 |
| Welcome to Elitist Jerks |
We're testing some new features on the site regarding OpenID registration and coordination with gamerDNA. If you experience any issues with registering an account, please take the time to fill out a report and send it to this e-mail address. We would appreciate any assistance you could provide in making sure everything is functioning as intended. Thanks!
If this is your first visit, please be sure to check out the FAQ and the forum rules. Users must register to post and new registrations are subject to a one day "mute" period to get acquainted with the community.
|
02/12/09, 11:04 AM
|
#51
|
|
Soda Popinski
Night Elf Druid
Frostmourne
|
Originally Posted by Ukerric
Except one: The PvP items. Each successive incarnation of PvP items is close to a linear 10% (modulo rounding effects) increase in each stat, and each family of items have exactly the same repartition in stats.
|
Cloaks - Items - World of Warcraft
[LK Honor 5 Caster DPS Crit Cloak]
Yes, I realize the second link is potentially placeholder. Even so, the 200 and 213 cloaks do not share a proportional linear increase in each stat.
|
#elitistjerks
<^clicker> nice job trying to troll but you're a fucking idiot because i wasn't responding to you
<^clicker> this is the channel for serious discussion of important world of warcraft issues i believe youre looking for /b/ get lost scrub
...
<^clicker> do you act like this all the time
|
|
|
|
|
02/12/09, 11:35 AM
|
#52
|
|
Don Flamenco
Undead Priest
Whisperwind
|
Originally Posted by bdew
This is false. All the random sufixes are stored centrally with your item having only the ID of the suffix. On the client side this is in ItemRandomSuffix.dbc
|
Interesting. Maybe that means the only reason they didn't change them in the past was not to upset players.
|
|
|
|
|
|
02/15/09, 12:33 PM
|
#53
|
|
Glass Joe
|
NM.
Last edited by A-KO : 02/15/09 at 12:43 PM.
|
|
|
|
|
|
03/06/09, 5:32 AM
|
#54
|
|
Glass Joe
|
Calculating Armor via Item Level
Does anybody know how to calculate base armor via the item level? From my searches I've found several different yet similar formulas. However, they all appear to be wrong (including wowwiki). I've tried creating a few formulas on my own but have been unsuccessful in producing an accurate model. Any ideas?
Last edited by Symgeosis : 03/06/09 at 6:18 AM.
|
|
|
|
|
|
03/06/09, 5:38 PM
|
#55
|
|
Glass Joe
|
When trying to calculate the value of a socket, has anyone compared the value of the blacksmithing ability to add a socket to other profession specific bonuses that are easier to value?
If we assume that each profession's bonus is equal in item budget we could solve for the value of a socket:
Enchant ring: (16 points) x2 = 32 points
Jewelcrafting: (11 points) x3 = 33 points
Leatherworking: 50 sta -> 33 points
Mining: 50 sta -> 33 points
Blacksmithing: (x points) x2 = 33 points
Seems like x is 16.
|
|
|
|
|
|
03/09/09, 11:47 AM
|
#56
|
|
Von Kaiser
Dwarf Paladin
Khaz'goroth
|
- Armor Formulas: I wrote the ones on wowwiki myself. However that was during BC and i haven't updated since then. There seems to be a (very) slightly different armor formula in WotLK.
I just plotted iLevel vs Armor on a graphics calculator and used a linear regression tool to find the line of best fit. I'll try it again for LK
- Sockets: The Smithing sockets are indeed worth 16 points. However from all the testing I did in BC, there did not seem to be a fixed value for sockets.
I tried thinking of each socket as a separate stat (Ie, item would be 50 Stam, 50 Str, 1 Socket, 1 Socket, 1 Socket). Didn't work.
I tried thinking of sockets together as one stat (50 Stam, 50 Str, 3 Sockets). Also didnt work.
I suppose its possible there is some exponential or polynomial function on the costs of sockets, but even normal epic items are difficult to determine their budget.
My eventual conclusion (which wasn't subject to thorough testing but is on wowhead anyway) was the assumption that each socket was simply X Points removed from one stat on the item. This meant that, if you socketed in the way blizz intended, you end up with a perfectly balanced piece. If you stack one stat, its entirely possible (and quite likely) that you end up with an 'over-budget' piece - this is why we love stacking Spellpower/Stamina/Strength in off-color sockets.
