◊ Kalman

Aldriana: Best data I'm aware of at present suggests a 25%/25% glancing mechanic for rate and severity. I'd use that for now.

 

◊ Shaker

A few things now that I'm looking over .4:

1. For some reason no matter what I do, I can't get the Thundering Skyfire Gem to activate, and even when I changed the formula to correctly count the colored gems out of their correct cells (and got activate to change to 1) it didn't add the EAP to the talent_gear sheet.

1a The activation formula for Thundering Skyfire seems to be set at "requires at least 3 gems of every type" - the screenshots I've seen (albiet, I might be missing something having been changed on test) show that it requires at least 4 each of red/blue/yellow gems.

1b Shifting Nightseye isn't being counted as a red gem, when it should be. (matches red/blue).

2. Suggestion that you add Drakefang Talisman to the trinket sheet. With my gear/spec, it's better than the Hourglass (though, that's because my crit rating is in the toilet, but still), and it's roughly equivalent to the Bloodlust Brooch. I've already added it to my sheet, so I suppose that's totally up to you.

3. Another suggestion: Changing the suggested upgrades so it shows the 3 next best pieces? My group is currently clearing Kara, so while I know the Slayer's set is better, I also don't have any realistic expectation to be seeing Slayer's gear in the near future. I totally accept that this could be a "this sheet is designed for people who are w/in a tier or 2 of the top end gear", but I thought it might be a nice addition for those of us who aren't.

 

Safid

Is there any intention of including some alternate specs, by which I really mean Mutilate?

Bleedo's sheet (actually Miq's latest sheet) has me at 90-95% of what I could theorycraft as combat daggers using a Sunblade offhand depending on raid buffs. To me that's high enough to be included as 'viable in PVE' especially with a lot of fights that decrease a rogue's TOT during which the extra yellow damage and fast CP generation of the build shine.

I appreciate all the work you're doing and just wanted to know if a) the cycles could cope with a radically different set of talents and b) if there was a plan to do so eventually.

 

◊ Shaker

Oops, found something else.

Current vs Optimal at the top of the Talents_Equipment page, Optimal doesn't include optimal trinket values, so the comparison is reasonably flawed.

This sheet really looks good though, and I'm excited to see it coming into its own. It's definitely a very useful tool, and all you guys' work is really appreciated!

 

• Aldriana

We're actually working towards a different solution here; the goal is to add filters based on zone, so you can pull up a list of "best upgrades from Karazhan" or "best upgrades from SSC". Not sure when we'll get around to implementing such a thing, but it's something we've discussed and are hoping to move towards.

 

• Aldriana

The cycles are currently coded to very specific rotations; they couldn't even really do Shiv or Hemo, much less Mutilate right now.

As for the potential for adding such a thing: it would certainly be possible to do so, but it would be a fairly significant amount of work. I was currently planning to try to clean up the existing features some more (i.e., working through the "known issue" list) before I worried too much about significantly expanding the functionality. However, if I hear a high level of demand for alternate builds, I might reconsider that course of action.

 

Graecus

I suggest keeping the option to upgrade in as it currently is, but expand upon everything available, or at least a top 5, but I think I realize on my own that with WW shoulders, t5 and above shoulders are an upgrade. :P

Precision in Paradise

 

Darlal

As Aldriana stated, we're working on some more fine-tuned gear upgrade recommendations.

In terms of trinkets being counted in the gear values, the current trinkets have a large effect on the relative values of the stats, so in that sense they are included.

I will look over TSD issues, I took my 3 number from wowhead (wasn't on ptr), but I'll double check the actual values tonight.

I didn't make any updates yesterday due to patch excitement. I'll have a small amount of time today, but I wouldn't expect anything new until tomorrow realistically.

In terms of mutilate (which i also use), it would be possible to re-do the cycles for mutilate, but that would be a totally different sheet.

Thanks for the shifting nightseye catch! All gems currently should be added to each color they correspond to (notice glinting noble topaz is there twice).



In general, keep in mind that all upgrades are scored using your current items. That means that if changing trinkets would seriously alter the relative value of stats, the optimal upgrades might change. There are ways for me to suggest overall optimization but they are (a)much better done with a database (if this seems worthwhile I wouldn't be opposed to it), (b) require cyclic arguments, making the sheet very sensitive and especially platform dependent, and (c) make the final results a bit harder to interpret.

Before making final gear decisions I would change balance-altering items (TSD, warp-spring coil, etc...) first and see if the optimal upgrades remain fairly steady.

 

• Aldriana

I've actually toyed with the idea of moving away from a spreadsheet format to avoid some of the limitations of the format; I think the ideal tool would have an initial UI akin to the old CTProfiles site, i.e., a full talent calculator and character sheet page where you can enter your gear and talents. It would probably also want to have an option to fetch your current stats and abilities from the Armory. Then, under the hood, it could go through and do all these calculations in painstaking detail, and provide summary pages indexed by more or less anything you wanted to sort by. On the whole, I think it would make a pretty cool tool, and while I'd be happy to code the underlying computations, my GUI programming skills aren't up to the level that would be required for the project. But if anyone wants to work on putting such a thing together, give me a whistle .

