Toots Hepcat

The 6.5% vs 5.x% dodge question looks like the detritus of an old calculation from back when weapon skill could exceed 350. Wowwiki being, well, wowwiki, nobody who knew what they were doing ever took the time to update the old page.

According to the rogues (who I trust on this matter, since not being missed or dodged is important to them), it seems that post 2.3, the base dual wield miss chance against l73 mobs went up to 28%, and the base dodge/parry rate went up to 6.5%. (This is opposed to 25.4% and 5.4% prior to 2.3).

If Yo's reading, I'd appreciate confirmation that these values are used by his simulator.

 

Yo!

There is an attack table on Forensic Report tab.
28% miss, 5.6% dodge are being used (9% miss for specials)
checked again if it outputs same as coded - yes it does, you can check it with different hit/exp values
Bug found - uses same 28% miss for 2H weapons
Just give me community approval to change dodge to 6.5%. Did not find explanation in rogues thread though it is easy to check with wws (down at the moment)

I will change model if needed according to your final test results.

 

Malan

I could have sworn that we ruled in on this a couple of weeks ago and it was decided that the rogue website was dyslexic and that it was actually 5.6%.

Shitting up every single thread on EJ since '06

 

Gehenna

I don't think I'll be attempting/aiming to obtain more than 4-5% anti-dodge.

Reason being that the extra few points will not be helping me on trash or regular mobs, so in the long run, the advantages of capping versus a boss might really just equalize that stat weight.

Might pick up an extra item to cap it just for boss fights though.

 

Myul

Well, you will really want the three new pieces of tier 6 (since there are no alternatives),
[Skyshatter Wristguards]
[Skyshatter Girdle]
[Skyshatter Greaves]
giving you 68 expertise rating or 18.5 expertise. That's 4.6% and should be enough.

For bosses or until you have your three pieces, you may want to wear [Shard of Contempt] and maybe add [Shoulderpads of the Stranger] and/or [Belt of One-Hundred Deaths].

Wearing Shard + DST should be the best combination until you are deep in sunwell content, where "Blackened Naaru Sliver" (that haste/battletrance one in the right bottom of the trinket section) could be nice, tough.

And for trash you will really want loads of expertise, since the mob's %parry from the front is higher than %dodge from the back and reducing it to avoid parry bombs on your tank is what you want (since you can't be on every camp 100% behind the mobs).

 

Grung

Being a former rogue i can confirm there was something with hit/dodge mechanics in 2.3

The change was from Weapon skill to expertise rating, which changed things around quite a bit for the rogues who previously had +10 weapon skill from talents.
*EDIT* Was it actually in 2.3 it was changed ? Seems so long ago.

I never had enough weaponskill/expertise on my gear to actually be near the cap, so i didn't research the final conclusion of it.

With this new shaman and the new trinket coming out in 2.4 it all of the sudden became more important to me.

Im sure some of the rogues did in fact research this, and passed wws logs, but so far i cant find the results.

On a totally different note, is there something about the testing of procs and things like that i miss?

It seems most people test on the mobs in blasted lands, but i would think the spirits in Diremaul North would be much better. They don't hit back, and you can hit them in the back for 10+ hours while being afk. Are they special, or do people just use blasted land mobs because they are easy to get to ?

 

Mindrila

Well for me, I just didn't know about those spirits, back before TBC no one wanted to go there with me and now, well no one really cares about old instances anymore. But that's a nice hint and for vast testing I will go there and not blasted lands anymore, sometimes there are some unfriendly people around there messing up the whole testing because they want to kill me...

 

Grung

They are also quite nice to skill all those weapon skills to 350.

Go there, equip weapons, go to work, and when you come home you have 350.

The client tries to log you out since you are "afk" but since you are also in combat it can't log you out, so you end up with a chatlog spammed by blizzard with something like "We will log you out"....."you can't log out while in combat"... repeat.

 

Malan

Here we go, from Nite_Moogle's "Working Theories of Theorycrafting as of 2.3" thread -
I'll add that to the expertise section.

Shitting up every single thread on EJ since '06

 

Yo!

