WoW-Europe.com Forums -> Hit Rating Theorycraft

Razoraxe

< Forlorn Hope >

Bronze Dragonflight

100. Re: Hit Rating Theorycraft | 03/08/2007 15:25:44 UTC

About the 2 roll system for yellow(/special) attacks:

I'm kinda fussed about using Olaf as a test case, BECAUSE he has that amazingly high block rate. It is fairly possible that the lil' Dorf has been granted a special ability by Blizzard making the next attack done to him always a block (if possible). That would mean he is guarenteed to block, regardless of the crit roll the server did for you. (Similar to the Cold Blood mechanics, but the other way around). I'm also curious about where you got the 93.84% block chance he is supposed to have against 354 weapon skill
It's always dangerous to use special mobs and bosses for tests as they may have different behavior then the general rules.
A far better test subject would be to take a healer and a crap grey weapon with you to Netherstorm and go to the Kirin'Var Village. Plenty of ordinary, shield-wearing mobs right there and afaik those guys don't have any special blocking skills.

Edit:

Another question: Also about the 2 roll system.
This one is about the lvl 60 Rogue in AQ example. Couldn't find exactly where he posted his hit rate, so let's assume it's high enough to rule out backstab misses. The -hit% debuff he got should be 85%. He was using backstab, so no parry or block, just dodge. If you keep it simple and give the mob a 5% dodge rate, he now has 90% total avoidance.
This leaves 10% left for glancing, crits and crushing. Since it's only a lvl 62 mob, glancing blow chance was most likely below 10% which should be filled up with crits. However, when I rechecked the Dust Cloud debuff, I found 2 versions on WoWhead. http://www.wowhead.com/?search=dust+cloud.
The 30% one is currently not used, but was in the game at some point. Maybe he got that one after a mixup during a patch?

Last one about crit chance:
You mentioned the following:

Q u o t e:

Like you said, you have 13.2% crit rate on ALL YOUR ATTACKS and strangely you have 13.2% crit chance on your char sheet. Well, this is no strange at all. It is what you expected to be. You said on all your attacks which probably includes white and special attacks. I don't say that your overall crit chance will be lower than what your sheet says. What i'm saying is that white attacks will have a bigger crit chance while specials will have a lower one. So, your white crit chance may be, let's say 15% and your special crit 11.4% which makes an overall 13.2% crit chance.

Now, let's assume this is true. How does the tooltip know what your white : yellow attacks ratio is? If I'm grinding or tanking I have totally different white : yellow ratios yet my crit chance seems to be equal(ly low).
Also, you did a test with a total of 1k swings. That's about 500 white hits and 500 yellow hits. With your crit chance of around 30% you should have around 150 white and yellow crits each. I find this a rather low amount for test results, especially if you're not that far off with your actual results (10 crits more or less would give you a 2% difference already).
I suggest a test with sword & board on those unkillable blokes in Blasted Lands. Try with a 2.6 sword first, then a 1.3 dagger. Use EXACTLY the same amount of yellow hits (preferably not HS or Cleave, which eat your white attacks) over X amount of time (say, an hour) and compare crit results. Same gear, same crit chance, same amount of yellow attacks. Your white attack crit chance should be similar in both cases, although in the dagger one it will be more accurate.

Oh, and yes, I can do these tests for you, but who am I to tell you how it works? Test and see for yourself, I await your results =)

[ Post edited by Razoraxe ]

Q u o t e:
You have no point. You have no meaning. You have no significance.

You will be forgotten.

- Someone

I blame Spunt!

Cryinfreeman

< Lex Taliones >

101. Re: Hit Rating Theorycraft | 04/08/2007 18:22:31 UTC

Olaf has a special ability which makes him block 100% of the landed attacks against a 40 lvl player. However, this doesn't mean that he isn't acting like the other NPCs do. He is following the same formulas. Cold Blood is not the same. Cold blood is an attack made by player and not a NPC skill. Now, about the 93.84% block. To begin with, weapon skill reduces the chance that your attacks will be blocked. Each point reduces the chance by 0.04%. Olaf is a 40 lvl elite which means that he has 200 defence. I was 70 lvl when i performed the test and i had 354 weapon skill. Now, the difference is 354-200=154. 154*0.04=6.16% so 100-6.16=93.84%.

