took 18 hours to get this kind of data. PTR is nice for this stuff

I'm gonna go try kill lots of lvl 73 non elites up in Blades Edge Mountain later today with 305 hit rating. I assume they don't have to be elites?

But you are saying +1 skill doesnt give +.2% hit based on data you do not even know is correct (ie base miss is 25.5% which is complete conjecture at this point, and a hitrating/hit vs a 73 of 15.77 which you are trying to prove or disprove that now). If you just change the assumption of hitrating/hit to 16.3 or whatever, then you need over 300 hit rating to cap out vs a 73.

dont know if you guys know Ippon on the WoW Us forums, but he is currently running with a hit rating of 315 and didint miss a single attack on a full kara run... One can maybe work backwards from this and see what happens or where this boundry is....

I just wish Blizzard can confirm exaclty how +skill works

Quick question, as an alliance rogue I was always under the impression that for sword/fist spec rogues Windfury was better then Deadly Poison MH, however right now for a Combat build the spreadsheet is showing that I am better off with MH Deadly/OH Instant or Deadly, then I am with MH Windfury/OH Deadly or Instant. Did Deadly Poison end up scaling up better then Windfury? Should I just ask the Shaman to drop Grace of Air instead and just use Deadly? Thanks in advance!

This threw me for a loop at first as well, but the difference is that windfury buff on your weapon will only affect the "buffed dps" number. See if that's changing...I bet it is

Thanks, I was getting concerned that I was having the Shaman drop WF for no reason... now I don't feel like a complete idiot.

Yes, the Warpscale Leggings were changed in the recent patch.

Okay, let me spell this out.

Lets assume for the moment that +skill gives .2% hit rating, as you assert.

Gogge was wearing Latro's, and had 5/5 Precision and WEx. Thus he had a base +5% hit, and +13 weapon skill, which, according to your theory, would grant 2.6% hit.

The highest base miss rate that anyone has conjectured in this thread is 25.5%. Thus, if he's getting 5% from precision and 2.6% from skill, it would require 17.9% from hit rating to min out your miss chance.

The highest number put forward for the amount of hit rating needed to gain 1% hit is 16.3.

Thus, according to this theory, the most hit rating you could possibly need to min out your miss chance is 17.9 time 16.3.

17.9 times 16.3 is 292.

Thus, even assuming the *highest* base miss rate anyone has proposed, and the *most* hit rating required for 1% hit, 292 is *still* enough to max out your hit rating if weapon skill gives .2% hit.

Gogge had 294, and was missing.

Thus, we can conclude:

1) the base miss chance is higher than 25.5%
2) It requires more than 16.3 hit rating for 1% hit
or
3) Weapon Skill does not give .2% hit.

As I've heard absolutely no reasonable theories that would support the first two, my inclination is to accept the third as the most likely conclusion. You're welcome to disagree with me, of course, but it seems most likely to me. Hence, when I devise experiments to further test what's going on, I'm going to assume that weapon skill does not, in fact, give .2% hit. If you'd like to propose a theory that would explain what's going on and includes .2% hit per weapon skill, by all means, do so. But as it seems pretty likely that that's not the case, I'm proposing tests for what seem to me to be the more likely cases first.

So, it appears that it is possible to miss at 305 hit rating. What can we conclude?

Well, lets assume for the moment that weapon skill gives .1% hit like it does vs 70s. This isn't necessarily true, but for the reasons I just descibed I think it's more likely than any other theory I've heard. Thus precision and WEx are giving 6% hit.

Under the theory that it takes 16.3 hit rating for 1% hit, 305 hit rating would contribute another 18.66% hit, for a total +hit% of 24.66%.

Under the theory that it takes 15.77 hit rating for 1% hit, 305 hit rating would contribute another 19.34% hit, for a total of 25.34%.

In either case, if the base miss chance were 24.6%, it would be impossible to miss. Therefore, it seems likely that the base miss rate is higher.

Since we're assuming that +skill gives +.1% hit, it seems reasonable to conjecture that +defense might give +.1% miss. If so, then the base miss rate would be 25.5%.

Under this theory, the miss chances would be .84% (if it's 16.3 hit rating per hit) or .16% (if it's 15.77).

Zito launched 2447 attacks all told. With a base miss chance of .84%, he'd expect to miss 20.6 times, with a standard deviation of 4.5. His data would thus be 4 standard deviations from the mean... hence this theory is incorrect.

If the miss chance were .16%, he'd expect to miss 3.9 times, with a standard deviation of 2. Under this theory, his data is 1 std deviation from the mean - well within acceptted statistical bounds.