(and obviously, when epic gems came out, items of that tier and above are itemized as if you socket with epics. current LK items assume blue gems.)
|
|
|
|
|
|
03/09/09, 4:19 PM
|
#57
|
|
Glass Joe
|
Thorgred, thanks for the response. I did do the exact same thing in fact; However, I found that the linear regression wasn't as precise as I desired. I may be going a little over board on being precise but I'm working on some theorycrafting tools for druids and all the multipliers for bear form lend themselves to precision. I also presume there must be an exact formula out there for at least T6 - T7 level gear since Blizzard redid the armor values for feral T6 near the end of BC.
|
|
|
|
|
|
03/09/09, 11:53 PM
|
#58
|
|
Von Kaiser
Dwarf Paladin
Khaz'goroth
|
Just some quick testing for WoTLK items got me the following plate formulas:
Greens: 8.843*X + 41.214
Blues: IF(X<180): 12.095*X - 121.881
IF(X>180): 5.728*X + 1020.709
Epics: 4.517*X + 1334.562
Note that with such a small sample size on epics (i had to use MMOC's Ulduar items) on epics it's hard to get an exact formula. All of these come to within 2 points of actual.
The discrepancies i think come back to how blizz actually calculates iLevel: is the iLevel *really* 213 or is it 213.?
The more i look the more i'm sure that items actually have non-integer stat values.
Ie, your '50 Str 50 Stam 50 Agi' chest is more closely 49.6 Str, 50.1 Stam, 50.4 Agi.
Which would mean the calculated ilevel if you plug in 50/50/50 is different to the actual ilevel calculated with the decimal values.
This also means that an item with 50/50/50 could have stats anywhere between 49.5/49.5/49.5 and 50.5/50.5/50.5
So you could potentially have two chests with 50/50/50 of three stats, a few iLevels apart. You actually see it a lot on low level (<50) greens. See where i'm going?
This is why its so difficult to calculate epic iLevels and see if they're underbudget. The stats that we *see* may cause us to think it is iL200, but the hidden 0.5's add up to make it up to 213.
You'll also need to make sure to factor in the turning points in itemization formulas. In BC it was particularly prevalent in weapons (they changed formula at 97 and 138), so you can't just put *all* values in to a LinReg tool - make sure that you at least have a look at the graphs.
I have lectures to attend but i'll do more deriving when i'm done 
|
|
|
|
|
|
03/11/09, 4:21 AM
|
#59
|
|
Bald Bull
Human Warrior
Turalyon (EU)
|
Blizz confirmed that some BR/BV items were underbudget so we know for sure now that the underbudget items are the exceptions and not the rule (IIRC there was some discussion if Wall of Terror was itemized normally and other items were just overbudgeted).
|
I don't want to derail a good discussion, which blue comments tend to do. However, I did want to offer that we did fix those Naxx block items in question.
|
World of Warcraft - English (NA) Forums -> Pre 3.1 Exhaustive Tank comparisons
|
Unexpected TankPoints error
|
Originally Posted by Ghostcrawler, justifying Druid health > Warrior health
To be generous, the warrior has 50K and the druid has 55K? How many times is that 5K going to make a difference when the boss hits for 40K? I know more Stam is always better, even in relatively trivial amounts. But until the magnitude is so large that the druid can survive one more hit than the warrior, it isn’t likely to crop up often.
|
Clearly someone doesn't understand how EH works. That, or upgrading from T8 to T9 is optional in beating Arthas. Clearly.
|
|
|
|
03/11/09, 5:32 AM
|
#60
|
|
Von Kaiser
Night Elf Druid
Dalaran (EU)
|
Originally Posted by Thorgred
The discrepancies i think come back to how blizz actually calculates iLevel: is the iLevel *really* 213 or is it 213.?
|
In the item database, iLevel and stats are Integer values. Also, it is clear to me that iLevel is never calculated by Blizzard, it is chosen, and stats are calculated to match that target iLevel.
|
|
|
|
|
|
03/11/09, 8:23 AM
|
#61
|
|
Don Flamenco
|
A nice example of that is the [Wyrmrest Band], which is quite clearly intended to be ilvl 213 (and drops from Heroic Sartharion), but is in the database with an item level of 200. From this, you can deduce that the item level is clearly a design target and not a rigidly checked constraint.
|
|
|
|
|
|
03/17/09, 9:48 AM
|
#62
|
|
Von Kaiser
Dwarf Paladin
Khaz'goroth
|
According to my graphics calculator (which has never failed me in the past), weapons in LK more closely follow a quadratic model rather than a linear. Taking X as the ilevel, this is what i came up with.