About the rogue. As he mentioned the debuff was 85%. Let's say that he had 90% total avoidance leaving him with 10%. The mob was 62 lvl which means that the rogue had 30% chance for glancing blows so this 10% is already filled with glancing blows leaving him no space for crits. To be more accurate,before the expansion,the chance for glancing blows was the following: Against a same lvl target you had 10% chance for glancing blows, against a +1 lvl than you, you had 20% chance, against +2 you had 30% and against +3 lvl you had 40% chance.

About the tooltip. The tooltip knows nothing about ratio. Actually, there is no ratio for white and yellow attacks because since you are a warrior you can never know the exact ammount of rage that you will have. What the tooltip does is to show you the crit chance you have depending on your items and your skills/talents. Let's say you have 30% crit chance. In your tooltip you will see 30%. Now, when you attack a target the server will calculate your crit chance depending on your attack skills and your opponents defence skills. Let's say that your server calculated your crit chance and it is 30%. This means that your yellow attacks crit chance will be 30% of your chance to hit. If your hit chance against this target is 90% your yellow crit chance will be 30% out of 90% which gives us 27%. The white attack crit chance will be 33%. In other words, when you will perform 100 attacks, 27of your specials will be crits and 33 of your auto will be crits too giving us a 30% overall crit chance. The rest will be normal hits.

I see no reason to perform such a test since in both cases my special attack crit chance will be the same since i will have the same stats and the same ammount of yellow attacks. The only thing it could be done would be a test in these mobs with a number of like 4k total attacks including both yellow and auto. However, this requires enough spare time and a healer willing to help. Personally, i don't have that much time now in the summer since i'm working but i will try to make the test in the near future. Anyway, i watched my crit rates during gruul and karazhan runs and the results were what i expected them tobe according to the 2 roll theory.

Jayna

< Niflheim >

Shadowmoon

102. Re: Hit Rating Theorycraft | 28/08/2007 16:09:31 UTC

Cryinfreeman, I'd like your opinion on this: http://forums.worldofwarcraft.com/thread.html?topicId=1272009577&sid=1

He says that it's better to go for AP/str after getting enough hit to never miss on special attacks. I kind of agree with him. You seem kind of bent on the hit stat being the > all stat, though. What do you think?

There is no right or wrong.
There is no true freedom.
The truth is never fully revealed.

Sometimes I'm so sarcastic even I don't know if I'm serious or not.

Ruma

Tarren Mill

103. Re: Hit Rating Theorycraft | 28/08/2007 16:17:37 UTC

Q u o t e:

About the rogue. As he mentioned the debuff was 85%. Let's say that he had 90% total avoidance leaving him with 10%. The mob was 62 lvl which means that the rogue had 30% chance for glancing blows so this 10% is already filled with glancing blows leaving him no space for crits. To be more accurate,before the expansion,the chance for glancing blows was the following: Against a same lvl target you had 10% chance for glancing blows, against a +1 lvl than you, you had 20% chance, against +2 you had 30% and against +3 lvl you had 40% chance.

you can glancing blow mobs of equal level?

Khayre

< Black Rat Society >

Earthen Ring

104. Re: Hit Rating Theorycraft | 29/08/2007 07:18:02 UTC

Q u o t e:
Cryinfreeman, I'd like your opinion on this: http://forums.worldofwarcraft.com/thread.html?topicId=1272009577&sid=1

He says that it's better to go for AP/str after getting enough hit to never miss on special attacks. I kind of agree with him. You seem kind of bent on the hit stat being the > all stat, though. What do you think?

That post is excellent and really sums up almost all of the stuff a fury warrior needs to know. Thanks a lot for the link. He is correct on the Str>>hit after 95 hit rating, and imo the rest of the post is also right on the spot.

edit: to Cryinfreeman, the miss chances you use in the OP has been found to be wrong, the true base miss chance for white attacks is 9% (with very high certainty) and 28% with DW. The effect of weapon skill is also wrong, the first 5 points of weapon skill is worth 3% of +hit (bringing the base miss chance down to 6%), making all skills/racials that increase weapon skill by 5 very very good for melee classes (almost too good imo, then again I'm just a bit jealous). The next 10 weapon skill points are worth 0.1% hit each, bringing the miss chance down to 5%/24% ( +15 weapon skill vs a bossmob makes the skill equal to the boss defense). Wowwiki has not been updated in ages on this subject, recent testing has pinpointed these figures (in EJ forums).

