• Aldriana

So, before I get to the actual theory, let me spell out the background:

Starting at about post 850 in the rogue dps spreadsheet thread (http://elitistjerks.com/showthread.php?t=8806), there's a section wherein the effect of weapon skill on hit chance is investigated. For those of you that want to read over the supporting argumentation, you might read posts 921, 930, 971, and 972 where I post various analysis of the data in question. For those that want to skip all that, what we found is as follows:

1) +1 weapon skill seems to increase your hit chance by .1%
2) The base miss chance for dual-wield against a level 73 is more than 24.6% - in particular, the data we've seen has been largely consistant with a 25.5% base miss chance.

Now, the current theory (as far as I know) is that mobs (usually) have 5 times level in defense - so a level 73 mob has 365 defense, and, in particular, 15 defense more than a level 70 mob. So, if you believe what the character screen says about defense, you'd expect this to reduce the hit chance by .6% beyond the base 24% miss rate. Instead, we notice a miss rate of 1.5% higher than versus a level 70 mob.

So, my immediate thought in reading this is: if +weapon skill gives +.1% hit, might +defense give +.1% enemy miss rate, despite what the character sheet says? This would account for the level 73 base miss rate correctly, and there's a certain logic to it.

As a first place to look to confirm or disprove this theory, I watched the miss rates of bosses attacking our tanks on a selection of fights. Alas, the sample size was inadequate to be conclusive.

However, tonight, I ran a full Heroic Underbog tonight, and parsed it with Recap starting just after the first boss. And over the course of the remainder of the run, Recap recorded 1087 attacks made against the tank, of which 189 missed, for a miss rate of 17.4%. The tank in question had 493 defense. Now, the mobs in heroic Underbog are all level 70 and up, so, logically speaking their base miss rate should be 5% or less, and, according to the character sheet, 493 defense should reduce their hit rate by 5.72%, so they should be missing 10.72% of the time or less (as some of them are level 71 or 72).

Now, if the miss rate is, in fact, 10.72% or under, then the 17.4% observed rate is roughly 8 standard deviations from the mean. Aka, totally impossible. So clearly one of three things is true:

1) The base miss rate of monsters is higher than 5%, or
2) Defense increases miss rate by more than .04% per point, or
3) There's some other factor reducing miss rate.

Now, I can't think of anything for #3 (although if someone has an idea, I'm all ears), and #1 would require more testing (notably, someone with a significantly different defense parsing a similar run. In particular, a feral druid tank, who can become uncritable at a much lower defense value, might be a good person to take data). However, lets look at #2, in particular, the theorized .1% miss rate per point of defense. Well, this would cause the base miss rate to be 5% + 143*(.1) = 19.3%. Since some mobs are higher level, they would have a base miss chance up to 1% lower - hence, the "average" miss rate would be somewhere between 18.3% and 19.3%. The observed rate was 17.4%, which, depending on which set of numbers you use, is somewhere between 1 and 1.5 standard deviations from the expecation - well within statistically accepted bounds.

Now, this isn't proof, of course - as mentioned, more testing (at a different defense value) would be needed to be sure. But it certainly looks possible that, despite what the character screen says, +defense gives +.1% enemy miss, much as +weapon skill gives +.1% hit.

 

Morsexy

tomorrow when i do a couple of heroics or 25 man, ill keep a log and a copy.

 

Boevis

As luck would have it, I'm currently running tests vs 71-73 mobs in the flight only areas to determine the base miss for druids (it's ~5%) As well as trying to confirm the effect of skill, thought it would seem you've already done that, .1 per point fits considering my relatively small samples.

I'll just throw on some tanking gear and kill 2 birds with 1 stone.

 

Tzan

Do any of the mobs dual-wield and could that be accounting for some of the difference?

 

Muggins

Heroic Underbog does have a dual wielding type mob, the name escapes me(beserkers?) but they're in the Broken mob packs just before the second boss. So it may very well be these that affected the sample.

 

• vpolden

Interesting theory. I'll run some tests in instances myself later. As for the dual wielding mobs, is there a generally accepted theory around right now about their missrate?

It is better to keep your mouth closed and let people think you are a fool than to open it and remove all doubt.

