This site makes extensive use of JavaScript.
Please enable JavaScript in your browser.
Live
PTR
10.2.7
PTR
10.2.6
Beta
Should a bear ever gem for crit?
Post Reply
Return to board index
Post by
405390
This post was from a user who has deleted their account.
Post by
Sunstreaker
No, you never gem for crit. Ever.
Post by
Toldu
I don't have any numbers to back this up, but I would think that agi would be preferred above crit and dodge. Obviously, get hit-capped, stam is still great, and def is pointless.
Post by
hashmel
yes, crit = threat
BUT
bears have quite a high crit% as it is and threat isn't nearly such an issue that our gems/chants need to be focused around it
i'd like to direct you/your friend
here
for further guidance on bear intricacies
Post by
285998
This post was from a user who has deleted their account.
Post by
Thror
The bears that like to match socket colors no matter what (i used to be like that..) are usually trying to find some stat for the yellow sockets. IF you are hit capped, the only yellow stats that are relevant to bear tanking are Crit and Defense. Both are crap, and its better to ignore the socket color, and pick +30 sta instead, but since some people just cant stand having a socket that isnt happy and sparkling because it has gotten a gem it likes, id advice... +10 defense and +15 stamina, over crit. Its kinda better, since your Crit is knee-deep in DR thanks to agility, but your Defense isnt.
Post by
MegaVolt
He is being stupid.
Yes, crit is threat. But AP would be much more threat so his argument is really pointless. If he wants threat then he would gem AP or even STR over crit.
The thing is that AGIL will give a fair amount of crit and also help with armor and doge. It is far superior to pure crit. Your priorities are right: Go for exp/hit if you need threat (which should not really be an issue in the first place, usually it is safe to just ignore those caps as feral tank), gem sta for a longer time to live and gem agil for mitigation/avoidance.
Don't gem anything else and slap your noob Druid hard.
@Thror: There is no DR for crit. Just a cap (because of the one-roll-system) which bears should not be that close to.
Post by
Thror
@Thror: There is no DR for crit. Just a cap (because of the one-roll-system) which bears should not be that close to.
That is a good thing to know, thanks.
Anyway, megavolt, while your gem suggestions are okay, you dont see the real problem with this particular bear tank we are discussing in this thread is, that he just wants to fill yellow sockets with matching gems. You know those people exist, and we cant do anything about it :B And all the gems you suggested have red stats in them~
The guy shouldnt be slapped, he might still be a good tank... Seriously, how much weaker can he be, opposed to a tank that gems pure sta instead of whatever/sta? Like 1,5% worse? These things are relative, and hard to measure anyway.
Aaaand if said druid tank is not in a HM progressing guild, he can really gem whatever he likes... and still manage to tank everything up to ToC25, excluding both ToGC's.
Post by
MegaVolt
He can probably even tank ToGC10 successfully in wrong gems.
But if he wants to match all gem sockets he really needs to be slapped. There just is no point in doing that. And if he got an insanely awesome socket bonus that he really, really wants to match (there is none that I know of that would be worth it though, serious slapping is justified) he can still use a sta/def gem instead of a sta/crit one.
Post by
Aadramelekh
He can probably even tank ToGC10 successfully in wrong gems.
But if he wants to match all gem sockets he really needs to be slapped. There just is no point in doing that. And if he got an insanely awesome socket bonus that he really, really wants to match (there is none that I know of that would be worth it though, serious slapping is justified) he can still use a sta/def gem instead of a sta/crit one.
About the yellow sockets with great bonuses - there is ONE piece worth socketing with a 10 defense / 15 stamina gem and that is the
T9 helm
. 8 Agility + 10 defense + 15 stamina > 30 stamina in this case. This is the only one I know is worth gemming anything else than stamina or agility.
@
Thror
: That cap is determined by the attack table. The normal swing attack table against a raid boss attacked frontally (by a tank) is this (considering 1-1000 roll and 30% crit chance):
1-80: miss (8%)
81-146: dodge (6.5%)
147-287: parry (14%)
288-538: glancing (25%, approximate rounded value)
539-839: critical (30%)
840-1000: normal (16%)
The roll precedence is as listed here. Glancing blows cannot be reduced in any way. This means that if yuo add 1% extra crit, your crit chance raises to 31% and normal chance drops to 15%. Adding 1% hit means miss chance dropping to 7% and normal raising to 17%.
