gregorian

Hmm good call on the focus stone. Don't care about the spider one because it wasnt in the original trinket file anyway. But yes, it should be sp/haste.

Some suggestions
-option to save stat scalars from gui
-option to export gear set as txt
-possibly add a box for CSD weight, otherwise just automatically blacklist all helms without meta sockets
-I don't feel that it's terribly necessary to design a complicated white/blacklisting system. The current implementation works fine. It's going to take multiples more time to code and and configure on the user end than to just do what I did: just run the app multiple times and click to exclude certain pieces from each subsequent calc run. I eventuallly arrive at a reasonable output within a couple minutes...
-Maybe a find-as-you-type searchbox for convenience, or a check box on each item on the output list to blacklist that item from future runs. eg. check conductive seal and starshard edge and upon clicking 'calculate gear set' the 2 items get shifted from the included list to the excluded list just before the actual run. Simple and reversible

 

Yeda

Regarding to Mac users, MONO seems to work ok with this app until now : )

 

acceleratum

Sundial of the exiled is missing in the item list.
I press get items from wowhead and it just crashes.
I excluded everything, then added only the items I possess, I click calculate and the list only shows 5 items , the rest stays N/A.

I use winxp 32 bits with a quad core q6600 g0.

 

thetrueavatar

FYI, works under linux with wine. Redirect to wowhead doesn't works on my computer but maybe I need some wine tuning.
Here are my suggestions:
-right click on BIS item could offer the possibility to directly exclude that item.
-add source of the item(boss, hard mode ?, normal/heroic).
- a preview of the item stats
- show the hit/non-hit item that have the higher score for a slot.

Very nice job otherwise.

 

Axodious

Staves are being enchanted with Mighty Spellpower rather than Enchant Staff Greater Spellpower. Also clicking on the Mighty Spellpower link takes me to an invalid WoWHead item number 60714.

Suggestions:
- Jewelcrafting Gem option
- Whether or not to use hit gems or choice of common Warlock gems
- Receive PTR loot from WoWHead

 

Todgerdodger

As is The Egg of Mortal Essence.

 

Viper007Bond

Default hit is way too high. The destro spec only needs 341 for example if you have full raid buffs.

 

ashtray

If you're Alliance. If you're Horde you need 14% from gear/talents. Also, not all raids have enough draenei to go around. I know mine doesn't. I often find myself having to gear for 14% as opposed to 13%. So I imagine setting the 14% default is the safest bet for optimizing gear. If it's a different value for you there's nothing stopping you from changing it.

 

Tauraherion

Since the patch, new items were loaded into the optigear. And now it keeps freezing when calculating item slot 18. anyone know how to fix this?

Also, Venombyte, do you have another version of this coming up?

 

Axodious

Another suggestion in addition to these would be to add MMO-Champion item database support.

 

Mezmaron

Whenever I try to get gear from wowhead the program crashes and nothing is added. Is there anyone who would be willing to put up a version that includes items from the newest patch?

 

VenomByte

I'll have a look at this.

Also, is anyone willing to update the trinkets list? I'm not entirely sure how to approximate best their values as straight spellpower/haste etc

 

mjball

Just downloaded and gave OptiGear a shot. Very awesome! Looks like it doesn't include any of the items added in patch 3.2. When do you think those will make their way in?
Thanks for a great little app! I can't wait to try it out on my Mac using Mono.

Edit: a feature I'd really like to see is an import/export gear set (i.e. as a wishlist), either to an XML file or just to a wowhead profile. Also, in the interest of efficiency, you may consider completely omitting (not even retrieving from wowhead) any obviously non-lock-y items, such as those with strength, agility, armor penetration, attack power, dodge/block/parry, defense... example: Unsmashable Heavy Band is included by default. Either exclude such items by default or don't even present them as options.
Yes, I know cloth items will never be itemized for physical DPS or tanking, but there are other items that can be: trinkets, cloaks, rings, and necks.