4 hr BT run:
(dodged normal attacks + dodged WF + dodged SS)/(landed normal non-crits + normal crits + non-crit WF + WF crits + non-crit SS + crit SS)
Shaman 1, no expertise
(232+88+16) /(2450+1370++1037+488+617+260) = 5.4 % dodges
Shaman 2, no expertise
(128+56+16)/(1144+836+496+282+252+156) = 6.3% dodges
Roques:
Rogue 1, expertise from talents only
(302+81)/(5762+3049+1206+626)=3.6% + 2.5% (talents only, no additional expertise) = 6.1%
Rogue 2, same story
(252+54)/(4821+2638+1005+520) = 3.4% + 2.5% = 5.9%
Meanwhile rogues stealthed to wowwiki and changed dodge there to 6.5%

Edit: 1) I took white + sisnister strike numbers for rogues
2) trash was included - this brings % down somewhat...

 

◊ Rob

So that's a 5.9% average between 4 trials there, hardly conclusive evidence that we've exceeded 5.6%. Hmm. Who's got a lot of spare time and wants to look through their guild's WWS parses?

 

Atren

I guess only way to find it out is someone with Expertise between those two values to check if gets dodge :P

Concerning rogues, did they have suprise attacks? That makes dodging their finishers impossible if i recall right? Since i can see Combat potency i am guessing they did have it.

Also if this is over 4 hours then it will have lot of thrash mobs and especially akama channelers cant dodge anything i believe and so on :P

EDIT: Added some thoughts :P

 

Malan

What is our allowable rate of error though? I had a parse that I showed Nite_moogle at one point where I had something like a 15% dodge rate on Windfury but it was simply because of the low numbers of procs.

Shitting up every single thread on EJ since '06

 

Ilmatar

Correct me if I'm wrong, but wouldn't trash skew the data? Looking at the specific boss fights though, for example, High Warlord, there were 69 melee hits with 6% of them dodged. There were 34 Windfury attacks, 1% of them dodged. I'm not sure the sample size is large enough to be conclusive.

 

◊ Rob

I don't remember my stats well enough to say how many samples we need to have a 95% confidence that the mean is greater than 5.6%, but I'm sure that such a formula does exist. If someone can chime in that would be helpful.

This would be somewhat useful -- sample size for estimating means:

n is number of samples,
Zc is the critical value for confidence level c (so if we want confidence 95%, we need the Z-value where 95% of the samples are inside the normal distribution,
sigma is s.d.,
E is maximal error tolerance (so this might be like 0.1 if we can tolerate a 10% error)

The formula for "confidence mean is greater than X value" is probably more useful but I don't have my stats textbook on hand.

@ Ilmatar:
Yes, trash would skew the data. I was thinking that Yo summed up the boss numbers, didn't check to see if that was the case.

 

Ilmatar

I just glanced at the numbers, but the hit counts that Yo was summing seemed to be the ones from the log overview, and not from specific bosses.

 

Gehenna

The only way to do this correctly is to add expertise as slowly as we can around where we feel the cap is (post 2.4 of course) and see when we push them off the table.

 

Toots Hepcat

It is difficult to "add expertise as slowly as we can," because it's always in big blocks (10 ER or more) and often in odd quantities. I suppose some braniac could mine all the gear with expertise and set us up with combinations that amount to even "steppings" between 20 and 25 Expertise, and then some poor sap could go farm all said gear. But we don't need it.

If we find enough parses of enough boss fights with no adds and consistent dodge mechanics from non-Orc shaman with no expertise, we can merge them all together in a spreadsheet and figure out the true dodge rate. One column for dodges, another for every other combat result. (I assumed the Rogue theorycrafters had already done this, that their number was accurate and that it would be the same for shaman.)