[ Post edited by Khayre ]

Jayna

< Niflheim >

Shadowmoon

105. Re: Hit Rating Theorycraft | 29/08/2007 10:02:41 UTC

Q u o t e:

That post is excellent and really sums up almost all of the stuff a fury warrior needs to know. Thanks a lot for the link. He is correct on the Str>>hit after 95 hit rating, and imo the rest of the post is also right on the spot.

edit: to Cryinfreeman, the miss chances you use in the OP has been found to be wrong, the true base miss chance for white attacks is 9% (with very high certainty) and 28% with DW. The effect of weapon skill is also wrong, the first 5 points of weapon skill is worth 3% of +hit (bringing the base miss chance down to 6%), making all skills/racials that increase weapon skill by 5 very very good for melee classes (almost too good imo, then again I'm just a bit jealous). The next 10 weapon skill points are worth 0.1% hit each, bringing the miss chance down to 5%/24% ( +15 weapon skill vs a bossmob makes the skill equal to the boss defense). Wowwiki has not been updated in ages on this subject, recent testing has pinpointed these figures (in EJ forums).

Do you have a link to the discussion on the EJ forums? I believe you, but I'd like the proof firsthand anyways.

There is no right or wrong.
There is no true freedom.
The truth is never fully revealed.

Sometimes I'm so sarcastic even I don't know if I'm serious or not.

Bogeyman

< Unbreakable >

Aggramar

106. Re: Hit Rating Theorycraft | 29/08/2007 11:12:13 UTC

ive read the theorycrafting threads on a 2 roll system and it seemed to be proven for a rogue, i dont remember reading that it was proven to be the same for all melee classes.

1k swings on ogres isnt a huge sample and a lot of testing would need be done to prove it

http://elitistjerks.com/f31/t11885-rogue_warrior_weapon_skill_adjustment_discussion/

ej discussion on wep skill jayna skip towards the end

Jayna

< Niflheim >

Shadowmoon

107. Re: Hit Rating Theorycraft | 29/08/2007 16:16:10 UTC

Q u o t e:
ive read the theorycrafting threads on a 2 roll system and it seemed to be proven for a rogue, i dont remember reading that it was proven to be the same for all melee classes.

1k swings on ogres isnt a huge sample and a lot of testing would need be done to prove it

http://elitistjerks.com/f31/t11885-rogue_warrior_weapon_skill_adjustment_discussion/

ej discussion on wep skill jayna skip towards the end

Thank ye.

There is no right or wrong.
There is no true freedom.
The truth is never fully revealed.

Sometimes I'm so sarcastic even I don't know if I'm serious or not.

Devvie

< nublets >

108. Re: Hit Rating Theorycraft | 29/08/2007 18:46:34 UTC

The hit rating theocrafy seem outdated.
I dont know how it really works, however.

Your Crit% only affects you hits. This means if you got 45% crit, then 45% of your landed attacks will be critical. Just think for a second, if the tables posted is correct, you would basicly aways crit and never get a normal hit, that just doesnt happen.

If a fury warrior adds hit, he will also get more crits not only more hits.

Cryinfreeman

< Lex Taliones >

109. Re: Hit Rating Theorycraft | 30/08/2007 03:02:43 UTC

Q u o t e:
Cryinfreeman, I'd like your opinion on this: http://forums.worldofwarcraft.com/thread.html?topicId=1272009577&sid=1

He says that it's better to go for AP/str after getting enough hit to never miss on special attacks. I kind of agree with him. You seem kind of bent on the hit stat being the > all stat, though. What do you think?

I saw his post. Well, his is wrong. He is completely wrong. To begin with, hit doesn't only affect white dmg. More hits means more white dmg and more white dmg means more rage which translates into yellow dmg. Hit offers a stable dps increase. Same goes with ap. On the other hand, crit offers burst dmg. Adding hit will increase your chance that you have rage for your special attacks. Adding crit is more risky. If you crit ok, you will get rage but if you don't crit?? You will have to queue your special attacks. Don't forget that as you have crit streaks you may also have miss streaks. Same goes for AP if you lack hit. AP will not help you at all if you don't hit. So in the end, you will starve rage and you will have to queue your special attacks.

Fury warrior with 95 hit rating + precision? Trust me, whoever do this will be in the last dps positions. And if you want, i dare any warrior to follow his instructions and then come and post here a screenshot with his position in his dmg meters. Get 2.5k ap unbuffed and 95 hit, you will still do no dmg. Look at this, when we say 24% miss chance for example, we mean 24% on MH and 24% on OH and not 24% overall. Note that hit is the only stat that can't be buffed. You don't trust me? Make a new thread here asking if 95 hit rating + precision is enough for good dps in raids. Hit is the worst dps stat after this point? I'm sorry, but this guy is clueless on how fury warriors work.