WEAPON DPS = Ax^2 + Bx + C
| | GREEN | BLUE | EPIC | | A | 0.0021806319 | 0.0029076347 | 0.0043624915 | | B | -0.2003571429 | -0.3587891408 | -0.7782980585 | | C | 70.792 | 85.438 | 124.535 |
2-Hand weapon DPS is still 1.3x 1-hand DPS.
DPS on spellpower-sacrificed green weapons seems to be 57.74% of the DPS on a melee weapon.
Spellpower weapon seem to also follow a quadratic function (green SP weaps seem to be in the order of 0.0091x^2 - 0.9184x + 126.7268). Just confirming to myself that they no longer follow a model of Y spellpower per sacrificed DPS.
And also, to calculate armor:
| ARMOR CLASS | GREEN | BLUE | EPIC | | CLOTH | 1.181*X + 6.000 | IF(X<182): 1.630*X -17.771 | 0.544*X + 191.597 | | | | IF(X>182): 1.058*X + 83.071 | | | LEATHER | 2.228*X + 9.643 | IF(X<182): 3.059*X - 33.331 | 1.038*X + 356.474 | | | | IF(X>182): 1.449*X + 255.287 | | | MAIL | 4.946*X + 23.543 | IF(X<182): 6.778*X - 69.918 | 2.308*X + 793.128 | | | | IF(X>182): 3.171*X + 577.865 | | | PLATE | 8.843*X + 41.214 | IF(X<182): 12.095*X - 121.881 | 4.517*X + 1334.562 | | | | IF(X>182): 5.728*X + 1020.709 | |
Edit: Fixed B value for greens
Last edited by Thorgred : 03/17/09 at 10:39 PM.
|
|
|
|
|
|
03/17/09, 10:36 AM
|
#63
|
|
Don Flamenco
Dwarf Priest
Dalaran (EU)
|
Originally Posted by Thorgred
According to my graphics calculator (which has never failed me in the past), weapons in LK more closely follow a quadratic model rather than a linear. Taking X as the ilevel, this is what i came up with.
|
Given that stat budgets are now confirmed to be exponential with ilvl, I'm not surprised that a quadratic curve is a better approximation than a linear one. But it's going to be a bad approximation: armor and base dps is highly likely to be exponential as well, and not polynomial.
|
|
|
|
|
|
03/17/09, 12:07 PM
|
#64
|
|
Von Kaiser
|
Originally Posted by Thorgred
WEAPON DPS = Ax^2 + Bx + C
| | GREEN | BLUE | EPIC | | A | 0.0021806319 | 0.0029076347 | 0.0043624915 | | B | -2.003571429 | -0.3587891408 | -0.7782980585 | | C | 70.792 | 85.438 | 124.535 |
|
I double-checked, and it still seems to me that the function is monotonically decreasing for greens, and yields negative results? Either I made some mistake copying the values (but as I said, I double-checked), or I think there's some mistake in those numbers.
|
|
|
|
|
|
03/17/09, 12:24 PM
|
#65
|
|
Von Kaiser
Undead Priest
Cho'gall (EU)
|
The 'B' value for green is missing a factor 10.
With a couple of tests, -0.2003571429 returns the expected results.
Example for [Blackened Blade]
Ax²+Bx+C = 69.091 -35.664 + 70.792 = 104.219
|
|
|
|
|
|
03/17/09, 3:14 PM
|
#66
|
|
Von Kaiser
Blood Elf Priest
Kirin Tor
|
Originally Posted by Thorgred
Weapons in LK more closely follow a quadratic model rather than a linear.
|
Out of curiosity, did you just calculate based on epic weapons from Wrath or did you include the epic weapons from TBC as well? When I set out to calculate weapon DPS, I thought that, like blues, there would be different equations for TBC and Wrath weapons, but when I actually did the numbers, I saw that they're using the same equation.