In this case our bear with zero hit and zero expertise has a 46% hard cap for critical hits concerning white attacks. Special attacks however don't follow this table. They are not affected by glancing blow chance. As such, their attack table would look like:
28.5% miss+dodge+parry
30% crit
41.5% normal
So the crit hard cap for special attacks is 71.5% here.
Assuming our bear is hit and expertise hard capped, the white attack table looks like this:
1-250: glancing (25%)
251-551: crit (30%)
552-1000: normal (45%)
Here the hard cap for white attacks crit chance is 75% (because glancing cannot be eliminated).
Attack table for special attacks for a hit/expertise capped bear:
1-300: crit (30%)
301-1000: normal (70%)
In this case, the hard cap for crit on special attacks is basically 100%.
Hope this shed some light on the mechanics of the attack table. ^_^
Post by
465564
This post was from a user who has deleted their account.
Post by
466571
This post was from a user who has deleted their account.
Post by
Heckler
So I agree with your White Attack table (a 1-roll system), but I was under the impression that yellow attacks followed a 2-roll system, that is -- whether a special attack lands or fails is one roll, and whether it's a crit or not is another roll, meaning that if you have a 50% crit chance, then 50% of your attacks
that land
will be crits.
This means with a 50% Crit chance and a 20% failure rate, after 1000 attacks, you would expect 400 Crits (half of the 800 attacks that did not fail).
(
Source: The Original "Two Roll System?" thread from ElitistJerks
)
Maybe it's changed and I'm just out of date, but back when I was really into numbers in TBC, two-roll was the accepted correct system for special attacks.
This means that while there is a functional "Crit Cap" for white attacks, in order to have a 100% Crit chance for Special attacks, you would actually need 100% Chance to Crit, even with a 20% failure rate.
As a consequence of this, capping hit and expertise actually increases the value of your crit chance. (Less failures means more crits for a given crit chance).
Apologies if you already knew about all this and it's just changed since I last did heavy research, in a quick 5 minute google-job, I couldn't find any "Two roll system verified false" threads anywhere, but I did find quite a few suggesting a two-roll system to be more likely true.
Post by
Aadramelekh
Of course capping hit and expertise increases the value of crit chance for both normal and special attacks. Because normally miss, dodge and parry take precedence over crit on the attack table and crit taking precedence over normal hits, pushing miss, dodge and parry off the table actually means an increase in the cap of crit chance by 28.5%, up to 100%.
The whole discussion about the two roll attack table is something I am uncertain about as well. And as a matter of fact I don't really remember any "blue" information about it. So I just assumed specials run on the same attack table as normal attacks, just without glancing chance. I *do* admit I am also not exactly 100% sure about how the special attacks work. But them working on the normal attack table doesn't seem so far-fetched to me either.
However
, considering that they are based on a two roll system, this brings the discussion back to where it was during an old and fun topic here, about the value of hit/expertise for cat DPS. And those principles apply here as well.
A bear tank with 28.5% total miss chance (miss+dodge+parry) will have only a 71.5% chance to hit the target. If he has 50% crit chance, then actually only 35.75% of his special attacks will crit (on average). This simple fact here makes capping hit and expertise a MANDATORY requirement for efficient tanking and DPS as well. Otherwise it severely gimps the excellent effect of
Primal Fury
.
Because a one roll system would actually be more forgiving, by making the special attack table look like 28.5% total miss, 50% crit and 21.5% normal hit, thus maintaining full benefit from
Primal Fury
.
Post by
Heckler
The only blue post I could find just seems to add more confusion to the topic (
Source
):
The way WoW calculates crit rate is over ALL attacks. Crit rate is not based on hits only. In other words, if you have a 5% crit rate, that 5% chance includes misses
Some have take this to mean it must be a one-roll system ("not based on hits only"), I think it could just as easily be interpreted to mean that a "critical miss" is a possibility, something that is only possible with a two-roll system ("5% chance includes misses").