In terms of optomization strategy: figuring out what, in the long run, is the highest dps option tends not to be particularly hard; the challenge is figuring out how to get from point A to point B. I generally take an incremental approach and start with my current gear, then add in the largest reasonably obtainable upgrades and see how much this changes things. In general, the answer is "not too much", but there are some effects (Dragonspine Trophy and TSD, notably) that have a huge impact on the value of stats (hit, in this case) that can change the order of things.

However, the general optomization strategy for someone in early 25man content tends to work out to:

1) get Dragonspine Trophy
2) get to the hit cap
3) profit

 

Darlal

My UI programming skills are pretty much nonexistant in terms of webforms, but I work as a databases (typically access) design consultant in my free time so that part of the project is totally possible. I've really been feeling the limitations of excel with this project, so if there is support for moving to a more robust format I'd be totally for it.

Advantages:

a) eventual merging of pf's dps caculator with gear optimization
b) easier to follow what's happening behind the scenes
c) vastly superior UI
d) real modeling options

Disadvantages:

a) less universal than a spreadsheet
b) real coding involved
c) requires more community support

Anyway, I'd like to hear what people think.

 

• Aldriana

I'm not sure I'd actually set it up as a webform - it's certainly an option, but I was originially thinking in terms of a program that you'd download and run locally - although that would require either bundling with a database program (probably sqlite) or getting by without a database per se (which would probably be possible if a bit messy). Running it as a webform would defenitely have some advantages in terms of setup, although also would require resolving issues of hosting and whatever.

I assume by advantage A you're referring to the ability to support the full range of specs and abilities? I'd personally be reinclined to redo all the calculations from scratch rather than using his particular formulas, to take full advantage of the flexibility that actual code gives you. That's to some extent what I was doing here, but the nature of the spreadsheet is such that I had to limit myself to a small slice of rogue-dom. With the power of actually programming behind it, it would be straightforward to have calculations at or above this level of detail with all possible specs - although it'd take some time to code.

Regarding the disadvantages: I don't think a) wouldn't really be a problem; we've already started to run into compatability issues with spreadsheets due to the differences between Excel and OpenOffice; I think getting to the same level of compatability would be pretty doable. A webform would certainly be able to do as well, and even a locally run program could be structured to work on most flavors of Windows and MacOS with relative ease. Point b) is something of a concern - it would take some time to put such a tool together, but it sounds like we've already found a couple people willing to work on the back-end; if we could find some people to write the front-end this might be possible.

 

Darlal

I'm a bit torn on the best method.

While my personal taste screams "build this in access in an afternoon!!", I know that not everyone has access. I really shy away from making people get a database program, although I guess shaking the rust off my mysql or slite wouldn't be bad.

On the other hand, making this web based would probably be the best--but also most time consuming--way to go.

The ability to really model all specs/gear is really, really tempting.

 

Graecus

We frequently use Frontpage to build an ASP frontend to our dbs in Access using VBScript. Then you'd just need a host (bleedo? :P).

Precision in Paradise

 

• Aldriana

I daresay that it'd take longer than an afternoon anyway; just to get decent models for all the specs would probably take a couple weeks of work - I might be able to do it in a week if I could work on it full time, but between job and raiding that just isn't going to happen.

 

Darlal

Oh yeah, the modeling will take longer than that. I more meant get a general framework setup.

Anyway, on another topic. I've been thinking alot about the zone-selection option. Instead of going with zone-selection I was thinking about using item ilvl to determine what items a player has access too. In general each zone is +5-10 ilevel over the previous zones (with final bosses being a bit higher than that still). For instance, if you select ilvl 120, you'd be selecting all gear from 5-mans, heroics, and kara (minus the prince and tier 4). There could of course be tweaks/exceptions (the prince's drops and tier 4 helms/gloves could be considered ilvl 120 for instence). What do you think?

 

• Aldriana

That might be a reasonable way of filtering, but I personally like the idea of having someway of saying "these are your upgrades, and this is where to get them". Knowing that something drops somewhere in Karazhan is useful; knowing that it drops off Netherspite is even better. I was thinking of adding two columns to the armor page, one for which general instance things drop in (i.e. SSC, TK, Kara, etc.) which could be used for filtering purposes, and then a second column that would list specific drop information ("Heroic Black Morass", "Netherspite", "Lady Vashj", etc.) So you could, for instance, search for all upgrades in Karazhan, and it would return a list saying:

Spiteblade is a 50 AP upgrade from Netherspite
Netherblade Helm is a 30 AP upgrade from Prince
etc.

(Broken down across cells, of course, but you get the idea). Does this seem like something reasonable to implement?

 

Cragzop

Some quick impressions and comments.
Overall I like the idea a lot. Especially once more specific filtering is enabled. It's tedious at best to sort through all the gear choices looking for upgrades. This tool has the potential to make it far less painful

However it needs either a simple set of instructions so users can enter gear easily or much more gear. Using me as an example - We're still in Kara, I understand that Black Temple gear is better than what I have now, but if I can't even enter the gear I have now its usefulness seems limited.