 

Tauraherion

I think that the best way to approximate the DPS contribution of a trinket (be it On use or a Proc Chance trinket) is to use the formulas:

For proc chance trinkets:
(stat_gained*specdpsmodifier)*duration/(internal_cooldown+frequency/proc_chance)

For on use trinkets:

(stat_gained*specdpsmodifier)*duration/(internal_cool down)


warlocomotiff used these formulas in this post:

Trinket Discussion

 

• Heeno

He is talking about the new trinkets in the Coliseum: Reign of the Unliving - Items - Sigrie.

 

mjball

When I run this, calculating item slot 18 (and some other item slots between 18 and 22) does take a very long time, but just let it run til it finishes. It'll get there eventually.

Venom - I think we could really use an update to this now, for the T9 set boni in particular. Also, it's taking way way longer to calculate gearsets than it should, IMO. Maybe an hour or day is what we should expect given the large number of items? I am guessing that, because Optigear is an exhaustive search, before pruning, this problem's complexity grows exponentially with the number of possible items.
A nice new feature might be a "cancel" button for whatever calculation is in progress.

 

VenomByte

Generally speaking, what really impacts the speed is
1) Tier bonuses
2) Meta gem calculations

It's not so much the number of items that increases the search space exponentially, but the number of 'competing' items - i.e, the number of realistic item choices for a given slot. Weaker choices will get discarded with negligible effect on run time. With the above features however, it becomes much more difficult to conclusively discard a weaker item early in the search process, hence increasing the time taken.

Without either of the above, it's many orders of magnitude quicker. Until I've got the T9 bonuses in, you might find it as well to disable the T8 & T7 entirely. On that note, can anyone confirm whether Conquerer's Regalia stacks with Triumphant Regalia for the purposes of generating T9 bonuses? I would think not from the wording, which is a pain because it means I'll have to track two sets of T9 gear.

I'm working on the incorporating the T9 bonuses at the moment. As soon as I have a working version, I'll upload it.

EDIT: Done now, so a new version has been uploaded

v0.69 - 18/8/09
- added T9 triumph tier set bonuses
- updated gems list
- updated gear list
- adjusted pruning process to speed up app

Note that the conquerer's T9 bonuses are not yet included. Nor have the trinkets been updated, though other gear & gems have.

The last point on pruning is worth explaining in some more detail. What I've done is to remove 'equivalent' setups during the optimisation process - gear sets which have identical weigh, hit, meta, tier x, etc to another different set. In practice this means identical items with the same gems in different orders (provided socket bonus matches) and the new horde/alliance colliseum gear which the app otherwise would include both of. This makes a huge difference to execution time.

On my 4 year old PC, a full search of all gear for the default affliction spec (including tier bonuses) takes ~2.5 min. It is approximately 4-5x slower for deep destro. This is roughly the execution time you should be seeing now.

 

mjball

Thanks, Venom! I'll give the new build a shot.
Incidentally, my newish loaded work PC (Vista x64, 6 gigs ram, 2.5 GHz core 2 quad, etc) had a lot of trouble running the last version of OptiGear whereas my 3 year old Macbook running Leopard had no trouble at all (using OG under Mono).

Edit: Glad to see you threw in more weight profiles but I don't think they're the very latest from simcraft (are they?). Then again, I'm not even sure that the 00/13/58 weights I posted (in simple questions/simple answers) are 100% accurate for 3.2. *sigh* You might want to update profiles with the t9 entries as well - they're not in there, but I added them myself and OG reads them out correctly.
Also, definitely need some trinket updates. I noticed Abyssal Rune wasn't in the trinket file; there are probably others missing as well.

Question: what's the difference between checking the "include Tn Bonuses" box and setting the MxTn Weight values? Not really clear on that one.

BTW, if you'd like to offload any of the work with further development on OG, I can lend a hand (even if it's just stuff like maintenance of trinket and profile data).

 

xonelith

Excellent tool, particularly for quick decisions on upgrades during raid.

