If I understand what's going on here, they need to understand those 2-8 pieces of it before they can understand the other 71 pieces. Those 2-8 pieces are the keystone to understanding the other 71. Trying to analyze all 71 at once is problematic and would be like trying to take all 71 pieces and just mash them together at once, instead of placing the puzzle together piece by piece.

There's multiple factors that go into reporting something like this to blizz. One is that the original guy who wrote the code may not be there anymore and his documentation may or may not be very thorough. Two, given that nothing seems to be proven right or wrong why would blizz spend hours upon hours of programmers time on a problem that might not exist (if it ain't broke, don't fix it). Right now this crit reduction on focused aim issue is just something someone noticed and hasn't really yet been proven to occur and if it does indeed occur may be the way the mechanics were meant to work in the first place and our understanding of the details has been flawed. Three, they keep the details away from us on purpose. It's a recreational game and in some ways I think all the min/max'ing that has gone on over the course of the game has marginalized some of its fun and marginalizes their fun in making it.

What they're doing is eliminating variables to try and understand what happens behind the scenes. It's like trying to determine your mileage in your car- if you record a number and want to see whether having the trunk full of dumbbells affects it, don't go changing the tires and switch from city to highway driving while doing that test.

More or less what I am describing, except with roll 1 and 2 bundled into one roll as Hit, Miss, or crit, with Hit rating pushing misses off the table. The final intent of this testing is to see if the hit from Focused Aim replaces a miss after it occurs, thus never allowing it to be a crit. The block mechanic is something we stumbled on as a factor that could be interfering with the original intent of the testing, so we decided to catalog it in order to set a control.

Do you have a copy of that screen shot with the blocked critical, since I have yet to see it occur. I am interested to see where the block shows up and if it is a true block, or just block mitigation.

This is in essence what I am describing with miss being the factor that is pushed of the table.


1: It is already established that equal level miss chance is 5 percent regardless of level. This occurs at all levels regardless of spec, talents (unless they effect hit), and gear.


2: The reason we "specced down" is because we originally found this occurring in fully specced tests, and we eliminated factors in order to come to these conclusion. First Lethal shots was eliminated, and now we are seeing if Focused Aim has an effect. The purpose of these low spec test is to set a control to test other factors against. Through the observation of WWS logs, we can find that block damage mitigation occurs on all physical attacks from the front, regardless of spec or class.

3: This seems to be a mechanic that occurs on all physical damage attacks, including our specials, not just our white attacks. If you parse any WWS report it shows all of our physical based shots can be block mitigated, and some can have block mitigation and resists. I think if we take this to blizzard, we will probably get the "working as intended" response, as attacking from the back simply bypasses this all together. I do find it interesting that tanking classes suffer from this block mitigated damage as well. It is something I will eventually bring up, but i want to have solid evidence behind it before I make a push. Right now i do not feel it is necessarily a broken mechanic, but just one we have not become fully aware of as of yet.

As for crit rising at the beginning of a test, this is something that occurs in both directions. At the beginning of a test you may get a long crit string that results in a return weighing in favor of higher crit, since one or two extra crits can swing a 100 shot cycle by 2 percent. The drop in crit to a more established level is a normalization of crit return, due to a larger sample averaging out high crit strings with low crit strings. In my test I observed some test starting low instead of high, and then the crit graduating upwards as it normalized. I even pointed out a low crit outlier sample, where I just got a lot of low crits at the beginning, and the normal crit returns afterward were just not enough to get me back to within the normal standard.

This is the essence of averages and where RNG comes into play. Sometimes it likes you, sometimes it doesn't, and in small sample sets, it can have a rather vast swing one way or the other. Anyone who PVPs a lot has had this help them and frustrated them, depending on which side of the coin toss they landed.

For example, if we do find that Focused Aim also contributes to Crit reduction, the best answer we will get is "use focused Aim to get hit until you can get that hit from gear, and then spec out of it." There is already an In Game solution to this (if we prove it to occur), and that is probably what is expected to be done by us.

Focused Aim doesn't show up on the character sheet or transfer to pets, but it has already been stated as "working as intended." It is very much possible that the additional side effect that we find may or may not be occurring is also covered by this statement.

Thanks for bringing me up to speed. I'm probably going to be respecing for next weeks raids to try out a surv build, I'll see if I can't get you some shot data before I do that, remind me: as close to an even % crit as possible and above 8% hit of all auto attacks, correct?