My short answer? Hit is crucial for endgame PVE. It's not just a coincidence that every fury warrior/rogue who does BT/Hyjal has at least 240 Hit. I'm not saying that Hit>all. It's just a trade off. It always depends on what stats and the ammount of the stats that you compare. 1% hit will definately give less dps than 300 AP and 10 ap will give less dps than 2% hit for example. The best thing is to have a balance between hit/crit/ap. Only then you will be able to do good dmg.

Q u o t e:
you can glancing blow mobs of equal level?

Yeah. Versus equal lvl mobs you have 10% chance for glancing blow. The new formula appears to be: GB= 10 + Mob defence - Player weapon skill, weapon skill being capped at 5 * level. So we have:

Vs Equal lvl mob: 10% GB chance
Vs +1 lvl mob: 15% GB chance
Vs +2 lvl mob: 20% GB chance
Vs +3 lvl mob/raid bosses: 25% GB chance

Link: http://www.wowwiki.com/Glancing_blow

Q u o t e:
edit: to Cryinfreeman, the miss chances you use in the OP has been found to be wrong, the true base miss chance for white attacks is 9% (with very high certainty) and 28% with DW. The effect of weapon skill is also wrong, the first 5 points of weapon skill is worth 3% of +hit (bringing the base miss chance down to 6%), making all skills/racials that increase weapon skill by 5 very very good for melee classes (almost too good imo, then again I'm just a bit jealous). The next 10 weapon skill points are worth 0.1% hit each, bringing the miss chance down to 5%/24% ( +15 weapon skill vs a bossmob makes the skill equal to the boss defense). Wowwiki has not been updated in ages on this subject, recent testing has pinpointed these figures (in EJ forums)

I looked into it. Actually, spended like 6 hours of reading in the EJ forums :D. Well, yeah. Miss chance is 28% for dw and 9% for 2handers. I'm gonna edit my first post too. However, it's not yet known how much hit does each point of weapon skill gives you. It is conifrmed that when the lvl difference is less or equal to 10, you have 0.1% miss per point of diference. What is not known yet is how much hit each weapon skill gives you when the difference is 11-15, in other words how much hit you gain when going from 350 weapon skill to 355. The theory of 0.6% per point has already been disproven which means that the first 5 weapon skill points don't give you a 3% hit.

Q u o t e:
The hit rating theocrafy seem outdated.
I dont know how it really works, however.

Your Crit% only affects you hits. This means if you got 45% crit, then 45% of your landed attacks will be critical. Just think for a second, if the tables posted is correct, you would basicly aways crit and never get a normal hit, that just doesnt happen.

If a fury warrior adds hit, he will also get more crits not only more hits.

It has been proven several times that special attacks use a 2 roll system while auto attacks use 1 roll system. What you say refers only to special attacks where crit is a % of landed attacks. In auto attacks, hit and crit are 2 different possible outcomes. In this case ( auto attacks), hit will increase your crit chance only if you are already crit capped. And yeah, the tables i posted are correct. Also, you can notice crits the moment that you don't have hits.

[ Post edited by Cryinfreeman ]

Khayre

< Black Rat Society >

Earthen Ring

110. Re: Hit Rating Theorycraft | 30/08/2007 12:04:04 UTC

Sorry to say that I really don't agree with you Cryinfreeman.

+1% crit will give you exactly the same amount of rage as +1% hit for white dps (the difference is that +crit will give more burst rage). For yellow dps, it does not matter at all whether it's a two-roll or one-roll system, hit is capped at 9% which means that +hit becomes worthless for all yellow damage after 9%.

In total, there's absolutely nothing wrong with having more hit rating than 95, but saying that it's imperative to have >200 for raiding is just wrong. If you have AP and crit to make up for it you will do good damage. Hit rating will be a good stat, but str/ap and crit will just be better for raw dps (total: yellow + white). If you want to trade some damage for smoother rage generation (which is what you get with choosing hit above crit).

The reason why almost all raiding warriors run around with +hit >200 is just that the best gear has +hit. You should of course not gimp yourself by choosing subpar gear just to get crit instead of hit. I I use a spreadsheet (including white damage, instant cycle damage and WF procs) and calculate the value of different stats I get that 1 str = 0.85 crit = 0.6 hit, i.e. the difference between hit and crit is not big. The end-game gear currently focuses on haste and -armor stats rather than hit and crit, these stats are both very very good. The gear at that level also has +hit, thus the warriors at that lvl of gear will have >240 hit. Still, if they know what they are doing they will choose str>crit>hit (in ratings). My impression is that end-game warriors know this (based on discussions in the EJ forums).

