The best fit came out as a Polynomial function to the order of 5 (minimum), however I dont have that graph available to me right now.
Feel free to critique as necessary
Looks like two straight line segments to me. The only weird thing is the transition point between the two, where the behaviour is currently undefined since you have too few data points.
Well thats the thing, prior to T9, its one segment, but it matches a logarithmic function mostly (though like you said, too few points), where as the 3.2 T9 sets match an exponential equation...I guess they have shifted the numbers to accommodate the final end game for Lich King.
With the change to Block Value in 3.2, is the statMod for BV now halved? Would it be 13/40 now, or is it some other value?
Also I plotted the armor values for plate starting at iLvl 200, and ending with the new T9 Heroic 258s, and came up with an interesting graph:
The best fit came out as a Polynomial function to the order of 5 (minimum), however I dont have that graph available to me right now.
To me it looks like 2 linear plots where they decided to basically roughly (probably including various rounding errors) double the increase of armor on items beyond the ilevel 232 boundry compared to the existing WotLK formula. Perhaps this is due to feedback on the slow scaling of Armor compared to previous points in the game's itemization.
Blizzard has tried many times to bandaid the problem of itemization that plate should be worn by plate classes and not leather/mail. First, by giving extra AP per point of armor through talents, favoring plate for DK's and Warriors. Second, by giving rather large +STR gains through talents for Warriors, Paladins, and DK's, favoring STR not AP items. Third recently for warriors, by increasing the effect of Armored to the Teeth. For DK's this is not much a problem, as they rely too heavily on strength. Paladins still can pickup a non-STR item or too if it has perfect itemization for them. For warriors though, it is still such a huge problem. They have talked about letting certain classes get 2AP from agility instead of 1 and removing AP from AP/AGI items, but I don't think that would be a smart change as it would reduce these items down to 4-stat itemization and crush leather/mail class's dps.
To fix the root of the problem without changing how Rogues/Druids/Hunters/Shaman get their AP requires a small change in the itemization of the STR gear. STR items need more itemization points first and foremost. Taking a double hit on less points overall and more of those "less points" towards stamina is the definite culprit. Let's see what happens when we give the STR item the extra 25 pts of stats and distribute them in three different ways to see which looks the most balanced
1. Evenly between the 3 dps stats:
Godbane
Sif's
New Sif's
Total Points
229
204
229 (+25)
DPS Points
187
156
181 (+25)
Stam Pts
42
48
48 (+0)
DPS Diversification
4
3
3
STR
72
81 (+9)
AP
126
AGI
46
Crit
42
42
50 (+8)
ExP
36
42
50 (+8)
STA
63
72
72
This item seems a bit too overpowered and looks like a 239+ iLvL STR item without the stam.
1. Across the 4 stats, keeping STR and Stamina matching:
Godbane
Sif's
New Sif's
Total Points
229
204
229 (+25)
DPS Points
187
156
174 (+18)
Stam Pts
42
48
55 (+7)
DPS Diversification
4
3
3
STR
72
82 (+10)
AP
126
AGI
46
Crit
42
42
46 (+4)
ExP
36
42
46 (+4)
STA
63
72
82 (+10)
This item looks balanced between STR and Stamina, but the DPS stats on it are not very diversified at all and it has way too much attack power compared to the other stats. Keeping STR items as 4-stat items is definately not the way to go.
Another idea would be to treat Crit rating like Agility on AP/AGI gear and give STR items the 5 stat rule. Since AP/AGI are two high priority stats for Rogues/Druids/Hunters/Shaman, and STR/Crit are high priority stats for Warriors/Paladins/Death Knights, it seems fair right? Let's take a look at what happens to Sif's Promise when we make it like Godbane Signet:
New Sif
Godbane
Stat
Amount
STR
63
Stam
63
Crit
46
Exp
36
Hit, Haste, or ArP
42
Stat
Amount
AP
126
Stam
63
AGI
46
Exp
36
Crit
42
Not only does the new STR/Crit item have the same amount of stat diversification as the AP/AGI one, it is as balanced in terms of itemization. Yet, the new Si'f Promise with this rule looks too overpowered in general. What could be done further to balance it is to take away some points from the secondary stats and put them back into stamina to retain the "STR items have more stam in general" rule. Also, when changing the items to a new version, you could even retain the amount of stamina on each item.
