sp00n

So, just letting you know, I'll be gone for a week now (no internet /cry), and I expect an answer by the time I return.


And I'd really strongly suggest mace specialization for further testing weapon skill.
Weapon skill rating is just a mess to test, a) we don't get even numbers with it (205 rating would be the first round number I think) and b) we don't know if fractional weapon skill does anything at all, or if its rounded up/down, or taken as is.

 

Mace

As a tank I'm pretty interested in the parry reduce part of weapon skill. (ASsuming a mob recieves the same parry attack speed bonus as a normal player, thus reducing spike damage)

the "famous" blue post:
The amount of parry reduction (1,6% for 4 skill points) took my interest.
I assume a mob starts of with a default 5% parry like everyone else.

Tests previously done tells something way different tho:

Therasiv should match the mobs amount of defense with his weaponskill, giving no extra parry for the mob.

So that mob had A) a higher base parry % (contribution from parry rating / contribution from talents)
or B) 512.7 defense to match the "defense skill-weapon skill*0.04"

It cant be B otherwise both were impossible to land crits on that mob. (mob would be crush immune at 490 for Beaglej)

That makes me wonder if the numbers of this test are actually representable.

 

Oscarvil

From what I recall of that thread didn't they find that weapon skill gave no bonus to -parry but double the bonus to -dodge? This effectively combines the -mitigation effect into a nonpositional stat which benefits everyone for damage, but does nothing for the tank trying to prevent thrash.

If I'm wrong point it out so I'm not just sitting here spreading misinformation.

 

• frmorrison

WWS, but it may be messing things up, idk.


Pets have a normal 11% miss rate (howver, that may be adding dodge in) on same level targets, and the extra 15 defense on the boss increases the miss rate reportedto 15.6%.


Here is a WWS from a BM hunter on Void Reaver Kodra - WWS with about a 15.2% miss rate.

Another WWS from another BM hunter same fight Tientzo - WWS with about a 16% miss rate.

I don't know if they have the 4% to hit talent for pets.

 

Wodahs

Thats an incorrect interpretation of WWS frmorrison. Click the config button and under the miss column, checkmark % and Nb. This will now show a column with the actual number of missed, and percentage. You are looking at the all missed aka failed to land due to miss, dodge, block, parry column.

Kodra's pet missed melee 5%, kc 2%, claw 5%, bite 6%.
Tientzo's pet missed melee 4%, claw 6%, kc 4%, bite 12%

 

Olgas

I'm almost certain that your 11% figure against equal level mobs must be adding in Dodge.

Recap, if you look at the details, will further break out the "misses" into dodges, parries, blocks, etc. So, you can discern the actual miss rate if you look at the detail.

I am a BM hunter and I do have the talent that gives +4% to hit.

I haven't been tracking my misses with my pet carefully enough up until to now. But if you read back in this thread a little bit, you'll see that my actual miss rate with my pet has (on the couple runs where I was paying attention) been about 4%. Meaning that my pet's implied miss rate is about 8% against Level 73s.

I actually think that with further testing, though, my pet's miss rate will turn out to be closer to 9%, consistent with the other results we have seen for 2H melee. But the testing hasn't quite shown us that, yet

In any event, I'm almost certain 11% against an equal level target isn't accurate. I believe (and at this point it's still only a belief) that a pet's base miss rate is the same for 2H melee:

v. Level 70: 5%
v. Level 71: 5.5%
v. Level 72: 6%
v. Level 73: ~9%

 

Olgas

These are both pretty small sample sets, but they are directionally informative.

According to Armory, both Kodra and Teintzo have 2 talent points in Animal Handler, which means they both have +4% hit for their pet.

Kodra's pet did a total of 223 normal melee attacks and saw 13 misses (5.8%).
Teintzo's pet did a total of 270 normal melee attacks and saw 14 misses (5.2%).

Add back in the 4% from Animal Handler and you see a base miss rate of 9.8% and 9.2%, respectively. But again these sample sets are very small, so these results are likely consistent with a 9% base miss rate.

 

Olgas

Hey, something just occurred to me.

I don't think we've actually seen a dual-wielding test yet, with exactly 363 hit rating. This would be a great next test -- if it demonstrates no misses at all -- that would go a long way towards confirming the 28% theory.

 

Dontmindme

I believe I have a good way to test whether fractional weapon skill applies or not, thus opening up more configurations to testing.

For openers, we know that the character screen (and subsequent hit percentages) round all fractions down with weapon skill rating. Thus, there are 2 main schools of thought:
1) That's just the character screen, but fractions probably still apply. OR
2) All fractions are dropped.

Now, the easiest way to accurately test hit ratings seems to be Haluum and determining the hit rating needed not to ever miss. So I ran a few numbers with 360 and WEx at various weapon skill ratings...