What is the number that it's actually calculating? I noticed that if I leave most of the talents alone at their default values but reduce the poison one to zero I get a value of say 3793. If I then put the poison value back to 3 the value decreases to 3790. I admit the change is very slight, I just don't know why it would go down by making poisons more effective.

 

• Aldriana

The scores are admittedly sort of abstract right now - they're based on AP-equivalent points. So removing the point in poisons drops your dps by about 3AP-equivalent.

As for adding items - it's admittedly a bit challenging right now. If there's specific things you feel should be added, feel free to list them and I'll add them; however, please recall that the goal of this sheet is not to include every item that any rogue might have; the itemization starts with level 70 5-man blues, with the idea that these should be obtainable by any remotely serious raider without undo effort. So if there are specific level 70 blue (or easier to obtain) items you'd like added, I can add them - but I'm not going to put in just anything.

 

Cragzop

Thanks for the quick response.

I'd have understood it if removing the points had dropped the score.
It was adding them that seemed to lower it.

Items I was referring to: Inciters cloak, Natasha's neck, Kings Blade, Mag Scout Tunic, the wraith cannon, etc - I forget their exact names. Until the sheet style is finalized it might just be easier to explain how people can add their own. Also alphabetizing the lists might help some.

 

Darlal

Adding your own items is not as easy as it sounds. I will try to come up with a writeup of how to do it, but you have to be very careful of named ranges. Also, given the limitations of excel (and my excel coding knowledge I'm sure), there may be places where items have to be entered muliple times--gems, for instance, need to be entered under each color they apply to.

In terms of RAP (relative AP) increasing after you removed poison talents, this could be because some of your other stats are slightly more valuable wihtout poison talents.

For instance, if you have full 5/5 imp 5/5 vile in a mutilate build, a faster offhand is superior in dps by a large margin. If you have 0/5 imp 0/5 vile, a slower offhand will net you similar, or slightly better dps. Thus given your weapons, removing poinson talents could actually increase the value of some of your items (sorry for using mutilate, it's what came to mind).

 

Fluker

What do you mean by "hit cap" and "profit"?

 

• Aldriana

Dragonspine Trophy is such a powerful proc that it greatly increases the value of hit - since misses don't proc. Hence, you want to miss as little as possible. In practice, it elevates the value of hit to the point where optimal gear generally allows you to get to the point where you never miss (once your miss rate is 0%, further +hit is useless). The amount of hit required for this to happen is known as the "hit cap".

The currently accepted value for this (and the one used in this spreadsheet) is that this requires 25.5% total hit; for a rogue with combat potency and weapon expertise but no other sources of +skill, this requires 307.5 hit rating. Depending on your stats, the exact number will be a bit different.

"Profit" is an allusion to underwear gnomes, representing the fact that performing the first two steps leads to an advantageous outcome.

In reality, the situation is a bit more complicated; not all items with +hit are created equal, and once you get to the hit cap their is still a reasonable amount of flexibility in itemization; as a general rule you want to be stacking AP and Agi more than crit past this point, but to get the precise rubrics, a gander at this or the other rogue spreadsheet is recommended.

 

Ghond

With all these haste procs and +hit, do you ever hit a threshhold where using IP on your OH is better than DP? If so what would you guestimate it is?

Also if this is a threshhold that is passable with current gear availalbe it would be nice if your spreadsheet would be capable to indicate when you do pass that point.

 

• Aldriana

Eventually, this might be possible. It's pretty darn hard to do, though.

First, the daggers case: the fastest possible OHs are 1.3 speed, and you have no poison talents to assign. DP thus ticks off for 225 (and basically never goes down) every 3 seconds, or 75 dps.

IP hits for 170, so you need to average 75/170 = .44 procs per second to do as much damage. Since it procs 20% of the time, this means you need to hit 2.2 times per second. Assuming you never miss or are dodged (which isn't possible with current itemization), it would require a 2.87x increase in attack speed over the base speed of the weapon. You get a 1.3x increase from SnD, 1.025x from pro-rated Blade Flurry, and ~1.02x from average mongoose uptime on both hands. This means that you need to get roughly a 2.11x increase in weapon speed from haste rating, which requires 1167 haste rating to obtain. Even if you somehow managed 100% uptime on both TSD and Dragonspine, that's only 565, and there's just not enough haste rating on other items to get another 600, nor do I expect there to be anytime soon.

Now, sword spec, optimally speaking, gets 4 points to put in poison talents, so the relevant comparison is DP with 4/5 Vile vs IP with 4/5 Imp. Deadly Poison with 4/5 Vile ticks off for 261 every 3 seconds, which is 87 dps, which requires .512 procs per second; with a 28% proc rate and assuming attacks never miss or are dodged, this requires 1.83 attacks per second. With the fastest possible sword OH (1.4 speed), this would require 2.56x attack speed increase, which works out to about.... 928 haste rating. Again, that's still well out of reach of current itemization.

So, briefly: From a theoretical perspective, it's possible, but in practice, it ain't gonna happen.