New Sif
Godbane
Stat
Amount
STR
63
Stam
72
Crit
46
Exp
36
Hit, Haste, or ArP
36
Stat
Amount
AP
126
Stam
63
AGI
46
Exp
36
Crit
42
Let's look at an example where a Body Armor piece was changed to give it the same amount of itemization points as a same iLvL leather item while retaining its stamina:
New Conqueror's Siegebreaker Battleplate
Conqueror's Siegebreaker Battleplate
STR
97
113
Stam
129
129
Crit
81
58
ArP
58
74
Hit, Haste, or Exp
58
Let's compare the new plate setpiece to the existing Rogue setpiece:
New Conqueror's Siegebreaker Battleplate
Conqueror's Terrorblade Breastplate
97 STR
97 AGI
129 Stam
90 Stam
81 Crit
162 AP
58 ArP
66 Crit
58 Anything
76 ExP
Compared to the leather, the Plate gets some stamina for some other secondary stats.
When I did this to miscellaneous plate items and compared them to leather items of the same iLvL they appeared to be about 10% better than the leather for a warrior, which is just enough for them to choose it over non-plate (since they want STR/Crit instead of AP/AGI).
New itemization Proposal:
Warr/DK/Pala
Hunt/Rog/Dru/Sham
STR
AP
Crit
AGI
Stam(+)
Stam
secondary(-)
secondary
secondary(-)
secondary
How would this affect DPS of plate wearers to non-plate wearers? If this itemization change put plate classes into doing too much DPS, certain talents could be slightly adjusted a tad to return the DPS back to what it was intentionally.
Is there any way to get a list of random properties for the item other than parsing the randomPropertyEnchant text in armory XML? E.g. is suffixFactor related to the item level in any way?
Edit, to clarify, can I use the inverse of item level formula in original post to reliably calculate suffixFactor given item level?
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.
Unfortunately the higher the iLevel the margin of error of your result increases considerably which is fine, considering it's been so long since your post.
One of the problems I see with calculating DPS is that there's a considerable margin of difference in-game for items of the same item Level. At first I thought this difference was caused by fluctuations in Attack Speed but there are cases of weapons with the same Attack Speed but up to .3 DPS of difference. which doesn't really help our calculations. As Thorgred pointed out, the itemLevel we know is probably rounded up.
Since picking up an arbitrary value wouldn't make much sense, I've reached a different quadratic formula for epics. First I calculated manually the true DPS from the weapons and divided by the sample amount:
Where x is Item Level:
DPS = 0.005789383933253384x^2 + (-1.3909657021268087x) + 189.9220636880327
With this formula we have the following results:
ItemLevel
Result
In-Game values
Rounded Result
in-Game Median
Difference
200
143.47
(143.2 - 143.5)
143.5
(143.4)
+0.1
213
156.4
(156.4 - 156.7)
156.4
(156.5)
-0.1
219
163.0
(162.9 - 163.2)
163.0
(163.1)
-0.1
226
171.31
(171.1 - 171.5)
171.3
(171.3)
0
232
178.86
(178.7 - 178.9)
178.9
(178.8)
+0.1
239
188.19
(188.0 - 188.3)
188.2
(188.2)
0
245
196.66
(196.4 - 196.7)
196.7
(196.4)
+0.3
258
216.44
(216.4 - 216.6)
216.4
(216.5)
-0.1
In every case, the resulting value falls within the known DPS range for a given itemLevel.
My opinion is that if weapons follow a quadratic model, we'll probably never be able to figure out the exact formula used because of the fluctuations.