I seem to remember Emmerald modeling his original feral numbers on a one-roll system, and someone in his forums convincing him to change special attacks to a two-roll system, and I believe they referenced a blue post saying that it is technically possible to "critically miss" -- again, this is only possible in a two roll system... I'll try to find that one.
And all I meant about the Hit/Exp thing is this, let's say you had 1000 attacks, a 20% failure rate, and a 50% crit chance.
In a
one roll system
, you would expect 200 fails, 300 hits, 500 crits. If you reduced your failure rate to 10%, you would
still
expect 500 Crits.
In a
two roll system
, you would expect 200 fails, 400 hits, 400 crits, but if you reduced your failure rate to 10%, you would expect 450 Crits, even though your crit chance didn't change.
I actually brought this up back in the old "Hit Cap not Necessary" fight, but I don't think it picked up any traction there because no one can "prove" it's a two-roll system, so no one wanted to think about it. But I do agree, Hit / Exp are worth more in a two-roll system than in a one-roll... still not worth enough to stress about a 1-2% total failure rate IMO, but more nonetheless.
Post by
Aadramelekh
Okay I did a brief testing this morning before I left for work. 2 tests, about 5 minutes each. Mangle spam while auto attacking.
For the first test I took any hit / expertise gear down, remaining with 0 hit and 10 expertise from Primal Precision. Technically I was left with 8% miss, 4% dodge and 11.5% parry (parry is included because I was attacking the boss frontally). For the second test I got hit/expertise hard capped and did the same thing.
Note: I'll repeat the same tests several times over later in the afternoon once I get home to get a wider sample collection. This time take the results with a grain of salt due to the RNG factor and the low sample number (2).
The results are here:
Attack Table Test
--------------------------------------
So what do we see there?
First test:
8% miss, 4% dodge and 11.5% parry, 43% crit (I rounded this because absolute precision is not my goal right now). Idol + Mongoose + DMC procs would add an average +5% (up to 48%) crit over time, but I'll consider 45% crit because crit is reduced by game mechanics against a lvl 83 target (as far as I remember, by about 1% or so per each level of difference). Side note: because miss, dodge and parry produce basically the same result (zero damage to the target), I added them all under the name of "failure" (23,5% theoretical failure rate). Results are as follows:
Mangle:
20.8% crit, 48.1% hit, 31.1% failure.
Melee:
43.1% crit, 8% hit, 27.7% glancing, 21.2% failure.
It is interesting to see how the crit rate for Melee is consistent with our assumed crit chance (45%) while the crit rate for Mangle is abnormally low (less than 21%).
What happened to Mangle? Let's see... out of 154 Mangles, 48 failed to land. This leaves 106 that hit the target. Out of these, 32 scored critical and 74 normal. That means basically a 30% crit and 70% hit chance utilizing a two roll attack table. Of course I am aware of the abnormally high failure rate in this test sample (31.1%) compared to the theoretical rate (23.5%).
But it is obvious that this sample already shows a high probability that the two roll attack table is true for special attacks. Like I said in my note above, extensive testing is required in order to obtain more accurate results.
------------------------------------------------------------------------------------
Second test:
hit/expertise hard capped, 51% crit chance. Due to the Mongoose / Idol proc it could be argued that it is more like an average of 54% crit, but I'll assume a 51% crit rate. Same reason as above. Results are as follows:
Mangle:
42.4% crit, 57.6% hit
Melee:
48.6% crit, 24% hit, 27.5% glancing
This time the crit rate is somewhat consistent with the number displayed on the character sheet for both Mangle and Melee. Taking into account the possibility that specials use a two roll attack table, this shows that having hit and expertise capped (or very close to cap) is mandatory for improving the effectiveness of crit chance.
Later edit:
ALSO this leaves an open room for interpretation, in the light of what you quoted,
Heckler
:The way WoW calculates crit rate is over ALL attacks. Crit rate is not based on hits only. In other words, if you have a 5% crit rate, that 5% chance includes misses, like you said:it could just as easily be interpreted to mean that a "critical miss" is a possibility
Working on your example of 1000 attacks and 50% crit + 20% failure rate, a two roll table on the principle first roll determines failure or success of the attack and second roll determines crit or normal, the result would be a predictable set like 200 fails, 400 hits, 400 crits.