So, the conclusion? Zito's data is totally consistant with weapon skill giving .1% hit, 15.77 hit rating needed for 1% hit, and a base miss chance of 25.5%. That is: the amount of hit rating needed does not depend on opponent's level, each point of skill gives .1% hit, and each point of +defense gives +.1% miss.

We have not proved this theory is correct, but it seems reasonable and fits all known data.

Next steps:

Under this theory the hit cap with WEx and precision would be 308, or 303 with Latro's. Hence, if one ever missed with hit rating above this level, that would disprove the theory. Running 10k attacks or so at this level without a miss would be nice evidence to support the theory.

Testing at 307 (or 302 w/Latro's) to confirm that a miss could occur might also be worthwhile. Note, however, that the miss chance is on the order of 1 in 3000 for this... even if you don't miss in the first 5k attacks, that doesn't prove much. This test should be run for at least 10k attacks, or until a miss occurrs.

Ideally, these tests should be run at at least 2 different levels of weapon skill, to provide confirmation to the assumption that +1 weapon skill gives .1% hit. If you'd like to know the proposed hit cap for an amount of +skill other than 10 or 13, let me know and I will compute it for you.

The other test that could be performed is confirmation of the theory regarding defense. Theoretically, this should apply to bosses hitting tanks as well as rogues hitting bosses. Hence, a log of attacks directed at tanks by bosses might be instructive.

For instance, on a Gruul attempt the other night, Gruul attacked the MT 123 times, and missed 21, for a miss rate of 17.1%. The tank had just over 500 defense rating (I didn't check which gear set he was using, he varies between 500 and 510 defense depending on gear). Under the old +.04%/-.04% theory, the predicted miss rate would be about 10.6%. Under the new +.1/-.1% theory, it would be about 19%. Unfortunately, the data set is too small to say conclusively what's going on, but with an observed miss rate of 17%, a theoretical value of 19% looks somewhat more plausible than a theoretical value of 10.6%. If someone could provide a parse of their tank getting hit a thousand times or so by bosses, and the defense rating of that tank, that would also give some insight into the validity of this theory.

Blue post stating that 15.8 Rating = 1% Hit @ Level 70 http://blue.cardplace.com/newcache/us/36573662.htm

Blue post stating that +skill is worth .2% hit vs lvl 73's
http://blue.cardplace.com/newcache/en/166546791.htm


Hopefully these numbers can help you narrow your conjecture down a bit.

Very solid post Aldriana. I had neglected to take into account the use of latro's before so you are correct on that matter.

If the info you are assuming is correct, then we can no longer assume everything in the blue post about +skill and further testing on parry/dodge/crit would need to be done. Considering it is reported that +3's parry at about a 12% rate (this came from a thread on EJ a while back, it might have been a crushing blow or glancing blow thread, but the subject was not about parry rate, it was only brought up as odd), this would mean that a default 5% +.6%*15 = 14%. This number seems to fall in line with what was told to us. Also, of note is that +skill has been associated as a tank stat (Maladath/THC) by blizzard even before the nerfing of skill, so this is probably not something new, but existed all the time. Dodge most likely follows the .1% rule as well. Could it simply be that the "blue" poster was simply misinformed and the person telling them meant that +1 skill is effectively +.2% hit from .1% miss and .1% dodge? I do not know, and the "blue" poster in question is no longer a blizzard employee and the information source of the post was never determined and interestingly enough, the thread itself was short lived. The problem I see with the blue post in question and the numbers that seem to be popping up through testing disagreeing is that the post could not have been made whimsically. Imo, either the poster was A) misinformed/miscommunication or B) posted the wrong numbers.

Anyway, I cannot reach the lvl 60 hitcap with the gear I have and cannot perform tests on 63s that would provide relevant results(I have about 2 weeks left on my account now), so it is up to others to prove/disprove the conjectures placed forth in this thread.

*wall of text hits Daemona critically for 1337 hitrating*

Ok, I am baffled... but i still want to help you

I have access to gear that goes up to 341 hitrating (including weapons15+18 rating). If i remove the weapons, its 308 hitrating, to which i could add up to 2 latros and/or a weapon without hit-rating. Specced precision and WEx and would like to avoid respeccing as i am in a major raiding guild and would have to do the tests in off-times only which would require me to respecc for testing without precision or WEx

give me something like "208 hit, 1 latros, mob lvl70, hit amount" and i see how fast i can deliver a result