I found a bug, it seems that iLvl 245 Two-Handed Weapons have their DPS calculated as iLvl251 in the game. Not sure if this is done on purpose or not by the team (to close the gap of 13 item levels to 7), but it's definitely weird.
Edit: they fixed it on PTR.
I also made an item calculator that's working perfectly with about every physical weapon in the game. I'll post if anyone's interested.
Any attempt to use a quadratic equation to model weapon DPS based on item level will fit well for the data used to find the equation, but outside of that it will be wildly off. Your equation works well for the items with levels 200-258, but if you try to use it to calculate the dps of an ilevel 100 one-hander like [Blinkstrike], you'll come out with it having 108.7 DPS as opposed to its actual 81.2 DPS. The reason for that is because weapon DPS as a function of item level is an exponential equation. Earlier in the thread I posted the reasoning behind that, but the short version is that the item slot value, the number that converts to item level, has a linear relationship with weapon DPS.
With the change to Block Value in 3.2, is the statMod for BV now halved? Would it be 13/40 now, or is it some other value?
Yes, I'll have the table in the first post updated to reflect that. It looks like this change is universal as even [Pauldrons of Wrath] have had their block values doubled.
Any attempt to use a quadratic equation to model weapon DPS based on item level will fit well for the data used to find the equation, but outside of that it will be wildly off. Your equation works well for the items with levels 200-258, but if you try to use it to calculate the dps of an ilevel 100 one-hander like [Blinkstrike], you'll come out with it having 108.7 DPS as opposed to its actual 81.2 DPS. The reason for that is because weapon DPS as a function of item level is an exponential equation. Earlier in the thread I posted the reasoning behind that, but the short version is that the item slot value, the number that converts to item level, has a linear relationship with weapon DPS.
Yes, I'll have the table in the first post updated to reflect that. It looks like this change is universal as even [Pauldrons of Wrath] have had their block values doubled.
I agree with you. It doesn't make sense and the calculations in the Op provide a better result considering the entire game, although they provide a margin of error.
However I believe that the weapon DPS and SpellPower don't follow the ItemSlotValue of the weapon itself, but rather the ItemSlotValue equivalent to the rounded value of the itemLevel. This seems to make sense if we observe some weapon DPS bugs (they were not calculated according to the true iLvl of the weapon, but rather some arbitrary value someone input wrong in the database).
I wrote a small app to brute-force some of the medians of the 1h weapons and I came up with a different value for the linear progression between item slot value:
For itemLevel to DPS we just invert the normal formula:
Regardless, there's still some oddities in the system that are not explained by anything, such as two identical weapons having a slightly different dps but this formula provides according to my calculations a slightly more precise approximation to what is seen in game.
Regardless, there's still some oddities in the system that are not explained by anything, such as two identical weapons having a slightly different dps but this formula provides according to my calculations a slightly more precise approximation to what is seen in game.
A theory I've had is that weapon DPS should really be rounded to the nearest whole number when it comes to weapon DPS. I imagine Blizzard's process is they give a weapon its item level, choose a weapon speed, then give it its listed damage range. If the average damage over speed is too far off the designated DPS for that item level, they probably tweak the damage range until it's around 0.5 DPS from the designated DPS for that item level. Blizzard's not afraid of using integer values for weapon DPS as evidenced by the heirloom weapons, but I imagine the slight variance in normal weapons is due to adding a bit of variation to existing weapons beyond just the differing stat spread.
In a different subject, 3.2 has updated all the greens, and presumably the blues, with random enchants to using the logarithmic formula that quest items had been using. While confirming this with the greens I saw last night, I noticed that an item level 178 cloth robe, the Vizier Robe of the Sorcerer, was using the new formula for stats, but treating Stamina as if it had a 1 point value rather than a 2/3 point value. Whether that's unique for this robe or all items with the of the Sorcerer suffix hasn't been tested.
I'll confirm that blue WotLK suffixed items (mostly dropped by rare spawns) have been upgraded to follow the logarithmic model in 3.2.