A few bugs that have already been mentioned, once corrected, would be perfect:

Not a huge deal, but could change the decision as to which item is better, based on a different enchant. Using a staff, it still selects the +63 spell power enchant. Removing the +63, removes all gear except for the staff (odd).

Another minor one I would like to see fixed. There are times when we don't have a boomkin or a SP and need to gear for the full hit amount (10 mans usually).

Thanks for all the work on this!!

-xone

 

mjball

Venom, here's some more feedback on OptiGear for you.
Bugs:

• OG is definitely not loading all of the items it should/could from Wowhead, particularly the latest-and-greatest from the new raids. For instance, it doesn't grab Felspark Bindings or Symbol of Transgression. I'm curious to know how OG queries items from Wowhead - could it be a caching issue, or just a result of a limited-scope Wowhead search/query? It would be a big improvement to be able to grab every reasonable item from Wowhead.
• After loading gear from Wowhead finishes, it looks like the textual progress meter leaves behind the string "[]" where there should be no text.

Some really nice usability improvements (not to add to your queue or anything):

• Ability to filter by item level (or other criteria I haven't thought of)
• Ability to grab & lock a full set of gear from a character's wowarmory or wowhead profile data
• Ability to select more than one item at a time for inclusion/exclusion

Additionally, the v0.69 update fixed whatever slow run & freeze issues I was experiencing running OG under Vista x64. I am guessing it did the same for others.

 

gurtinu

I've created another program which takes all the gear and sorts it according to a provided scale for each gearslot. I have not spent a lot of time on the program since I just wanted to achieve this basic functionality figuring it might be useful. However I have provided the source code so you are free to customize it if needed (and recompile if you don't thrust me It's created using java and as I said it was done hastily so I did not bother with sticking closely to nice coding standards so expect the code to be as ugly as the gui. It also utilizes the same gearlist.xml file as optigear so when it gets updated it should be usable for this program to. So feel free to try it out if you wish.

http://rapidshare.com/files/269633128/locklist.rar

 

banginosas

Nice program indeed. After reading about BiS sorting problems using scale factors due them scaling non linear and being cross-linked, I came up with idea using Simucraft. What do You think, would it be worthwile write code, which tries every possible gearset with simucraft and then picking one with highest dps?

 

mjball

It depends on how long a simcraft run for a single gearset would take. There are a HUGE number of possible gearsets, so a brute force approach like this could easily take a prohibitively long time.
As a very simple example, we could imagine that there are only five items we would consider per slot (ignoring constraints like multiple rings and trinkets, and one 2H weapon vs. one 1H and one OH). With 17 relevant gear slots (shirt and tabard not contributing), there would be 5^17 = 762,939,453,125 possibly sets. Unless simcraft can run, say, 1 million or 1 billion possible gearsets in some reasonable amount of time, there are just too many sets.

 

banginosas

True, I have thought about that. One thing - if program counts hits (considering at least 13.8% or so) AND some pre-made assumptions (like 258>245 ilvl gear) needed calculations could be reduced considerably AND program could be used in cloud computing - if 10 people would contribute, then calculating time would be 10 times shorter

I even thought run simucraft for 100 loops and then re-run it for 10k loops for best 10 sets or so?

 

VenomByte

Any time you make an asssumption (such as 258>245 always) you're defeating the purpose of creating such an application in the first place. It's easy enough to make a pretty good assumption about an entire BiS gear set, and you likely won't be too far off. The point of using a computer is to avoid such assumptions.

Programmatically, a better idea would be to record spellpower/haste/hit/crit/spirit/int against a given gearset instead of the weight/hit OptiGear uses, and prune accordingly. Then you would run your analysis program of choice against x resulting gear sets returned. Even so, I'd estimate you'd still have an unfeasibly large search space unless you could also trim the possibilities according to provable mathematical rules on the relative weights of attributes, e.g 'haste can never be worth more than spellpower', or 'int will never be worth more than 0.3dps'.

I just don't think you can attack such a problem without far more solid knowledge of the relationship between each of the attributes and dps output to assist with the pruning process.