Quick reminder:
1) A single parse isn't going to give us enough data to make a determination with any certainty -- we all know you can have bad fights or even bad nights where the "dice" aren't on your side. It would be nice to get something impressive, like a hundred thousand swings worth of data.
2) Some mobs have abilities (e.g. spells) which, when up, removes their dodge chance. Any parse used to determine the dodge rate will need to remove these from the dataset.
3) Trash mobs and level <73 boss adds will have a lower dodge chance than bosses. These would need to be removed from the dataset.
4) Windfury and auto attacks should have the same dodge chance, and I believe that Stormstrike as well. This means that results of all attacks can be merged together for the purposes of determining dodge chance.

 

• Chicken

If Stormstrike is comparable to Mutilate I'm pretty certain that only one of the two attacks it makes can fail to connect, or in other words, if the main hand strike hits, the off-hand strike will automatically hit as well. That'd skew the numbers a bit, since you'd get double Stormstrike hits recorded for each succesful Stormstrike, but only one dodge recorded for a dodged Stormstrike.

buff /bʌf/ Pronunciation[buhf]
–verb (used with object)
- to reduce or deaden the force of

 

Kahdrick

Oh believe me, I have talked to them many times. The problem is that my guild isn't big/reliable enough to have a 10-man pro ZA team, so instead we take the first 10 people who fit the raid slots. This often leaves us with 1 good/geared healer and 2 baddies. It's enough to clear, but the rest of us are definitely picking up the slack.

I generally want to play at as pro a level as possible, regardless of the shortcomings of my fellow raiders, so any information that I can bring to the raid with me is helpful. Whether or not faster weapons will regen more mana during SR, for example :P

Shamrogue FTW!

 

Yo!

Checking if a coin is fair - Wikipedia, the free encyclopedia
Assuming that we know definetly that dodge chance is < 10% (therefore E<0.3*Z/sqrt(n), n<0,09*(Z/E)^2):
Total number of proper strikes (white + WF for shamans, white + sinister strikes for rogues) that are needed to determine dodge chance with maximum error of 0,001 (that is 5.7% instead of 5.6% for example) with 95,45% level of confidence = 360,000...
10,000 strikes with same level of confidence provide 0,6% maximum error and that is not acceptable (6% dodge found with 10,000 means it can actually be 5.4% or 6.6%)
100,00 hits provide 0,19% maximum error.
Approximately 13 long (BT) parses with 4-5 dpsers and removed trash and casting (non-dodging) bosses statistics are needed for 360,00 strikes.
Any dodge% reduced by expertise that players have (orcs+axes, gear, talents) has to be added as additional # of dodges for the combined record (if rogue performed 10,000 strikes with 2.5% dodge reduction - add 250 dodges).

 

Grung

Is there anything special that happens around level 70, or would a level 57 tested against level 60 mobs give the same results?

If a level57 could work, we should find someone with a level57 shaman, pop him into DM north, and have him hit the level 60 spirits in there for 24 hours.

 

Veric

Does the Crusade Card's effects both stack at same time? If they do wouldnt that add 120ap and 80sp, now if mental quickness is talented(like it should be) wouldnt that then convery 30% of add ap to more sp? It would make the card add 120ap and 120sp which is a phenomal sp increase, both effects are easily kept up in an FS/ES rotation, and would greatly benefit personal overall dps. The additional SP would increase FS ticks damage dependent to SP at time of cast, while this would take a full minute to reach this effect, it would still be increasingly large during longer fights, please comment back on this. i have not tested this effect yet as im still collecting cards.

 

Illundai

Read.The.Thread.

Or at least search in it. The card is less good than it looks because of our high spell miss rate. One shock resist and your spelldamage stacks will fall off.

 

Nightowl@KJ

Hi guys,
need some help here ;-)

the last days i compared my Dragonstrike & Vanirs Right Fist of Brutality, using Yo!´s simulator several times...

To my surprise Dargonstrike was everytime ~10 dps better than the new fist.

This shocked me a little bit, because after reading this thread i thougt the two new fists, are the best choice
for shamans behind the Season3 weapons...

In my Opinion there are 3 solutions:

1. My gear (ssc,heroic badges) is to bad to support vanirs right fist correctly.

2. I´m too stupid to use Yo´s sim correctly

3. Dragonstrike is really better, but only at fights where you don´t move too much
and are able to take advantage of the haste proc every time...