 

• Apate

Evidence is currently leaning towards DW mobs having similar missrates to DW players. I don't recall how close the missrates were, but they didn't appear to be far off from the testing that I recall.

See you, auntie.
"You don't need a machine to make a rainbow. For rainbows are made of happy thoughts, and dreams, and chocolate unicorns, and gumdrops, and licorice sunsets, and fuzzy gumdrop bears, and sugar-coated chocolate gumdrop land."

Originally Posted by DeeNogger I am coming for you Apate.

 

• Aldriana

I can imagine this having an effect, but over a sample encompassing 3/4 of the instance I wouldn't expect it to be a major one. Assuming the DW mobs make up, say, 10% of the attacks and have a base 24% miss chance, the overall expected miss rate would still only be about 12.5%, well under the observed value.

However, I admit that my data isn't the ideal set to be testing on due to the diversity of mob types and whatever. The best test would involve either a) a tank parsing all his boss fights except Prince, Moroes, and any other DW bosses for a couple of nights, or b) A very patient tank, healer, and hunter.

 

Athinira

Mobs 1-3 level above you get, in addition to the 0.2% miss from the extra 5 defense per level, 1% less chance to be hit. In other words it looks this way:
Level 70 mob: 5% chance to be missed
Level 71 mob: 5% + 1% for level difference + 0.2% for defense difference
Level 72 mob: 5% + 2% for level difference + 0.4% for defense difference
Level 73 mob: 5% + 3% for level difference + 0.6% for defense difference

I sadly can't link this since its quite old data, but it was proved way back when Molten Core (and BWL) was endgame, both a hunter and warrior tested it, the hunter was able to achieve 100% hitchance when going from +8% hit to +9%, so the only logical conclusion from that is that mobs 1-3 levels above you gain an additional 1% chance to be missed per level. I believe this misschance scales exponentially if the mob is 4 or more levels above you.

Moderator and Organizer on The Druid Wiki
http://druid.wikispaces.com
The Druid Wiki is currently outdated and is scheduled for a major WotLK overhaul. If you are looking for information on druids, i would suggest browsing these forums for now.

 

◊ Kalman

Athinira: that data was reliable for *special attacks* (and possibly for 2H auto). For DW autoattacks, no 1% per level missrate was observed in any of the data I've seen; a 1% per level would imply a 27.6% base miss rate.

http://mmorchive.net

The WoW forums, explained:

 

sp00n

That's interesting, since I can only remember that 8.6% miss chance for special attacks was brought up when I posted that screenshot fighting Vem with +8% hit and missing a backstab.

That knowledge seems to have avoided this forum for a long, long time then.

Item Ranking Rogue [horribly outdated]

 

Athinira

This might be true. Remember that we are talking data over 1 year old, which means that while they are most likely are still accurate, my memory might not be. I believe, however, the warrior was using a 2h = No DW test results.

The tests back then was run over a long period btw, we were talking a hunter running the mod through 2-3 MC clears with 8% hit, then 2-3 clears with 9% hit, so it was a pretty large data sample on his part. I remember the hunter result for ranged attacks though, with 8% hit he had a total of 216 misses over the course of those 2-3 runs, 0 misses with 9% hit on both autoshot and specials. Warrior had 6% hit and was sporting around 2.5% miss, can't remember if it was white attacks only or white and specials.

Moderator and Organizer on The Druid Wiki
http://druid.wikispaces.com
The Druid Wiki is currently outdated and is scheduled for a major WotLK overhaul. If you are looking for information on druids, i would suggest browsing these forums for now.

 

• Vandemar

I managed to read this thread just after tanking a Heroic Underbog run on my feral druid alt (413.3 defense, 152 rating, with 4 resilience to eliminate the remaining crit chance on 73s)- I had reset Recap about 2 pulls into the run, after forgetting about it at the beginning. Incoming melee on the tank was as follows:

758 hits, 140 Misses, 447 dodges (I was seeing about a 37% tooltip dodge chance with BoK and GoA, for the most part), and 0 crushes / crits / parries / blocks. Assuming 1 def = .1% incoming Miss chance and 5.0% base Miss, the predicted value would be 11.3% Miss against lvl 70 mobs, 10.8% against 71's, and 10.3% against 72's, and I recorded a 10.4% rate against mobs ranging from 70 to 72, with a sample size of 1345 attacks.