I found that with 15 weapon skill rating [15/(2.5*82/52) or 3.8049 weapon skill], you get the following:
If weapon skill rating fractions apply, you would need 278 hit rating to theoretically not miss. If fractions are dropped, you should miss 0.0707% of the time and you would need 280 hit rating to never miss.

15 weapon skill rating is pretty easy to achieve. Grips of Deftness or Whispering Blade of Slaying, for example. I hope to be able to test this on the weekend, but if anyone has the free time and inclination before then, go for it. Probably need at least 10000 swings without a miss to be reasonably confident at 278, but any miss implies fractions are indeed dropped.

 

Dontmindme

Well, I did a quick test before going to bed.
Dual-wielding daggers, Precision, WEx and 15 Weapon Skill rating from Claw of the Netherwing Flight (Yields 263 weapon skill on the tooltip), 278 hit rating.
278*10*82/50 = 17.6293 + 5 + 1 = 23.6293%
If weapon skill fractions count = 23.6293 + 0.3805 = 24.0098% to hit
If not = 23.6293 + 0.3 = 23.9293% to hit
vs. Haalum level 70 from behind (theoretical 24.0% miss rate)

857 attacks against Haalum
856 immunes
1 miss (so I stopped)

Seems to indicate that fractional weapon skill from excess rating gets dropped.

When I next find time I plan to try 279 hit rating (which still theoretically should miss if fractions are dropped), followed by a large number of swings at 280.

 

Olgas

This is a great test you're doing. Really smart approach. If you can demonstrate that the fractional Weapon Skills get dropped, then it really does open up a lot of other testing capability.

One question, why do you want to the test again at 279? If I did my math right, 279 should give you 23.9927% +hit. Since that is so very close to 24%, I worry that you might have to do a very long test to see the misses that you predict. And I'm not sure that it'd tell us that much more than your 278 test has already shown us.

I think I might jump right into the test at 280. I agree that you'll want to do a lot of swings at 280 to prove you doesn't see misses there.

Again, great test -- can't wait to hear the results.

 

stampy

Might want to keep an eye on the possibility of rounding fractions as well as truncating. I have a couple of rather ad-hoc spreadsheets for comparing gear and enchants, and the first thing it does is grind out the base stats.

Pure base hit points (eg, how many hit points I would have if my stamina was 0 and I didn't have the tauren racial) had some interesting results. Lots of modifiers (vitality, kings, tauren) applied to a large health pool made errors really stand out, and flooring underestimated my total health when I added gear numbers in; while rounding made the spreadsheet match up to in-game numbers.

It was something I just threw together in a half hour or so to play with resist gear stats, so I'm not professing that its solid fact. But, I do think it was pretty accurate -- and its probably worth keeping the possibility of rounding the intermediate numbers in mind, as well as flooring or keeping decimals.

If I had to guess, the rating is converted into the skill number, rounded, and the integer skill number is used to modify the combat tables. Once that integer is converted to miss or dodge percent adjustments, it is likely in full decimal glory (basis: there used to be blizz-standard script commands to return dodge/parry/block percentages, and they would spit out 10+ decimal places). But then again, guessing sucks.

 

Olgas

The great thing about the test Dontmindme is doing, is that it should distinguish between whether the fractional skill is rounded or dropped.

The +15 weapon skill rating he is adding should add 3.805 to his Weapon Skill. Which, if rounded, would round to +4 to his Weapon Skill. But on his test, it seemed to behave as though it was only adding only +3 to his Weapon Skill.

We probably need to do a few more tests at different weapon skill ratings to confirm, but it's very helpful knowledge.

 

Dontmindme

To be perfectly honest, I didn't expect to see any misses at 278 because I thought fractions would carry through. My guess was obviously wrong, but I guess that's why we test it.

Well, the truth is that showing a miss at 278 really says a lot. It seems to eliminate weapon skill rating as is (carrying fractions) or rounding up leaving only the possibility that all fractions just drop as in the tooltip. I'm not sure what testing at 280 really proves at this point. I can think of no reason we would see a miss, thus it would be primarily academic.

Showing a miss at 279 would at least have the benefit of narrowing down the hit rating to the predicted value for all fractions dropped and if I don't get lucky and have an early miss, the few thousand immunes before the miss will show that the miss chance is very small and consistent with expected.

I'm thinking I may try other weapon ratings first, before going through with the 280 testing, unless I get an early miss at 279.

 

Olgas

Yeah, I hear you. And I think your reasoning is sound. The only reason I think the test at 280 might be useful is just to confirm that we do NOT see a miss at that point. If we were to see a miss, then it would mean we made a mistake earlier somewhere, in part of our assumptions or analysis. I don't think we did, though.

But totally understand if you want to test other stuff first. Just grateful that your sharing the results. Thanks.