This graph is two-handed weapon DPS as a function of item level. It goes from item level 105 up to 226. Seeing the shape of it, I thought it looked like an exponential curve so I tried an experiment and took the item slot value and graphed that against the weapon DPS.
My suspicions were confirmed with it showing what looks to be a linear relation between the item slot value and the weapon's DPS. I'd made these graphs before the new weapons from Ulduar had come out, so they aren't reflected in the equations.
On the topic of sockets, I came across and interesting find while out questing on my death knight. [Stability Girdle] and [Thorim's Grasp] are the same item but with a socket. Thorim's Grasp is short 16 points of Strength as compared to the Stability Girdle. What I wrote in the first post would suggest that the socket would be red, but the socket is blue. Adding in 24 Stamina to Thorim's Grasp leaves it under budget compared to the Stability Girdle. If I add in the socket bonus, however, Thorim's Grasp is much closer to the expected item level.
Another possibility could be that socket color doesn't correspond to a specific stat, but could have 16 points taken from any stat in order to create the socket. As adding 16 Strength using the blue socket on Thorim's Grasp puts it at the exact same as the Stability Girdle.
|
|
|
|
|
|
03/18/09, 1:43 AM
|
#67
|
|
Von Kaiser
Dwarf Paladin
Khaz'goroth
|
The formula i used doesn't fit for weapons back to iLv100 - but that's to be expected, given the small initial sample size. A few changes and the addition of all the 100-145 weapons gets 0.0032535x^2 -0.35255x + 84.9395.
Whilst its probably more likely to be an exponential, i found i got less variance in results with quadratic apporximations (o^2 = 0.99996 as compared to 0.992)
Good find there with those belts - they might be good to experiment with. however, wouldn't you be adding 12 points (as they are green items = green gems?). And i would say that perhaps the socket color doesnt matter: i think thats more likely than the socket bonus counting as part of the item budget,
|
|
|
|
|
|
03/18/09, 2:50 PM
|
#68
|
|
Von Kaiser
Blood Elf Priest
Kirin Tor
|
I think the reasons sockets are so valuable on epic gear is that they're a fixed cost across all qualities of items. Adding a socket to a green item or to an epic item would cost the same amount. This lets sockets become worth more when epic gems come out. I can't remember where I read it, either here or back when 2.1 came out, but sockets used to consume a percentage of an item's budget which was then changed to a fixed value.
|
|
|
|
|
|
03/30/09, 5:47 AM
|
#69
|
|
Glass Joe
|
Socket worth in Item Levels
I have been using a method for calculating item level with sockets, it works for me but its not perfect.
First, calculate your itemSlotValue without counting the gems. Then add 12 for every gem, yes even 12 for the meta gems. Using this new value as the itemSlotValue in the calculation for the itemLevel you get relatively close to Blizzards version of the Item Level.
It doesnt work for every item, but it gets close.
|
|
|
|
|
|
04/18/09, 2:16 PM
|
#70
|
|
Don Flamenco
|
I believe that [Pulsing Spellshield] is underbudgeted for a item level 219 item. Can someone confirm that I performed the calculations properly?
Assuming that the following stats have these modifiers:
Stam = 2/3
Int = 1
Sp = 6/7
Crit = 1
Haste = 1
and assuming the equation for itemvalue is
= [((stat[1]*statMod[1])^1.7095)+((stat[2]*statMod[2])^1.7095)+...+((stat[N]*statMod[N])^1.7095)]^(1/1.7095)
as well as assuming that the itemSlotValue = (itemValue/slotModifier) and that a shield has a slotModifier of (9/16).