However the Blizzard quote above could mean that the first roll determines crit / normal hit and the second roll determines failure or success.
ANYWAY in both cases hit and expertise would have a tremendous importance for both tanking and DPSing.
As for the comparison between the one roll and two roll attack tables regarding special attacks, you have to admit that the one roll table is the best case scenario.
Let's say you had 1000 attacks, a 20% failure rate, and a 50% crit chance.
In a one roll system, you would expect 200 fails, 300 hits, 500 crits. If you reduced your failure rate to 10%, you would still expect 500 Crits.
In a two roll system, you would expect 200 fails, 400 hits, 400 crits, but if you reduced your failure rate to 10%, you would expect 450 Crits, even though your crit chance didn't change.
And with hit/expertise caps you would reach up to the one roll attack table in an artificial way. If we were to think of the two tables as "good cop / bad cop", the one roll table is the good cop here, the one that gives you full access to your crit chance and only penalizes normal hits in case of failure. The two roll table is cutting equal shares of both normal and critical hits (and since critical hits are not just 200% normal damage but 226% normal damage, much more is lost this way in terms of DPS when there is a chance of failure).
PS: sorry for the wall of text Q_Q
Post by
Thror
Hmm. I find all these numbers quite interesting, as in... i enjoy reading theorycrafting in process. However... hmm, i wonder whats the point of all this? Let's say we would know for sure whether WoW calculates white/yellow hits with a one or a two roll system... could the knowledge be practically used to... i don't know, min/max someones DPS/TPS or something?
So far it seems like the only effect it has is proving the importance of Hit and Expertise. But... there was noone who would even doubt the importance of those stats, i think pretty much everyone now knows that those stats are really important.
Well, i guess the point could also be "we are just curious", and that would be fine, for i am too. :3
Post by
Aadramelekh
Well yes, we are all curious around here, that is why we like crunching numbers and debating theories. After all, depite what they say, this kind of curiosity
DIDN'T
kill the cat and in the end it is what sets good players apart from the mediocre/bad players, amirite? ^_^
PS: "Hardcore/casual" have nothing to do with this :P
Post by
Heckler
Hmm. I find all these numbers quite interesting, as in... i enjoy reading theorycrafting in process. However... hmm, i wonder whats the point of all this? Let's say we would know for sure whether WoW calculates white/yellow hits with a one or a two roll system... could the knowledge be practically used to... i don't know, min/max someones DPS/TPS or something?
Right off the top of my head, I'd struggle to come up with anything beyond curiosity. It's obviosuly vitally important when building models / simulations, if you don't have this right then your calculated worths for crit, hit, exp (and therefore everything else) will be wrong for anyone who isn't capped.
As far as min/maxing goes, it then becomes important because FbN (which models specials as a two roll system) shows that if you could throw away all your Hit and Expertise and trade it for an equal amount of Agility, your DPS would actually increase, except in the case of movement fights which have less than ~50% time-on-target (from a prior pointless thread).
PS: sorry for the wall of text Q_Q
I agree with everything you said, especially the high failure rate test -- very interesting stuff. Also I'm not sure which dummy you were attacking (83 or 80), but your crit chance on the char screen is only accurate vs. an 80, and the parry chance against the 83 dummy looked like 15% base in the last test I ran, I'm not even sure what the 80 dummy's parry chance would be.
Either way, the general trend of your data points more toward a two-roll than a one-roll, as you said.
PS: "Hardcore/casual" have nothing to do with this :P
Very much Truth. I've never beaten Anub10(Normal) for example, and I probably won't anytime soon. I consider myself good at what I do, but I'm not in a "hardcore" guild, and I don't think I'd fit in with one anyways due to scheduling. But I do enjoy thorough understanding of the games mechanics, and I think it helps be a better player.
Post Reply
You are not logged in. Please
log in
to post a reply or
register
if you don't already have an account.