I don't know if anyone has posted these yet, and frankly, I'm too lazy to search for it.

Here are my formulas for crit_chance per crit_rating per level and for hit_chance per hit_rating per level.

Curve = 0.00010897435897*LEVEL^2 - 0.08716666666667*LEVEL + 6.83769230769229

crit/crit_rating = Curve/28

hit/hit_rating = Curve/20


The coeficients are in long format and can be truncated to meet reasonable tolerances. These formulae were determined under the asumption that the numbers Eonyx announced in his blue post are absolute.

You should use significant figures. The world is a better place with significant figures, especially with theorycraft that people tend to multiply in their head when deciding to roll or pass on an item. No one is able to multiply those twelve sig fig monstrosities in their head, so use one or two sig figs and trade a little bit of precision for a whole lot of convenience.

While I'm ranting here, I'd also like to complain about the Ratings systems ( http://www.wowwiki.com/Combat_Rating_System ) not rounding to whole numbers at level 70. People are going to be doing this in their head, and doing most of it at level 70. It would have been kind of Blizzard to make the ratings come out such that 1% increases were nice round numbers of Hit Rating, Crit Rating etc. instead of things like 15.8 hit rating for a 1% hit increase at 70.

Round numbers are easier to work with, but I would rather need 15.8 hit rating to get 1% to hit than need 16 hit rating. I am glad they started with even number at 60 and worked out scaling to 70.

Also, the UI mouseover makes seeing the benefit pretty easy without needing to do much math.


BTW, here are the formulas from wowikki, where L is q(L) = (L - 1)/(70 - 1), so q(L) is 0 at level 1 and 1 at level 70:

Dodge = 17.9 * q(L) ^ 3.11 + 1
WeaponSkill = 2.9 * q(L) ^ 4.21 + 1
MeleeHit = 14.8 * q(L) ^ 3.18 + 1
SpellHit = 11.6 * q(L) ^ 3.23 + 1
MeleeCrit = 21.1 * q(L) ^ 3.09 + 1
SpellCrit = 21.1 * q(L) ^ 3.09 + 1
MeleeHaste = 14.8 * q(L) ^ 3.18 + 1
SpellHaste = 14.8 * q(L) ^ 3.18 + 1
Defense = 1.4 * q(L) ^ 6.58 + 1
Parry = 30.5 * q(L) ^ 3.02 + 1
Block = 6.9 * q(L) ^ 3.48 + 1
Resilience = 38.4 * q(L) ^ 3.00 + 1

Both of those formulas look overly complicated. Were they even based on anything besides level 60 and 70 information? If not, you could even fit a linear function and get perfectly good results for those two levels. If there are more numbers behind those, I’d like to see them.

Edit: Oh, just noticed that level 34 number on Eyonix's post. Will look into this.

http://www.wowinterface.com/download...ng_Buster.html

Author of this mod reverse engineered formula based on about 20 samples for each value, and his formula is correct to 13th decimal place

Very nice info, thanks. These numbers make much more sense.

The level 8 to 60 formula follows item budgets pretty closely. Basically if equal percentage of the stat budget on your items was spent on Crit Rating, your Crit chance would stay the same.

Am I correct in the assumption that each point should be roughly the same DPS(or avoidance) and that unless a stat is specifically bad/good the person doing the "is this item worth it?" math simply has to total up all of the points? I guess reverse engineering everything is neat, but it seems blizzard simplified it for us.

Speaking of crit rating, I noticed something amusing the other day - whatever automated process Blizzard used to convert legacy items to the ratings system, it did all the calculations based on the level 60 numbers, even for lowbie gear. So a pair of Shadowskin Gloves, if equipped at the minimum level, will give nearly double the benefit that they currently do. Not a huge issue, as there simply aren't that many low-level items with crit, but those few that do exist become immensely more powerful for their level.

Yeah, that's what I went to check pretty early on too (since Warden Staff was like the centerpiece of bearform tanking for a looong while >_>) but since people usually spend 1-2 days at any given level and 100+ at 60...

Ohwell.

Does anyone else find this silly?

The whole "rating" thing that is - I mean, I understand the point, but surely there are better ways to put this out there?

Whats silly about them?

Its a good decent solution else than its slightly confusing to the average person?

I think the nomenclature is lame yes, and perhaps the implementation a bit wonky, but it's really hard to argue too much since I think everyone agrees the current (retail) system won't work.

No I don't find this silly. What other ways would you suggest to do this?

I just find it silly that it isn't a standardized value across all ratings to reach a percentage chance in that rating.

That's simply because 1% dodge is not worth as much as 1% parry, but a 1 rating is always worth 1 rating.