If you could post your data, that would be helpful - for instance, 26.1% may or may not be within the statistical margin of error. On point #2 in particular: current consensus at least among rogues is that there's a 3% crit reduction to spells against boss-level mobs, and 4.8% on melee attacks - which seems more or less in line with your results.

a 0.6% less avoidance for each avoidance stats and 4.8% less crit chance are in line with my test
Expertise cap and expecially glancing will need more testing. I'll post my data soon (damn me I've cancelled my combat log going to raid yesterday :/ ).

HUNTER CLASS MECHANICS:
If you are specced 6/14/51, then a raptor outdps's a cat by a small margin because of the synergy of talented-kill command and savage rend.

Frost traps DO NOT trigger the ICD of LnL, making trap dancing still viable.

I've done some extensive test at Iron Forge Boss-Target-Dummy post 3.0.9.

Here the expected results:

I've done 4 different test with different gear set-up. Assuming 1 roll system for white attacks, I can easly sum up the results from different gear setup.

In every set-up I have between 264 and 280 hit rating.

With 260 hit rating I had a miss (in line with the 264 = 8% hit supposed).
With 25 expertise I had a dodge. (in line with 26 expertise = 6.5% chance to dodge supposed).


Hit 3462
Crit 2618
Glance 1975
Parry 162
Total 8217



Diminushed crit chance (difference between recount and paper doll): 4.67% +/- 0.15% (in line with the supposed 4.8%)

Glancing Blow: 24.01% +/- 0.94% (in line with the supposed 24%)

And now the unexpected:

The new expertise Hard Cap


Expertise__Parry________Error
26________6.2% ________1.4%
36________5.0%________1.3%
52________0.9%________0.3%
55________0.3%________0.6%
56_______0.0% ______2*10^-18 (2k hits -> 0 parried, I've used the poisson approssimation to compute error here, likelihood assuming old cap of 16%, if we assume 14.25% as the cap the chance is 0.3%)


Sum Up

The expertise hard cap it's 56. -14% chance to parry. It's not 16%!!!!

I had yet to see anything more fractional than 0.5% in this game (for what concern hit table: 8%, 24%, 27%, 6.5%, and so on..). But surerly I'll give you the data:

Parry/sample/expertise rating (+10 expertise from talents)

80/1291/137 (not 100% sure about this one, because I've started taking exp rating value after this test) (+10 exp from talents)
62/1237/218 (+10 exp from talents)
29/3350/347 (+10 exp from talents)
1/318/375(+10 exp from talents)
0/2031/380 (+10 exp from talents)

I'm not entirely sure, but somehow I don't think that's the intended mechanic for LnL and might get changed. I will however make a note of that.

Will update expertise hard cap.

According to everything I´ve read so far the hit cap is at 263 not 264 hit rating (8*32,78998947=262,31991576 using Whitetooths values). If I´m wrong with this assumption this would be pretty major for me because it would mean a whole percent less hit for my pet with exactly 263 hit rating (as pointed out by Rosamonde).

For the defense section, it's probably worth noting that Death Knights who use the Runeforge of the Stoneskin Gargoyle need 566 Defense Rating (115 Defense Skill) to become uncrittable.

Edited to reflect new information...
Here is a link to some past data where a warrior tested parry rates on this forum.
http://elitistjerks.com/f31/t18771-p...ior_guide/p75/
Here we have 10191 swings with a tooltip showing 13.75% reduction from Expertise. We do not however know the exact expertise rating. Worst case scenario, if this test were conducted with 377 expertise rating (with +10 from talents). This would still show 13.75% on the tooltip and gives a parry rate of 13.9974%. In this case one would see a parry in ~38640 attempts so its still possible the rate is 14%.

Above we have a test showing a parry within 318 attempts at 375 expertise rating and +10 expertise showing that the parry rate is greater than 13.9364%

I did a few followup tests on the PTR. It should be noted that currently expertise rating is giving 125% conversion on the PTR.

With 298 expertise rating and 10 expertise from talents I found 1 parry in 1290 attacks showing the rate to be greater than 13.8602%. If 14% expected 1 parry per ~715 attempts.

I then tested with 300 expertise (the equivalent to 375 above) and got a single parry in 1917 attacks. Which confirms the above rate greater than 13.9364%. If 14% expected 1 parry per ~1573 attempts.

Unfortunately, my expertise gear on the PTR I cannot test at 301, but everything is pointing to 14%.

This is a really late catch, but there is no hit penalty for live Titan's Grip nor none planned at this time. Thanks for this compendium Enova.

Being parry gibbed by Malygos or Sarth is not something I really like and probably into Ulduar will be more important, that's why is important the expertise hard cap. A tank should want to reach it (it's not that difficult every tank in my guild is from 45 to 55 expertise). The main answer is: does parry chance change from boss to boss? Is it always 14%? We have a 2-wield Frost DK tank that tanked malygos 10 (he is an alt) yesterday and he has "only" 48 expertise but he is sure that he never was parried by the boss.