Example:
My DK had a [Titanium Brain-Gear] of the Champion that weighed in at 29/44/29 Str/Stam/Def pre-patch. It is now 36/55/36 which works out to an ilvl of 182.25 of an expected 183 using the logarithmic equation for rares.
Has anyone written all of this into an addon? I'm using the since-updated ItemTips from earlier in the thread (or another), but that was still many months ago and has broken because of API changes in recent patches.
Hi,
I got a question related to item-ids and unique/unique-equipped. As I am not sure if this is topically ok to post it here I am sorry if it's wrong.
Here's my problem:
As we all know there are lots of items which are unique so that you can have it in your inventory (and so on your actual equip) only once, for example [Blood of the Old God]
Then there are items, that are only unique-equiped, like [Loop of the Agile], so items that you can wear only one at time.
Than there are items that are (e.g.) "unique (100)" like the marks from battlegrounds. You can only have the given number of times that item.
First conclusion of mine
All these "unique" things are related to the item-id of these items. so that you only can have/wear one of these item ids at the same time
Let's go on:
In 3.2 there is/are the new raid instance(s), trial of the (grand) crusader and there are normal (ilvl 245) and heroic drops (258) [for 25man]. As we know from the PTR we go different items with different stats (related to ilvl) but with the same name. The normal item and a ilvl 258 item from the grand-heroic mode.
Now the are slowly getting to my confusion
First thing I noticed: There are items which are "Unique: <itemname> (1)", e.g. "Unique: Juggernaut's Vitality (1)" 245 , 258 (both links mmo-champion-db)
Second conclusion of mine
Items that have an Unique: <itemname> (number)" you can only own (= have in your inventory) an item of this name for the given number, unrelated from the itemid. Obviosly it's a mechanic from blizz to prevent players wearing a very strong combo of normal and heroic items. So far, so good, so clear!
Now my main problem/confusion
There are also items with same name, different ilvl and different itemid which are NOT "unique: <itemname> (number)", for example Band of the Traitor King 245, 258 (both links mmo-champion-db)
As we see in the example: Both rings have the same name, but one (ilvl 245) is unique-equiped, the other one (ilvl 258) is unique.
If I go straight from my first conclusion where I think, that unique and unique-equiped are based on itemid, that would mean, that I can wear these normal and heroic drops that got the same name, but are nor "unique: <itemname> (1)" together and get a combo of normal and heroic drop?
I am far from being 100% sure, but on the one hand I am sure about my conclusion from the unique, unique-equiped thing, but on the background of such strong combos I am not sure if these thouts are still right. So I am sure on a logical way, but not on an balanced way [or should I say blizzards point of view?]
The question for me is wether or not an unique item is also unique-equiped. If thats true than you could have as many ilvl 245 Band of the Traitor King as you want but you could only wear one at a time. But as sone as you git an ilvl 258 Band of the Traitor King you can no longer loot any new Band of the Traitor King including ilvl 245 ones but you could still only wear one of them.
Of course that makes no sense whatsoever with the Juggernaut's Vitality you posted
It seems likely that the intention is to preclude anyone from wearing two of the "unique-equipped" items of the same name, i.e. the heroic and non-heroic versions. It's a lot less likely anyone was intentionally precluded from looting one. In fact, there might well be tooltip glitches or item-database errors that cause trouble. If the 258s are "unique" opposite their 245s, that would make looting a 258 when you have the 245 already a nightmare. "Object cannot be looted."
I'm really quite unclear what place "Unique" has in the game other than to remind you that you already have the item even if you're oblivious. Unique-equipped seems to solve the power-balance issues, which is where Blizzard has been going with that tagging for some time. I don't think anyone really objects seriously to being unable to use the same ring in both versions -- i.e. they might want to, but they understand why it's not allowed. But if there are looting problems once heroic opens, that will be highly unpopular.