 

Dontmindme

Well, tried 279 for a bit, same parameters as above.
5303 swings
5303 immunes
But given the 1 in 13666, 0.007073% expected chance to miss, this doesn't say much.
I also had another 2500 or so attacks, no misses (or so said swstats) but I apparently forgot to turn the combat log back on which made getting an exact number on just Haalum impossible.

 

sp00n

Disclaimer: I know I said I'd be away for a week, but we had trouble with our car so we postponed this for 2 days and are leaving tonight.

So I just tested against Haalum as well.
Wearing [Shoulderpads of the Stranger] with 10 additional skill to daggers and wielding 2 1.3 speed daggers I went up to him.
I had 280 hit rating, 2 WEx and 10 skill rating.

10 skill rating are 10/(2.5*82/52)= 2.537 weapon skill.


(280/(10*82/52))+(10*0.1)+((10/(2.5*82/52))*0.1)+5= 24.00976
With the fractions counting I should have never missed.

(280/(10*82/52))+(10*0.1)+(2*0.1)+5= 23.956
And with weapon skill rating being floored I would have had a very small chance to miss.

And I did miss (Screenshot). Seems I had a lucky roll, since theoretically it would have taken me about 10000 attacks.


On the other hand, I am not so sure that weapon skill *above* the defense skill yields the same benefits as when being below the defense skill. Remember, Haalum is level 70 with a presumed defense of 350, and I was having 362(.5) weapon skill.

It *could* be that weapon skill is offering less or more hit to such targets, just as it is offering a greater benefit when the difference from weapon skill to defense is larger than 10.

To test this, I engaged him again, this time without the 10 additional skill rating and with 283 hit rating. If WEx gives 1% to hit also against level 70 targets, I should have seen a miss with 283 hit rating, but not with 284.
(24-5-1)*(10*82/52)= 283.846

I did miss with 283 hit rating: Screenshot
So we can be pretty sure that weapon skill does not give more hit rating than 0.1% (resp. at least not more than 0.10537%) against a mob with 10 defense less than weapon skill.

As for 284 hit rating, I'm currently working on this, I should see no misses, but it could go on forever so at some point I'm going to abort it.

 

sp00n

4173 attacks and no misses, I've stopped now.
So it seems that with 10 more weapon skill, it is still at 0.1% hit per point of skill.
If this holds true, when attacking a level 57 Servant of Razelikh you would need 24-5-(360-(57*5))*0.1)= 11.5 hit or 181.346 hit rating to never miss.

So being in testing mood, I've gone out to do so.
With 181 hit rating, I saw a miss: Screenshot

And another 4000 hits later with 182 hit rating, no misses. I stopped although it was going *so much* faster than against Haalum, because of SnD.

Well, I'm pretty sure now that weapon skill does NOT change its effect on hit chance when you have equal or more weapon skill than the target's defense skill.
Therefore we can also be pretty sure that fractional weapon skill is dropped, i.e. weapon skill is floored to the previous natural number.


Also I'm 100% convinced now that the miss rate is solely dependent on the weapon skill to defense ratio.


// Edit
Just in case somebody wants to see the logs:

Haalum

280 Hit, 362 Skill
Immune: 1846
Miss: 1

283 Hit, 360 Skill
Immune: 1979
Miss: 1

284 Hit, 360 Skill
Immune: 4173
Miss: 0


Servant of Razelikh

181 Hit, 360 Skill
Hit: 463
Crit: 174
Miss: 1
Total: 611

182 Hit, 360 Skill
Hit: 3047
Crit: 1097
Miss: 0
Total: 4144

 

Dontmindme

Yes, I double-checked your math. Theoretically, you are correct. You should not have missed if fractional weapon skill applied, yet you did have a miss. The only correction I have is that your miss rate is about 1 in 2278 for that test not 1 in 10000. Thus, seeing a miss in 1847 attempts is much more plausible.

So, now we have 2 tests that indicate that fractional weapon skill rating gets ignored.

 

Fugazor

Is not ~9% miss for 2h/hunters/druids incorrect? I checked WWS for miss on bosses (white attacks only) and compared it to hit in armory.

2h retardin with 4.57% hit, 350 ws:
- VR: 108 attacks, 0 miss
- Solarian: 220 attacks, 3 miss
- Al'ar: 260 attacks, 1 miss
- Hydross: 228 attacks, 2 miss
- Lurker: 177 attacks, 3 miss
- Morogrim: 186 attacks, 6 miss
- Fathom: 116 attacks, 0 miss

Overall 1295 attacks, 15 misses -> 1.16% miss. 4.57 + 1.16 < 6%. Of course some fights had adds so real miss is probably ~1% higher. But it fits in theory that 2h miss rate is around 6.0-6.5% (25.5 - 19 as it was 27.6 - 19 before TBC) not 8.6%.

