Focused Aim reducing Crit Return
 

This is a continuation from another thread, since the title of that thread no longer lends itself to the current discussion.

This is the original thread:

http://elitistjerks.com/f74/t38309-l...t_steady_shot/



Lethal shots was eliminated as a factor, and along this road the 4.8 percent crit reduction versus level 83 mobs was reinforced. This was true if the hunter testing was fully hit capped from gear, and ran a long enough test. In this process I started discovering that if I had Hit from Focused AIM, my crit was lower then the 4.8 (new) standard.

I ran a 10,000 shot test with 164 hit (5.00) percent and 3 percent hit from Focused Aim. There were no misses. I had no other talents allocated.

http://i179.photobucket.com/albums/w...al/10ktest.jpg



My paper doll crit was 21.72, with no crit procs available. My Crit return after 10k Auto shots was 15.72, or exactly 6 percent. This would indicate a 1.2 Percent Crit reduction from gaining Hit from focused aim. AT the point of 7000 shots, the crit only fluctuated between 15.6 - 15 .7, indicating a less then 0.05 margin of error.

My current concern is that I always saw a range of 1.2 to 2 percent crit reduction , but not all the tests were of the same length and even at 5000 shots, I saw some RNR. It may be possible that at higher crit values, there is more reduction, indicating this is somehow scaling with gear.

I have to test at roughly a 30 - 32 crit rating, but currently, the only way I can get that high also puts me at 264+ plus hit, so Focused Aim is eliminated. I may have to talent 5 percent crit and try to switch some gear to get it up to 30 percent.

Well, the talented hit could still impact crit if it overwrites the Hit from gear.

sorry for the noob question, but I seem to be confused about the hit cap figures. I have been under impression that the hit cap for a raidboss (lv83 mob) is 9%. However, your exercise suggests that the cap is 5+3=8%, and the dummy on the picture is the lv83 one. What am I missing?

See the thread in this forum titled "Hit Rating."

yes, I have considered this, and I will run a test at 264+ hit, as well as with Focused Aim, in order to see if the hit from Focused Aim has priority.

The first test took me approx 5 hours, so I need a chunk of time to devote to it.

But since our pets don't get the benefit of FA, isn't it pretty much a waste of points anyway?

Are you doing your test from in front of the dummy? Could there possibly be a 2-roll system where block chance would be lowering your overall crit chance?

Here are my tests so far:

Paper Doll:
26.46% Crit
169 Hit
3/3 FA

Test 1 from the front:

Steady Shot:
2581 Hit
740 Crit
3321 Total
22.28% Crit
4.18% Discrepancy

Auto Shot:
2315 Hit
585 Crit
2900 Total
20.17% Crit
6.29% Discrepancy

Overall:
4896 Hit
1325 Crit
6221 Total
21.30% Crit
5.16% Discrepancy

-----

Test 2 from the back:

Steady Shot:
2446 Hit
652 Crit
3098 Total
21.05% Crit
5.41% Discrepancy

Auto Shot:
2153 Hit
610 Crit
2763 Total
22.08% Crit
4.38% Discrepancy

Total:
4599 Hit
1262 Crit
5861 Total
21.53% Crit
4.93% Discrepancy

Either way, my numbers are not supporting your theory.

For an MM / SV hunter, this trade off for losing pet hit could be advantageous, since 3 percent hit on 80 - 85 percent of your total damage is still effectively 2.4 - 2.55 hit, while for a BM, since the pet is 40 - 50 percent of their damage, it is only a gain of 1.5 - 1.8 percent hit. For MM / SV that is an acceptable lose in exchange for freeing up itemization points on gear.

Basically this talent is less effective for BM hunters, and the additional lose of 1.2 percent crit, would make it a poor use of talent points. The loss of 1.2 percent crit would also make it less attractive for MM / SV hunters.

(percentages range is approximate)

If you look at my screen shot, there were no blocks.
Why would I be behind the dummy at all?

I only tested Auto shot, and your first test supports this in the case of Auto shot.

Did you have any other talents specced?

At 3000 shots (the approx. length of your individual skill tests), I have seen RNR cause fluctuations of 2 percent. Run to at least 8000 - 10000 shots of an individual skill i.e Steady Shot, Auto Shot, in order to reduce the influence of mathematical patterns of low crit and high crit strings.

I prefer to have screen shots posted, to verify gear and other statistics, such as if other points were allocated. Indeed the reason I ran a full untalented test was at Rosamonde's observation that my previous tests had other points allocated. Eliminate all other factors, including proc trinkets, and proc talents.

I have had to consider if Master of Anatomy was effecting my crit returns, but in my previous tests, even with the crit from that talent, at 264+ gear, I was showing 4.8 percent crit reduction only. Even so, Master of Anatomy is only 25 crit rating, or about 0.54 percent crit.