You will not see any warrior running around with 95 hit rating and 35% crit and 2000AP, the gear to get these stats just doesn't exist (maybe it does but it will probably involve using a lot of leather). Sure, most of my reasoning is theorycrafting, but until I see a some careful tests I will believe that this type of gear will do excellent damage. If it turns out that I'm wrong, then the current state of theorycrafting is completely wrong. I agree that balancing the stats is a good idea, the available gear will do this for you automatically, most of us will use the gear we find and can't really choose an entire set just for fun.

You're right on that we still don't know the exact +hit per weapon skill for the first 5 points but the 3% for the first 5 points is definitely proven, you will have a 25% miss rate when DW with +5 weapon skill (without precision and +hit).

Jayna

< Niflheim >

Shadowmoon

111. Re: Hit Rating Theorycraft | 30/08/2007 13:18:26 UTC

Still, I'd rather choose 4st+4crit gems over +8hitrating gems, wouldn't you? Hit rating is thrown at you on epic plate dps gear, so you likely won't have hit problems anyways.

I was GOING to write something more, but I forgot what it was. o.O

[ Post edited by Jayna ]

There is no right or wrong.
There is no true freedom.
The truth is never fully revealed.

Sometimes I'm so sarcastic even I don't know if I'm serious or not.

Cryinfreeman

< Lex Taliones >

112. Re: Hit Rating Theorycraft | 30/08/2007 21:05:53 UTC

To begin with, indeed for yellow attacks hit is worthless past 9%. This means that crit/ap are more important past this point. +1% crit will give you the same ammount of rage with +1% hit. This is true. There is a difference though. Hit offers more steady rage regen than crit does. This becomes more important when you enter the execute range. See it this way. You perform execute leaving you with no rage. Then, a miss streak follows your execute which leaves you again with no rage. In other words, you will have to queue your special attacks till you get some rage. This will lead you in loosing some dmg. If this happens a few times during the execute range, you will loose way too much dmg.

Getting at least 200 hit rating for endgame pve is not wrong at all. The wrong would be to say that you don't need more than 200 hit. See it this way since you talk about itemization. Why all items provide so much hit since it's not that important? Blizzard could remove the hit, some of it, and give more ap and strength instead since it would be better. Well, i'm not talking about haste/ armor ignor since these are stats that probably are better than hit/crit/ap, especially haste. About the spreadsheets you mentioned, the spreadsheet doesn't calculate your miss streaks. It assumes you will have a steady dps and rage regen and under this assumption all the calculations are made. I'm not saying that hit is the most important stat and that we should stack it at all costs. I'm saying that hit is a very important stat and should be given much attention. Here is a small comparison i made on tankspot:

Hit: Hit means less misses which means more successful normal attacks which by it's turn means more steady dps and therefore a steady rage regen. Adding hit will decrease your chance to starve for rage so you won't have to queue any special attack. Also, more successful attacks mean more rage which means by it's turn, more special attacks. So, in the end, hit increases your yellow dmg too indirectly. It is important to note that when you have 24% miss chance for example and you dual wield, both your MH and your OH have a 24% each and not a 24% miss chance overally.

Crit: Crit is the burst dmg. It was and it will always be a very good stat for dps. However, due to the nature of the attack tables and the way that attacks behave, you may have crit streaks but you may not as well. You can have miss streaks instead which means no rage at all.

AP: Ap is the middle way between crit and hit. It offers steady dps increase but only when you land your attacks. So, we can see that ap is indirectly connected with hit and miss chances.

Yet again, if you are not convinced about the importance of hit, i dare you to come here and post your dps meters after a raid when you have 100-150 hit. I'm 101% sure you will fail in doing dmg.

Anyway, about the 5 weapon skill. Could you post the link where it proves that the first 5 points give +3% hit? I took a look in the Ej forums but after reading all this theorycraft for like 6+ hours i got really tired and i can't find it :)

[ Post edited by Cryinfreeman ]

Khayre

< Black Rat Society >

Earthen Ring

113. Re: Hit Rating Theorycraft | 31/08/2007 07:54:08 UTC

Ah well, I agree with you that +hit will give steadier rage generation than +crit, but there are other ways to get a steady rage generation (and guarantee execute at each gcd), e.g. use fast weapons (you lose on ww damage but gain faster rage generation).