Even doesn't matter so much, as long as you keep it consistent in tests.

I wonder if you could get that log thrown up on WWS reports to see if it catalogs that as a direct Block, or Block mitigated damage? I get immune a lot on my Explosive shots, and I found it can be block damage mitigated and resisted at the same time.

Ah, it is on WWS, sorry, I forgot to link: Wow Web Stats

I ran some quick tests on Explosive Shot this morning on the dummy, but I did not have time to edit down my log to show only Explosive Shot.

I did see lines that shows both (critical) and (resist X value), as well as normal hits with blocked damage, but I didn't see one as an example like you posted.

After parsing the log, I found several instances of this occurring, on both explosive shot and Auto shot, but I did not find it occurring on auto shot. Furthermore, the blocked value shots were NOT negating crit on specials. This would point to the mechanic we are finding only effecting Auto shot. I will note that WWS still does not record any of these as true blocks, but only block mitigated damage. When our guild gets our WWS up, I will check it out as well, and I will also do some explosive shot test runs on the dummy to try to recreate this event.

I would have discovered this earlier, but it took me a bit to figure out how to use the new combat log scan tool, since it no longer has a pull down menu. Turns our if you do "spell=explosive shot" you get a syntax error, and you need to do "spell=explosive". To parse a log you input as follows, with "hunter" meaning the name of your hunter, and "shot" meaning the name of the shot you are looking for, but only the first identifier of the the shot i.e steady, explosive, auto, etc.:

source=hunter and spell=shot



Since the attack was on a gargoyle, and the block value was 70, it seems that block value goes up by 3 block points per mob level, i.e. 67 at level 80, 70 at level 81, 73 at level 82 (assumed), 76 at level 83.

I did find instances of different block values though. Death knight cavaliers always blocked 69 damage, and Gothik blocked 74 in one instance of block I saw. This only occurred on Steady shot.

There very well could be a different system for Auto shot, or it could actually be that Auto Shot block mitigation is not functioning properly.

You can use the entire name of the ability, just put quotation marks around the word.

E.g:
source=Thayer and spell="Explosive Shot"

Reading through this thread, I noticed some people were using cheap ammo to save costs. I did some testing a while back and cheap ammo actually does damage to the ranged weapon. If you use level-appropriate ammo, it may actually cost you less overall as the weapon will take little or even no durability damage.

I get different amounts of weapon damage per 1000 shots, regardless of the ammo I am using. I think it may just be a chance thing.

Sorry about the lack of info lately, but i've had to get some work done. I should be able to get some more testing done on Sunday.

Unless you have some paradigm changing data for us to look at, I'm going to say that this is categorically untrue.

Here is a WWS report from my patchwerk kill this weekend. Wow Web Stats

I stood behind him as I tired to trap dance. I got 5 LNL procs in a 4 min fight, so I failed a bit at that, but I had no blocked damage on any shots, only resists.

I have begun my test on the level 83 dummy from the front, with the highest crit I can attain. I reached 46.22 crit through spec and gear, though I am using a Elixir of Accuracy (45 hit rating, and I bought twenty of them), along with some regemming to sit at 264 hit rating as well. If the results go too far off the expected scale, I can also run tests at about 44 percent crit, and full hit from gear, in case the Elixir is a factor.

This is my crit heavy, but utterly useless build. I avoided anything useful except talents that would boost my crit.

http://talent.mmo-champion.com/?hunt...h=000000000000

I will post my results here, doing 12 sets of 1000 shot tests.





46.22 crit rating, 264 Hit rating (0/0 FA, 45 hit from Elixir), level 83 Dummy, from the front

01: 38.1 crit return --- 8.12 crit loss --- 8.12 average crit loss --- 45 blocks (4.5%)
02: 38.4 crit return --- 7.82 crit loss --- 7.97 average crit loss --- 60 blocks (6.0%)
03: 39.1 crit return --- 7.12 crit loss --- 7.69 average crit loss --- 46 blocks (4.6%)
04: 37.3 crit return --- 8.92 crit loss --- 7.99 average crit loss --- 53 blocks (5.3%)
05: 40.6 crit return --- 5.62 crit loss --- 7.52 average crit loss --- 45 blocks (4.5%)