[Pulsing Spellshield] has an item budget of 89 stam, 41 int, 46 crit, 36 haste, and 37 sp.
itemValue = 111.66 = [stam(1075.11) + int(571.55) + SP(368.46) + crit(695.8) + haste(457.61)]^(1/1.7095)
the final itemSlotValue = 198.5 = (111.66/(9/16))
Thus meaning that the [Pulsing Spellshield] is itemized for Item level 200 as oppossed to 219 as is it supposed be. If the Spell power were bumped upto 63 then the item level would be 219.
|
|
|
|
|
|
04/21/09, 3:14 PM
|
#71
|
|
Piston Honda
Draenei Shaman
Shattered Hand
|
Originally Posted by PDXMarcos
I believe that [Pulsing Spellshield] is underbudgeted for a item level 219 item. Can someone confirm that I performed the calculations properly?
|
Read the first page again. You need to take the itemslotValue and put it in this formula:
Epic:  - 342.12)
to get the iValue of the item. Do this and you get a value that is approximately 219.
Doesn't make the itemization any better though.
|
|
|
|
|
|
04/22/09, 6:33 PM
|
#72
|
|
Von Kaiser
Blood Elf Priest
Kirin Tor
|
I've been looking at sockets some more and I've come to think Blizzard just takes 16 points out of whatever stat they feel like and turning that deficit into whatever socket color they feel like. Take [Archaedas' Lost Legplates] for example. Strength, dodge, parry, and expertise all correspond to red gems and stamina is a blue gem. There's no yellow stat to sacrifice to convert to a yellow socket. Adding 16 points to any of the red stats puts the item underbudget, but adding 24 to stamina puts the item right on budget.
When it comes to socket bonuses, the point value of the bonus changes depending on how many sockets there are on the item. One socket tends to give 4 points, two give 6 points, and three gives 8 points. For some reason, Blizzard's been giving 8 point bonuses on items with two sockets and there doesn't seem to be any pattern to it. [Conqueror's Worldbreaker Faceguard] has a 6 point socket bonus, while [Conqueror's Worldbreaker Helm] has an 8 point bonus along with the rest of the Tier 8 helms. And then there's the most likely a bug [Guise of the Midgard Serpent] which has a 16 point socket bonus.
|
|
|
|
|
|
04/27/09, 1:01 AM
|
#73
|
|
Glass Joe
---
Undead Priest
No WoW Account (EU)
|
Is it me or Blizzard miscalculated something with mail DPS boots from Ulduar hardmodes? Compare Tempered Mercury Greaves with Greaves of Swift Vengeance - the first one are supposed to be ilvl 226 while the second one should be 13 levels higher. And yet the comparison looks more or less like this:
+12 Agility
-11 Stamina
+9 Intellect
-6 Hit Rating
+4 Crit Rating
-24 Attack Power
which makes them roughly equal as far as item budget is concerned.
Did I miss something here? If so, feel free to correct me.
Last edited by Vysogota : 04/27/09 at 8:57 AM.
|
|
|
|
|
|
04/28/09, 9:57 AM
|
#74
|
|
Shadow Word: Cloud!
|
Originally Posted by Vysogota
Is it me or Blizzard miscalculated something with mail DPS boots from Ulduar hardmodes? Compare Tempered Mercury Greaves with Greaves of Swift Vengeance - the first one are supposed to be ilvl 226 while the second one should be 13 levels higher.
[...]
Did I miss something here? If so, feel free to correct me.
|
I punched the stats from these boots into this excel sheet that I made to evaluate item budgets using the data from this thread... and they both came out to the exact same budgeting, right around ilvl 219. I use a socket value of 16 for budget purposes, however this causes pretty much every item with sockets to come in grossly "underbudget" so something is amiss with that... Regardless, the two items you linked end up using the same total budget (within 0.1 points) Tempered Mercury Greaves = 218.37 (should be 226) and Greaves of Swift Vengeance = 218.32 (should be 236).
What is going on with this? Are socket bonuses factoring into item budgets more than previously thought?
Also, I need to do some of my own figuring on stuff, but the current school of though on sockets does not seem to be working, at least when I do my calculations. For 226 items with sockets, my calculations has pretty much all of them coming in at 218-222. This is with using 16, as I stated above. I have to use a socket budget of 28-30 to get items that are perfectly budgeted or even slightly over budget.
Any one have any thoughts? I can post my rudimentary excel sheet if that would help.
|
|
|
|
|
|
|