The spreadsheet I'm using can't calculate the effect of miss streaks, that's true, but on the other hand you will then also have to compute the effect of "non-crit" streaks (with effect on the flurry uptime). However, the spreadsheet does compute the mean rage generation and uses specials according to this (including Execute spam). Using this I do get that crit and str is better than hit in terms of dps.

You might say that I should go and test this in the game, but the fact is that there are so many factors determining the total damage I do in a given raid that it becomes pointless to look at my results to say the gear I'm using is bad. Perhaps I'm just crap at doing dps, I'm lagging badly or just having bad luck. What needs to be done to test this properly is for a well geared and experienced dps warrior to run exactly (or close to exact anyway) the same raid with different gear and report the output damage. This is a test that will likely never be done, it takes a long time to collect gear. I certainly don't have time to do this, therefore I trust the spreadsheet instead, and to me it produces fairly accurate predictions.

Anyway, as you say +hit is not a bad stat and it's on the end-game gear anyway. Then again, perhaps Blizzard is designing this gear with +hit instead of +crit/Str because they are balancing the output damage...

The weapon skill discussion in EJ forums is growing very long indeed, the tests done to get the 6% with 355 weapon skill is reported in posts:
- http://elitistjerks.com/429032-post142.html Rogue with 1/2 weapon expertise (i.e. +5 skill) and precision.
He finds that using 313 hit rating he gets 0 misses in 2175 swings (translating to an upper 95% confidence limit of 0.14%) => base miss chance <25.02 % (c.l., 25.16).
- http://elitistjerks.com/432304-post196.html Hunter with dwarf racial +5 skill guns.
Using a mix of 100 and 98 hit rating (I'll use 99) he gets zero misses in 2096 shots
=> base miss chance <6.27 %
- might be more in there, these were the posts I remembered

So from these tests the base miss chance with +5 weapon skill (or vs a lvl 72 mob) is ~6 % and for sure <6.16 (95% confidence).

Surprise

< Anathema >

Wrathbringer

114. Re: Hit Rating Theorycraft | 01/09/2007 00:44:35 UTC

Ah, some nice discussion here as well (although in the wrong forum for me).

Anyway:
1)
I'm one of the rogues (besides pf) mentioned in the EJ thread that were trying to test the 2 roll theory (http://elitistjerks.com/f33/t9104-backstab_two_rolls/p2/ - sp00n). As the question came up about my hit rating, the link to the character planer is still working, albeit being totally outdated now, it still represents my gear at that time, which had +20% hit.

2)
Current testings make me believe that 1 point of weapon skill is worth 0.4% to hit while the difference between defense and weapon skill is larger than 9.
I.e. the first five points (until 355 skill) is worth a total of 2% hit. And since the 0.6% theory sadly was disproved, there has to be a "jump" when you reach this value, which with the 0.4% is exactly 1%.
So you have 5*0.4% through the weapon skill + a "bonus" for reaching 5 skill points.
I can't exactly explain why it is this way, but the data strongly seems to indicate this. Anything above 0.4024% hit per point of weapon skill has been disproved, and anything below 0.3707% seems rather unlikely by now.

For anything below 10, 1 point is worth exactly 0.1% hit. This goes even down to low level mobs.
I.e. with 200 hit rating, dual wielding and 350 weapon skill, you will miss a level 58 mob, but won't notice any misses against a level 57 mob.

We have no real clue on how weapon skill affects dodge, parry, block and crits so far.
The only source of information was Crezax' posting (http://forums.wow-europe.com/thread.html?topicId=166546791&pageNo=1&sid=1#2) about weapon skill, which has been disproved by now, and the tiny entry in the patch notes (http://www.wow-europe.com/en/patchnotes/patch-2010.html), which is neither helpful.

3)
It would be *reallyreallyreally* helpful if Blizzard deciced to actually reveal the mechanics behind weapon skill.
I mean, it is present on our gear. And we have no real clue what it actually does. In tests lasting several weeks we meanwhile managed to get a somewhat decent formula concerning one single aspect of its value.
Following this attitude, they could have named the hit rating stat "stat of massive damage increase" and crit rating the "bigger numbers for you stat". We would have been in the exact same situation as we are now with weapon skill. The name does say something, but we have no clue what it actually does.