06: 39.8 crit return --- 6.42 crit loss --- 7.34 average crit loss --- 52 blocks (5.2%)
07: 38.9 crit return --- 7.32 crit loss --- 7.33 average crit loss --- 52 blocks (5.2%)
08: 37.0 crit return --- 9.22 crit loss --- 7.57 average crit loss --- 64 blocks (6.4%)
09: 38.6 crit return --- 7.62 crit loss --- 7.58 average crit loss --- 54 blocks (5.4%)
10: 39.7 crit return --- 6.52 crit loss --- 7.47 average crit loss --- 48 blocks (4.8%)

11: 39.2 crit return --- 7.02 crit loss --- 7.43 average crit loss --- 25 blocks (2.5%)
12: 39.5 crit return --- 6.72 crit loss --- 7.37 average crit loss --- 46 blocks (4.5%)


The overall average for this test series is 7.37 percent crit reduction. Removing the 4.8 percent standard, this leaves us at 2.57 percent additional crit reduction. If I take the outliers out of the equation, then there is virtually no fluctuation in returns, putting the average at 7.36 percent crit reduction. Using our formulation with a standard 5 percent block mitigated shot percentage, we would estimate a 2.31 percent additional crit reduction. This is a -0.26 RNG fluctuation, and is barely on the outskirts of what we would expect. If I apply the formula using the actual percent of block mitigated shots (4.92) then the expected crit reduction is slightly lower, at 2.28 percent, but not enough to really cuase any major adjustments.

I will note that at a higher crit percentage, the swings seemed more extreme, so I would expect a large RNG fluctuation to be applied. We are still within 0.03 percent of expected returns, which is pretty in line with previous test.

Assuming that the Hit elixir had any effect on our return, since I only gained 1.37 hit (1.32 to reach hit cap) from the Elixir, it would be significantly harder to test for a fluctuation this small. The implied effect, if Elixir Hit was not able to Crit, would be only a 0.61 percent swing, and would require quite a large sample pool to test this. It would actually put us within the RNG threshold in the other direction if this was the case, so it would not even effect our results significantly. I am fairly confident that this is not a factor because unlike the Hit we are going to test from Focused Aim, the Hit from the Elixir shows up on the paper doll.

These 12k shots worth of testing took 8 hours (40 minutes a set by my potion timer), so I am going to push off testing Hit Food / Elixir / Potions till a later date. Right now in the same gear set up, I will do a smaller series of shots from the back of the dummy just to confirm the control we have set up before in our previous tests. If anything, if there is a disparage because of the elixir, it will show up here.



46.22 crit rating, 264 Hit rating (0/0 FA, 45 hit from Elixir), level 83 Dummy, from the back

01: 40.6 crit return --- 5.62 crit loss --- 5.62 average crit loss --- no blocks
02: 40.6 crit return --- 5.62 crit loss --- 5.62 average crit loss --- no blocks (same return for both sets, not an error)
03: 40.9 crit return --- 5.32 crit loss --- 5.52 average crit loss --- no blocks
04: 40.3 crit return --- 5.92 crit loss --- 5.62 average crit loss --- no blocks (oddly consistent so far)
05: 42.6 crit return --- 3.62 crit loss --- 5.22 average crit loss --- no blocks

06: 42.3 crit return --- 3.92 crit loss --- 5.00 average crit loss --- no blocks
07: 43.1 crit return --- 3.12 crit loss --- 4.73 average crit loss --- no blocks
08: 40.2 crit return --- 6.02 crit loss --- 4.89 average crit loss --- no blocks
09: 44.0 crit return --- 2.22 crit loss --- 4.60 average crit loss --- no blocks
10: 40.4 crit return --- 5.82 crit loss --- 4.72 average crit loss --- no blocks


I am not going to bother with the next two sets, since we can see that we are getting very close to the expected 4.8 crit reduction. You can see from the two sets that at this crit level, the reduction by attacking from the front is so vast that there is only one set in the first series that even crosses over into the numbers in the second series, and this is the outlier for the first set. Getting my Hit from the Elixir seemed to have no influence on the returns here.

I will make a post tomorrow about these preliminary conclusions, and I think we can finally move on to testing if Focused Aim is effecting crit returns as well. It seems that since it does not show up on the paper doll, like the food or potions do, it may have some time of interplay that is different from the standard functioning of Hit. We will see. It has to be tested a high levels of Crit for any difference to be substantial enough that it is not consumed by RNG.