 

• hiwrac

Did you take into account the extra miss rating from defense?

I also believe the number of attacks they do could be quite a bit higher than 10% for several reasons.

-) They are often killed after the rest of the pull is dead, giving them more times to get hits in.
-) They hit fast.
-) The trash from the 2nd to 3rd boss spends a lot of time casting spells such as acid spit instead of hitting the tank.

 

• hiwrac

Update!

I sent out to find out if DW mobs do infact suffer from the same miss penalty that players do. I did a simple, relatively controlled test. All the mobs are the same level and same type, all DW and have no special attacks or offensive spells.

Mob: Demon Hunter Supplicant (located in large groups near Black Temple), Level 70.
Player: 492 defense (additional 5.68% chance to miss).
Projected Miss Rate: 24+5.68 = 29.68%
Actual Miss Rate: 28.5% (2,763 samples)

This seems to confirm that at least some DW mobs suffer the full miss rate. It also doesn't seem to support the first poster's claim of increased miss rate for defense, at least not for level 73 mobs.

I am going to start watching this more closely in the future to see if I can get a more definite answer for mobs 71+.

 

IceBox

As stated on WoWWiki a Mob with 365 defense would have a 1.5% higher chance that your attacks miss than one with 350 defense. But I don't know if this is still correct.

 

• Aldriana

Yes.

All true. But the point remains: it's still just a couple of mobs spread across a full instance clear. They certainly don't make up 10% of the mobs we fought, nor did they take 10% of the time. Regardless, they'd need to make up, like, 30% of the incoming hits for this to be the answer, and I seriously seriously doubt that that's the case.

 

Boevis

I don't see how this proves anything by itself. For all we know they could have 14.3% base miss and your +142 defense increased this to the 28.5% you observed.

A few hours of fighting 73's in Blades edge with +24 Def (.96% increased miss according to tooltip) So I should see 5.96% Miss minus their contribution from +Hit. 4.46% At .1%, 4.76% at .08%, and 5.36% at .04% per skill.

479 Hits, 8 Crits, 42 Misses, 419 Dodged. 4.43% Missed. That matched almost perfectly with the OPs .1% per skill but within probability of others.

 

• Aldriana

I think you misread the intent of my post. It's been pretty well confirmed that +weapon skill is .1% hit per point; the purpose of this post was to assert that it's .1% miss chance per point of defense.

Hence, a level 73 mob should have a base 3.5% miss rate against you, increased by .04% (if the tooltip is right) or .1% (if the tooltip is wrong) per point of defense, for an end miss rate of either 4.46% or 5.9% - both of which are within statistical bounds of your data.

To get more accurate data on this would ideally involve more defense and larger data sets.

 

• hiwrac

Point taken. Will return later today with no +def and test again for a control.

 

• Cluey

I haven't had a chance to test this at 70 but as one of the people listing 8.6% miss rates I thought I should chime in.

When this was last brought up the thread started with lots of people quoting the old data of 5.6%, as the thread progressed we ended up with people actually testing a bit culminating with sp00n's screen shot of a backstab miss on Vem with 8% hit.
Further on in that thread was a rogue who got to 25% hit and never missed anything.

From that I think you concluded there was no additional 1% for auto attacks with additional mob levels, I see in this post you have a possibly for 2H auto in there too.

I am expecting dual wield auto attacks don't get the additional 1% per level but other attacks do, it fits the data better than a difference between specials and auto attacks.

My testing was done on my druid in MC after the 1.10 patch. I did the entire run with 8% to hit, the attached screen shot is the total. The misses were from six auto attacks and one shred (cat backstab).
I ended up testing this as before the 1.10 patch I used to run with 6% and not get any misses, I had changed a few items around and lost 1% hit and then the 1.10 patch came and the numbers just didn't add up so I tested it more. To get to 8% hit I had to use the 2% hit food from one of the holiday events and some of the gear I had upgraded from.

Attached Thumbnails