Just that with these two examples it would have been even easier to find their impacts, since they have only one effect. With weapon skill we are dealing with multiple possible, but not proven impacts, such as miss, dodge, parry, block and crit. And the different effects don't seem to share the same formula at all either!

Again, I plead Blizzard to reveal the formula behind weapon skill. Hope dies last, does it?

// Edit
One more note. For the hit rating <-> hit % conversion, I prefer 10*82/52 instead of 15.77. It's the actual conversion formula used by Blizzard from level 60 to 70 (or at least very near to it) and works for most other stats as well (e.g. weapon skill is 2,5*82/52).
The first number is the amount of rating required on level 60 to match a 1% increase.

[ Post edited by Surprise ]

Cryinfreeman

< Lex Taliones >

115. Re: Hit Rating Theorycraft | 01/09/2007 07:51:45 UTC

From my experience so far, i think that nothing can guarantee you an execute in each GCD. It's not about fast or slow weapons or having much hit or crit. It's that we have to do with % in each attack. You can dodge, miss, hit, crit etc etc. You are talking about having steady rage regen in the execute range. However, it's not only the execute range. Best is to try to keep a steady rage regen during the whole fight. To manage this, you need a balance in your stats. I'm not saying that ppl should stack hit gimping other stats. NO. I'm saying we need a balance between ap/hit/crit. Fury warrior is a very demanding class in terms of stats. They need everything and in great ammounts. 40% crit with 5% hit is not good. 20% crit with 28% hit is not good neither.

Also, fury warriors don't look for crit streaks. A warrior wants to keep flurry up as much as possible. With a decent ammount of crit like 27% for example plus the full buffs which will get you up to 30% at least, this can guarantee you that flurry will be up almost all the time. Crit streaks can boost your dmg a bit but they don't help the flurry uptime. Also, since we have %, if you have a crit streak, then you are going to have a non-crit streak because this is how the mechanics work. Before saying that crit>hit or whatever, there is a big factor that we should consider. It's the procs. Procs can really modify the ammount of dps that hit/ap/crit will give you. Mongoose for example, it benefits more from hit than crit.

A final note. AP and Crit can be boosted very much through raid buffs. Hit can't. This is why you must have a good ammount of hit. Simply because you can't buff it.

About Weapon skill. Surprise, i believe too that this is the way weapon skill works. So far, the 0.4% plus 1% seems to be working very good but yet, it's not 100% proven.
Unfortunately, Blizzard will never reveal the mechanics of this game as they never did so far. The only exception is the post that Crezax made and that's it. I guess, the philosophy in Blizzard is to never reveal the secrets of the games and they are doing it very good so far, but as Surprise said, hope dies last :)

EDIT: Surprise, since you are a guest in our forums, and always welcome to take part in this discussion, i would like to know your opinion about hit in end game raiding. For me as you probably saw, hit is a very important stat, one of the most important tbh but since i see ppl disagreeing, i would love to hear your opinion from the rogue's perspective. Ofc, rogues and warriors are not the same and they differ in many ways but i would like to hear your opinion :)

[ Post edited by Cryinfreeman ]

Surprise

< Anathema >

Wrathbringer

116. Re: Hit Rating Theorycraft | 01/09/2007 10:37:16 UTC

Well, rogues have it relatively easy due to that there are two very good excel DPS spreadsheets available on which we can base our AEP (agility equivalence points) on.

And these 2 spreadsheets show quite clear that 1 hit rating is the best stat in terms of long time DPS for (raiding) rogues, only matched by haste rating in its current form, which will change after the next patch of course.
Even (especially) with my not so standard combat-hemo spec, hit is the best stat for me, despite the lack of Combat Potency.

As for warriors, I have no real clue. The only DPS spreadsheet I've seen for warriors was quite some time ago and (the creator may forgive me) was not as detailed as the rogue ones by far. Of course I started my WoW career with a warrior, but never actually seriously raided with him.

You can ask me about my opinion all the way, but to be honest, and that is my real opinion, one would need a very decent calculation to exactly measure the various impacts (rage generation, flurry uptime, procs from enchants, etc) for the different stats.

My gut as a rogue tells me to always prefer hit (of course that's what personalized AEP are for, to compare different items with different stats!) over other stats, but this may not apply to warriors.