First thing I noticed: There are items which are "Unique: <itemname> (1)", e.g. "Unique: Juggernaut's Vitality (1)" 245 , 258 (both links mmo-champion-db)
I would caution against drawing conclusions until we can actually see the items on live. WoWHead, for example has [Juggernaut's Vitality] just marked normally unique, but it also has the ilevel 258 version listed, so I'm not quite sure where the data is coming from. A screenshot of the in game tooltip on live would be helpful if someone has one.
Originally Posted by Crowl
If you have to control a robot dinosaur that fires lazers and there's a time when you shouldn't be shooting those lazers then the encounter is clearly flawed beyond hope of fixing.
Please add armor to the charts to cover cloak enchants, glove engineering, armor kits, bonus armor on PvP armor, some tank items like Saronite Plated Leggings, and bonus armor on jewelry and trinkets. Its value is 1/14 according to popular belief.
Has anyone written all of this into an addon? I'm using the since-updated ItemTips from earlier in the thread (or another), but that was still many months ago and has broken because of API changes in recent patches.
I updated the ItemTips modifications that I made before, it doesn't seem to be kicking errors every time you mouse-over something. Unfortunatly, it is still kicking an error on an Heirloom level item.
Is there a formula known to calculate the expected armor value of an item based on iLvl/armor class/slot ? The way the addon parses the default tooltip, I can't figure a way to determine if there was bonus armor on a given item, and as such various items such as [Saronite Plated Legguards] will show up as under budget.
I updated the ItemTips modifications that I made before, it doesn't seem to be kicking errors every time you mouse-over something. Unfortunatly, it is still kicking an error on an Heirloom level item.
You should be able to put some logic in your updated ItemTips to make it not throw errors on heirloom items as a stop-gap measure.
Field 3 of GetItemInfo() will return an integer for the item quality. 7 is for heirloom items.
First thing I noticed: There are items which are "Unique: <itemname> (1)", e.g. "Unique: Juggernaut's Vitality (1)" 245 , 258 (both links mmo-champion-db)
Assuming these items do work as they sound like they would, preventing you from using 2 of the same name but different iLvl, once faction changes become available it would potentially become possible to equip 2 heroic versions (one of each faction) because they'd have different names in addition to different IDs.
That issue will likely be avoided by converting all faction-specific ToC items to the other faction's equivalent once you do a faction switch, but if they overlook it and continue with faction-specific items (I'd expect the current ones will be rather dated by the time swaps are available), it would make for some rather ridiculous ring/trinket combinations.
I updated my item spreadsheet. These are the new changes:
-- Added a comparison tool. It allows you to compare between two different items, and to compare different item upgrades. I made it because I wanted to compare different items of same itemLevel through different stat weightings, something I haven't found yet outside of addons.
-- Added a Load and Save functionality.
-- Changed some calculations and fixed some bugs.
My TODO:
-- Add more functionality to the item types other than weapons.
-- It should be easy right now to create and modify the items and stat weights, but I intend to make a macro to allow easier edition.
-- Add more comparison templates and fix the graph so it will work with templates other than DPS.
-- Expand the main calculator so it will show all attributes.
-- Finish a functional trinket calculator.
Not sure if anyone's interested, since this is more of a personal project, but it's considerably better than the first version. Anyway suggestions are appreciated.
Edit: Just updated it again. Should have some new fixes.
Back from my ban. I have something to share this time.
I am studying software engineering, and learning about Java GUI's as I read this forum, so i decided to try my hand at a GUI interface program for constructing items and checking their iLevel.
You will need Java 1.5 or later installed on your machine - to run, you may have to right-click "Open With" and select Java from the list.
Here is a picture of the interface:
This is still in the early stages but the basic functionality (set up the item quality/type/slot, add stats, add gem, check the iLevel) is there. No calculations for armor or DPS as of yet. In future builds users will be able to use custom constants as well, but for now I'd like comments on the current functionality.
Is it clear? Easy to use? Any bugs or odd behaviors? please list here and i'll endeavor to update.