Rawr Retribution Model

Endoscient · 04/02/09, 10:29 PM

Rawr Retribution Module

Note the OP is out of date, I would recommend reading the whole thread to get up to date information.

Latest Version: 2.2.10 - Download from Codeplex

Changes

3.2 mode is synced with latest PTR build.
Fixed partial resist calculations. It was slightly underestimating the effect of partial resists, so you will see dps go down a little bit.

Introduction

For those who don't know what Rawr is, it is a theorycrafting program used for various various classes/specs to help them maximize there effectiveness. You load in your gear/enchants and it computes how much dps you do, and how much of an upgrade items are for you. You can learn more about it and download it at Rawr site on Codeplex.

This thread is about the Retribution Paladin model for it, please use the Rawr model development thread for anything not specific to the Ret module.

Some of you might have used it some post 3.0, and found it to not be reliable. The previous creator of Retribution module rerolled to DK at the time, and got much busier at work, so didn't have time to keep it up to date. So recently I took it over and essentially rewrote the whole thing, so please give the new version a shot even if you had problems with the older one.

How to use Rawr

Load your character in from Armory.
- File -> Load from Armory
Make sure your talent selections are correct.
- Go to the talents tab, and make sure it is accurate.
- This is needed until Armory updates to 3.1 trees, since there have been large changes from the live ones.
Set all relevant buffs. If you do not do this, the results will be much different then expected.
- Go to the buffs tab.
- Select all buffs that you will have in the raid.
Modify calculation options.
- Go to the options tab.
- Change any desired setting to model any given fight. The default settings should be good enough for novice users.
Have fun looking for upgrades you need!

Methodology

Rawr computes how much damage each of your strikes will do on average. It then uses the defined rotation to compute how much dps you will do. There are two ways for it to model your rotation.

FCFS Simulator - You enter in a priority scheme, and it simulates a fight following that scheme. This can more accurately the benefits of effects that change your rotation like 4T7 or Glyph of Consecration.
Effective Cooldowns - You enter the effective cooldown of all of your abilities. This doesn't compute the value of effects like 4T7 at all, but it gives you more fine tuned control.

Most trinkets have there procs averaged out to what they provide, its not totally accurate but close enough.

Upcoming Feature List

Improved information in the calculations pane.
Ability to define multiple FCFS priorities and have the model test all of them and use the highest dps one.

Known Issues
(none)

Frequently Asked Questions

Why do the character stats in Rawr don't match my character sheet?
Rawr averages to proc rates of some items (normally trinkets like [Mirror of Truth]) and counts that towards your stats.

Why did I get this <inset strange result>?
This is most likely caused by you having some error in your character settings, most likely not having all of your buffs or talents set correctly. Please make sure those are correct, if you are still getting weird results see the next section.

Something is wrong! HELP!

If you think something is wrong, either an incorrect calculation or whatever, please post here with the problem. But remember to ALWAYS post your Rawr XML character file when you do. It is significantly easier for me to diagnose the issue if you do. How do you post the XML character file? You can either upload it to a place like Filefront, or you can copy & paste the text into a PM to me.

Cavemanz · 04/03/09, 6:41 AM

Wrathstone is calculated at 856 AP constantly, and thus best in slot by far.
I guess at 2min cd 20sec duration, it should be calculated with about 16.6% uptime, giving around 142 ap average.
Changing that manually puts it at 3rd after greatness and fury. Tho assuming 320 ap for fury like rawr is doing is probably overly optimistic for the average fight too.

Also, the strength updates on ulduar gear havent made it into the simulator. Its using old values with more stam, less strength.

• Zurm · 04/03/09, 9:34 AM

Originally Posted by Cavemanz

Also, the strength updates on ulduar gear havent made it into the simulator. Its using old values with more stam, less strength.

You can run the update tool, assuming wowhead has updated its values, that should correct those items.

Endoscient · 04/03/09, 11:09 AM

Originally Posted by Cavemanz

Wrathstone is calculated at 856 AP constantly, and thus best in slot by far.
I guess at 2min cd 20sec duration, it should be calculated with about 16.6% uptime, giving around 142 ap average.
Changing that manually puts it at 3rd after greatness and fury. Tho assuming 320 ap for fury like rawr is doing is probably overly optimistic for the average fight too.

Also, the strength updates on ulduar gear havent made it into the simulator. Its using old values with more stam, less strength.

The item text is broken, it says "Equip: Increases attack power by 856 for 20 sec." which is confusing the item parser to think it is a flat amount of ap, and not a proc. You are going to need to manually edit it, until the item text is fixed.

[Fury of the Five Flights] is something I am going to try to get an accurate model for soon.

Rawr takes all of its stats from Wowhead and they haven't updated the items to there new strength values yet.

Endoscient · 04/03/09, 12:20 PM

One thing I forgot to mention. Make sure when you load your character from Armory you check out the talent pane and make sure everything has the proper 3.1 ranks. It can be messed up at first, because of the large amount of talent changes from live.

Sideshow · 04/03/09, 2:09 PM

Looks to be much improved and up to date. I used to use it a lot at the end of TBC, I'll be keeping it updated again now - thanks!

One thing, is there no way to change enchant data in UI? I seem to remember there was before..