If you're getting "immune" it's probably the TNT stun proc. Admittedly I have only played the new survival in it's PTR form on test dummies but I definitely saw immunes from this. Immunity doesn't make any sense as part of a mitigation/roll table. If you mean a full resist, were you hit capped? The spreadsheet originally had explosive subject to a physical hit table, and then it could be partially or fully resisted (17% cap) - but after some recent testing it appeared that it could not be fully resisted. If you notice anything differently, please make note of it - it could definitely use more testing.


Aside from that, I don't really know what you mean by direct block. As I understand it something can either be blocked completely (shows up as "blocked") or partially blocked (this is mitigation). Obviously his example could only have been partially blocked.

I have definitely noticed that explosive shot can crit, be blocked (partially), and partially resisted at the same time. Whatever it's doing with hit and crit, it seems to be passing through two mitigation checks after that (one physical, one magic). It's also not limited to explosive shot - I have seen partial resists on arcane shot, serpent sting, wild quiver, chimera shot (both direct and serpent effect) and physical mitigation in the form of partial blocks for arcane shot and chimera shot (direct damage only). This test started out to show that serpent sting could be resisted partially and ended up showing more than I intended. (Sorry for the tangent here, but it seemed relevant).

I have been reading this thread as often as I could, and have a few questions. Unfortunately, I do not have the time to attempt to work out these questions myself, nor do I know the feasibility of these questions. They are but mere thoughts.

1. Has anyone tested with the Hit Food instead of/or with Talents?

2. Has anyone tested with a higher Hit Rating then necessary to see if that effects the output of the Crit Return?

After switching from Beast Mastery to Marksmanship, I found myself sitting at roughly 11% Hit Chance. Unfortunately, I was not reading these forums yet, otherwise I would have tried to get Wow Web Stats of the raid that I did with those numbers.

I am sorry if these questions are irrelevant, they have just been sitting in the back of my mind and I had not seen them mentioned yet.

these are all things that will be tested in the future, but to get the right information, it takes time to test each parameter. I am testing with an Elixir right now, and if I get a result that is too far off the formulation we are considering right now, I will go back to getting all the hit from gear and retest again.

Alright! Thank you for the response and I will try to do some testing myself when I get the time. This thread has been a big help for me thus far with information.

I performed the following two tests:

========== TEST ONE ==========

http://home.comcast.net/~morgasaurus/crit_test_02.jpg

Information
Specialization: 0/0/0
Hit Rating: 299 (9.12%)
Paper Doll Crit: 16.10%

Auto Attack Results
Shots: 6047
Crits: 705
Sample Crit: 11.6587%

Analysis
We are 95% confident that the true crit rate is between 10.85% and 12.47%
Thus, we are 95% confident that the true crit differential is between 3.63% and 5.25%

========== TEST TWO ==========

http://home.comcast.net/~morgasaurus/crit_test_01.jpg

Information
Specialization: 0/3/0 (Focused Aim)
Hit Rating: 212 (6.47% + 3% FA)
Paper Doll Crit: 17.33%

Auto Attack Results
Shots: 5494
Crits: 658
Sample Crit: 11.9767%

Analysis
We are 95% confident that the true crit rate is between 11.12% and 12.84%
Thus, we are 95% confident that the true crit differential is between 4.49% and 6.21%

=========== INTERPRETATION ===========

Both confidence intervals overlap and both confidence intervals contain the expected crit differential. Focused Aim is not proven guilty by this test.

If we examine the effect of Focused Aim on crit rate, and it does cause a problem, then that would result in a 3.0 percent crit rate reduction at 100.00 percent crit. At ~16.0 percent crit rating, the effect of Focused Aim would only be 0.48 percent. At 95 percent confidence, that can get eat up entirely by RNG. You can see in my previous tests that RNG causes swings of up to +/- 2 percent, so the test has to be longer in order to get any reliable information.

This is why we suggest breaking test down into 1000 shot sets, as well going 10k - 12k shots. RNG can really mess up shorter results. Did you record block mitigated shots as well?

Also, since we have not tested the effects of going over the hit cap, testing at 9.12 percent, and 8.47 percent (with FA) may still be subject to some anomalies.

I appreciate the test though, and I am starting to become fairly confident that FA may not be the factor, and block mitigated damage shots are the culprit.

I shot from behind the heroic doll in Silvermoon city which seemed to eliminate blocks. Both tests were broken up into irregular intervals because I didn't do it all at once.

Also, what is the actual hit cap now? I've been assuming 9%.