Recount does not record our blocked shots. Watch your logs as you go. You will see that ranged attacks do get blocked by the target dummy when attacking from the front.

You can armory me to see what I'm wearing and my talents. There's no proc trinkets and no talents that would cause any fluctuations in crit or cause a variance between Steady and Auto Shot crit chance. Those tests were done at those numbers because that's when my gun broke. I'm sorry but I don't have a magical method of not having to repair. I don't really care what you 'prefer'. You can take the data as it is or not. These 2 tests combine for 12000 shots with about a 5% reduction in crit. I'm confident that FA is not reducing our crit chance.

To elaborate on this slightly, Recount records blocked shots, but it does not express them as a separate percentage from hit, crit, or miss. If you have a ranged shot that hits but is blocked, it is counted as a "hit" in the pie chart and similar tables. Instead, Recount will show the total amount of damage that was blocked, if you go to a more detailed view with a damage breakdown.

Yes. Sorry if I was unclear. It records blocks as hits. However, I was doing some further research on this topic and it was reported that blocks can also crit. This proves 2 things. There is a 2-roll system. The first being Defensive (dodge, block, miss, etc) and the 2nd being offensive (regular or critical hit). It also proves that blocks count as hits on the 1st roll so they have no effect on crit chance.

here is a copy of the details log for that test.

http://i179.photobucket.com/albums/w...detailcopy.jpg

blocked shows up only as blocked amount of damage, and does not show up on crit and hit. It is a total of 0.006 percent of damage blocked.

On further investigation this is 496 shots with block damage (76 blocked per shot) and results in a 4.96 percent blocked damage shot percentage.

That is fine and all, but in my previous thread, Rosamonde pointed out that other talents may be effecting crit returns, and that is why I did the test specced 00 / 03 / 00. Also it may very well be only Auto Shot that is effected this way, that is a possibility. Again you may have 12000 shots, but is is of two different abilities and from two different positions, so each counts as a separate test.

I repaired and restocked on Ammo while testing in order to get 10000 shots. If you have yourself set up in a particular spec, there may be things effecting it. This is not the first tests that have shown me this return.



Here is a test with Focused aim:

http://i179.photobucket.com/albums/w...tFocused83.jpg

Here is a test without Focused Aim.

http://i179.photobucket.com/albums/w...NoLethal83.jpg

This is where I found the original difference. Even though these support my theory, I have eliminated then from my own test sample base, as the test where not long enough. Here is another 10k Auto Shot test:

http://ordure2000.free.fr/10k.jpg

This shows a 6.1 percent crit reduction on 10k auto shots. He has 169 hit from gear, the rest from FA.


There is something while using focused Aim that is reducing crit return, and my test which ISOLATED the talent, showed that I was losing an additional 1.2 percent on top of the 4.8 from level difference. If I hit cap from gear, I am not losing it. I was confident in my original findings, but after scrutiny of my tests, I decided to retest in order to eliminate all factors, including any talents that effect crit (such as lethal shots), in order to be sure of my theory.

This variance in crit returns has been show before, but in all the other tests, there were other factors to cause variance. There is cataloging of anywhere from 2.56 percent crit reduction to 6.34 percent crit reduction, but the length of all these test was not sufficient enough to account for RNR. In your own tests, the crit reduction varies from 6.29 percent to 4.18 percent. that's a 2.09 percent swing.

I appreciate your attempt to test this as well, but you did not go to the same lengths to isolate this talent that I did, including re-gemming and un-talenting, in order to eliminate any direct or indirect effect on crit. I also had to watch the dummy in order to make sure no one was putting a debuff on it that increase crit chance, and to make sure I was not get buffed by people passing by. This is why I test in Darnassus, much less of a chance of that happening.

If I am not losing crit from Focused Aim, then where am I losing it?

Since I made this spot by accident...again, I will show some RNR samples.

I am currently running a 10k, 263 hit from gear test. This is to show the crit variance at intervals. Paper doll crit is 27. 86

Shots-----Crit rate

1000-----22.0

2000-----23.1 --- this means if in my first 1000 shots I crit 22 percent, and in my next 1000 shots, I crit 24.2 percent.
2500-----22.5

3000-----21.8
3050-----22.1 --- this is right after a long crit string. Right before it, I had watched myself get basically 1 crit per 10 shots for a good 200 shots.
3500-----21.9

4000-----22.1

5000-----21.9
5500-----22.1

6000-----22.0

At this point there is less fluctuation, but by this return, I am losing 5.86 crit, which is contradictory to the 4.8 percent excepted crit loss. I am going to start a 2nd series, this time breaking it into 1000 shot blocks in order to see what the variance per 1000 shots is, and why I had such a high spike for 1000 shots.

Right now, in order to reach the standard 4.8 percent reduction, I would have to maintain a crit rate of 24.50 percent over the next 4000 shots.
This is starting to look like a more complex pattern...