I also did a quick check for a 2h warrior with 6.98% hit and 354 ws and could not find any fight where he was dpsing and missed anything.

Basically, unless WWS data is incorrect, we all need ~2% less hit in TBC (27.6 and 8.6 before, 25-25.5 and 6-6.5 now).

 

Olgas

I'm sorry, but this isn't correct. If you read back through this thread a bit. You'll see some pretty good testing has been done on this over the last couple weeks.

There is a good body of evidence to support the conclusion that the base miss rate against different level mobs is as follows:

v. Level 70: 5.0% (24.0%)
v. Level 71: 5.5% (24.5%)
v. Level 72: 6.0% (25.0%)
and
v. Level 73: ~9.0% (~28%)

These have been controlled tests, that have, so far, validated these numbers under different conditions and for different classes.

As for the Level 73 number, we haven't definitively established the 9% and 28% figures as the correct numbers, but we have multiple points of evidence that it is higher than 8.6%/27.6%. And, indeed, we have evidence of Rogues still missing with +hit gear in the 27.8% ranges. So, we think the number is close to 9%/28%, and we are continuing testing on this to nail down the exact number.

 

Dontmindme

It should be noted that if you have at least 5 additional weapon skill the miss rate seems to become much lower. Are you sure the armory gear is what they used in the raid? Did you have a Dranei aura? Did you have Improved Faerie Fire up? Did they buff with a Fire-Toasted Bun?

I'm not sure you can really take a WWS log and compare it to gear as listed in the armory and consider this accurate. Now, those doing the tests are paying close attention to what buffs they are getting to be sure they aren't getting some other special effect.

 

Olgas

OK, ran Gruul's tonight with 8.624% +hit (136 hit) and no +Weapon Skill.

As expected, this test confirmed that the base miss rate for Ranged DPS against a Level 73 is greater than 8.6%.

Here were my results for the evening:

Auto Shot: 441 hits, 127 crits, 3 misses (.555% miss)
Steady Shot: 422 hits, 130 crits, 4 misses (.719% miss)
Arcane Shot: 8 hits, 0 crits, 0 misses (0% miss)

Total Shots: 1105 Total Misses: 7 (.633% miss)

So far, Ranged DPS is exhibiting the exact same characteristics as melee DPS. I don't know how to do that 95% confidence range calculation, but it seems to me like this test again supports a base miss rate of around 9% (9.23%)

Even though this sample size was not that big (only 1105 shots), I don't feel the need to repeat it. The important thing is that there were misses at all.

______________________

Meanwhile, I thought I'd also share the results of my pet attacks from tonight's Gruul's run:

Melee: 1339 attacks, 58 misses (4.332%)
Specials (all combined): 919 attacks, 42 misses (4.570%)

Total for Pet: 2258 attacks, 100 misses (4.429%)

As a reminder, my pet has +4% hit from talents. So this test suggests a 8.429% miss rate for the pet, which is a bit lower than our expected 9%.

______________________

For my next test, I'm going to jump up to exactly 9% (142 hit) and see if I see misses at that point. Unfortunately, I won't get the chance to do this in a raid until Monday...

 

gatz

Yesterday I noticed a miss in scrolling combat text with 85 hit rating and 309 weapon skill while hitting with my 2 hander on a boss.

Today I've replaced a 8 crit with a 8 hit gemm and ran combat log while doing 3 attempts and a kill on Void Reaver.

359 Weapon skill
93 hit rating (5.88%)

auto attack : 131 hits, 90 glancing, 88 crits
slam : 152 hits, 83 crits
mortal strike : 145 hits, 60 crits
whirlwind : 83 hits, 31 crits
0 misses
(not counting executes because I switch to DW for executes)

863 attacks ; 0 miss



last week I ran combat log with 95 hit rating on an attempt and a kill on Lurker and a kill on Morogrim.

359 Weapon skill
95 hit rating

auto attack : 107 hits, 61 glancing, 95 crits
slam : 76 hits, 30 crits
mortal strike : 84 hits, 43 crits
whirlwind : 62 hits, 24 crits

582 attacks (some on non lvl 73 adds) ; 0 miss

 

Karmon

309 weapon skill? I will assume 359 as in the following 2 tests.

This findigs are consitent with the proposed theory:
1) 359 skill, 85 hitratiing : skilldifference ist <= 10 so
5+(365-359)*0.1 = 5.6% missrate
85 hitrating equals 5.39%hit so you have a 0.21% residual miss rate

With 89 hitrating (5.64% hit) you should not see any misses

The 5.6% basemiss from above is the same
93 hitrating equal 5.89% hit
95 hitrating equals 6.02% hit

Both are higher than the 5.6 base miss so you should not see any misses.

Interesting would be number where your weaponskill is between 350 and 354.