Varg

< Descendants of Yggdr… >

Draenor

117. Re: Hit Rating Theorycraft | 01/09/2007 13:08:22 UTC

Q u o t e:
From my experience so far, i think that nothing can guarantee you an execute in each GCD. It's not about fast or slow weapons or having much hit or crit. It's that we have to do with % in each attack. You can dodge, miss, hit, crit etc etc. You are talking about having steady rage regen in the execute range. However, it's not only the execute range. Best is to try to keep a steady rage regen during the whole fight. To manage this, you need a balance in your stats. I'm not saying that ppl should stack hit gimping other stats. NO. I'm saying we need a balance between ap/hit/crit. Fury warrior is a very demanding class in terms of stats. They need everything and in great ammounts. 40% crit with 5% hit is not good. 20% crit with 28% hit is not good neither.

I agree that a steady rage generation is preferable, if you can choose it. With my stats (129 hit rating, around 32% unbuffed crit, 2000AP), I have no problems getting and keeping rage. I always manage to keep enough rage for when the BT and WW cools down, but I suspect that I could work with a lower "rage buffer" with a higher hit. However, it is not a real problem for me, as I always aim for keeping at least 30 rage, only going below that point when doing a BT/WW. It's not less DPS, just requires a bit more paying attention. This is why I am not personally weighting the steady rage generation as a very important attribute, possibly with the exception of the execute phase.

Q u o t e:
Also, fury warriors don't look for crit streaks. A warrior wants to keep flurry up as much as possible. With a decent ammount of crit like 27% for example plus the full buffs which will get you up to 30% at least, this can guarantee you that flurry will be up almost all the time. Crit streaks can boost your dmg a bit but they don't help the flurry uptime. Also, since we have %, if you have a crit streak, then you are going to have a non-crit streak because this is how the mechanics work.

I'm going to be bold and say that this is partially wrong.

From a statistical approach over a longer fight (which I assume is what this thread is about), the probability P of an attack being flurried is:

P = 1 - (1 - crit%)^(3+n)

where n is the number of yellow attacks (excluding HS/Cleave) made over the time period of three white attacks. This number varies depending on your weapon speed.

However, nowhere will this imply that a crit streak will create a non-crit streak afterwards. Quite the contrary: Each of the samples of this probability distribution are independent, and one crit will not affect whether the next hit will be crit or not.

(The only exception to this might be the mongoose enchant, where a hit might proc the enchant, giving a higher crit chance, and thus a higher flurry probability.)

Q u o t e:
Before saying that crit>hit or whatever, there is a big factor that we should consider. It's the procs. Procs can really modify the ammount of dps that hit/ap/crit will give you. Mongoose for example, it benefits more from hit than crit.

This is true. Also account for Windfury totem.

Q u o t e:
A final note. AP and Crit can be boosted very much through raid buffs. Hit can't. This is why you must have a good ammount of hit. Simply because you can't buff it.

This is true if a high hit is actually needed. I am not convinced that it is (as long as you get over the yellow-miss-minimum), although it does seem like the importance of AP decreases as your overall DPS increases (including hit and crit). This might advocate weighting hit rating a bit higher in relation to AP, but the same would be true for crit.

Cryinfreeman

< Lex Taliones >

118. Re: Hit Rating Theorycraft | 01/09/2007 13:26:14 UTC

Well, Varg has some good points here. I need to clarify some things though, at least from my own view.

About crit and non-crit streaks. Let's say we have 30% crit. Over a long fight, 2k swings for example, you will notice 600 critical attacks. If you have a crit streak with 100 attacks on a row being critical, then you are going to have a hit streak with 100 of your attacks on a row being hits or 2x50 non-crit streaks or 5x20 non crit streaks etc. This is almost sure since in such a long fight, the % will be exact. You can't have 100 crits on a row and no non-crit streaks since this would lead in a extremely high crit chance. I repeat though, this refers to long fights. In short fights, like 100 attacks, this doesn't mean that if you have crit streaks you will have non-crit streaks. This is indeed a very extreme example and i may be wrong but my impression is that this is the way that it will work.

[ Post edited by Cryinfreeman ]

Varg

< Descendants of Yggdr… >

Draenor

119. Re: Hit Rating Theorycraft | 01/09/2007 13:33:46 UTC

Yes, I understand that it is easy to note such non-crit streaks, but I do not believe there's a statistical theory to support that it will happen for the binomial distribution that these flurried attacks are. The average crit chance is not something the events strive to achieve, it's just an average measurement of the situation. There will be crit streaks, and there will be miss streaks, but they are completely independent of each other.

Although ...

A non-flurried streak will only follow a flurried streak, as there is nothing else it can follow, and that might be what you are experiencing =)

[ Post edited by Varg ]

Warrior