Hyperspeed Accelerators are way off in this version - the hover tip appears correct (56 haste rating), but it's ranked like half of what Minor Haste is (10 haste rating)

Also, I'll play around in depth this weekend with it.. but I noticed that Bellator's sheet (which I also like) shows in 3.1 it's better for me to drop T7.25 gauntlets (and break set) for Zeliek's by about ~50dps. This sort of made sense to me since judgement damage is going down as a percentage, and expertise is going up, and a new priority system causes effective CD of judgement to be over 8 anyway. (using CS, HoR, J, DS, Con, Exo)

But it shows it's almost bad to break T7 bonus for every slot, even though effective CD of judgement is coming up at 8.11 - does that make sense? I wouldn't even be using the bonus effect then... and stat wise it seems liek Zeliek is better (I'm only at 11 expertise)

thoughts? Is T7 bonus overweighted somehow since it seems like new priority puts it's CD over 8 anyway

Exemplar · 04/03/09, 2:32 PM

Trick to remember with effective cooldown is that it's a weighted average. If 50% of the time a cooldown is 7, 25% of the time it's 8, and 25% it's 9, then the effective is (7 * .25) + (8 * .5) + (9 * .25) = 1.75 + 4 + 2.25 = 8. If you could only use it every 8 seconds, then it wouldn't function - you couldn't use it 25% of the time you needed to use it, which would throw everything else off and generate new effective cooldowns.

Raencloud · 04/03/09, 2:34 PM

In my paladins current gear, I'm showing 8.08 effective cd on Judgement, but it says the 4p bonus is worth 71.62 dps. My conclusion is that the priority list must be wildly fluctuating the Judgement usage between close to 7 seconds apart to up to 10 seconds and it happens to average out to 8.08.

Djay · 04/03/09, 2:59 PM

I'm a little bit confused looking at this version. The module put out right after the SoB changes was showing that it was worth dropping 4 piece T7 for stronger offset pieces and now that is no longer the case?

Sideshow · 04/03/09, 2:59 PM

Okay that makes sense, I wonder why I get a huge difference though regarding upgrades between spreadsheet/rawr then. In the spreadsheet almost every slot says break T7 4 piece for a new item (slightly in some cases, but still an increase). But in RAWR every slot says keep T7 4 piece. I know the spreadsheet has the updated STR/STAM on the new gear, I'll see if that's all it takes to push it over in RAWR.

RAWR also said to drop betrayer for jawbone for a ~30 dps increase (spreadsheet did not). I know expertise is worth a little more in 3.1, and I'm a dwarf, and I believe weapon speed weights changed to favor slower even more right? But still shocking to see Jawbone up on top of Betrayer. Going to look into that some more before I shell 1000g for berserker on jawbone going into Ulduar..

Last comment, one thing I love about exemplar/bellators sheet is it's so easy for me to go twink trinket modeling or libram modeling to what I think they should be. Like I'd have no idea how to put in the proc for Bandit's Insignia in RAWR, or to update the libram changes in 3.1. It seems like RAWR was upgraded to include a lot of actual proc effects from wowhead that I guess you can tap into in your c# model right?

Excluding wrathstone it says top trinkets for me are
DMC (Str) - 306
DMC (Agil) - 237
Mark of Norg - 211
FoFF - 210

2 DMC won't work because the proc CD is linked right? I think I'll pick up the next Mark since nobody else wants them just in case it does turn out to be that strong for non expertise capped set

well great work endo & zurm, we got 3 solid tools now I can use

edit: It also ranks the Quick Autumn's Glow gem (16 haste) at 79 dps and the bold scarlet ruby gem (16 str) at 26 dps, but as far relative stat values it seems like haste is where it should be (below agil and above arp) - that might explain why the mark of norg is so high

• Zurm · 04/03/09, 3:09 PM

Originally Posted by Sideshow

Last comment, one thing I love about exemplar/bellators sheet is it's so easy for me to go twink trinket modeling or libram modeling to what I think they should be. Like I'd have no idea how to put in the proc for Bandit's Insignia in RAWR, or to update the libram changes in 3.1. It seems like RAWR was upgraded to include a lot of actual proc effects from wowhead that I guess you can tap into in your c# model right?

That's just a function of application knowledge/comfort level. Items in Rawr are fully customizable. The catch is, Rawr handles procs by averaging their effects (it's a theory tool, not a simulator). As such, a proc of 1000 ap that's up 1/10th of the time is valued at 100ap (and this is also why your Rawr stats do not equal your paper doll stats). So if you want to modify a trinket, simply adjust its ap value to the averaged amount, rather than trying look for a "proc" value.

As a side note, don't give me too much credit. Although I used to be the developer back in BC, now I'm little more than a glorified tester (we are and have been in the same guilds for a while, so its easy for him to harass me into testing). I still have access to make changes should I need to, but really this is all Endoscient's doing. Endoscient has been responsible for the complete overhaul and improvements. I am the former developer referred to here:

Originally Posted by Endoscient

Some of you might have used it some post 3.0, and found it to not be reliable. The previous creator of Retribution module rerolled to DK at the time, and got much busier at work, so didn't have time to keep it up to date. So recently I took it over and essentially rewrote the whole thing, so please give the new version a shot even if you had problems with the older one.

Endoscient · 04/03/09, 3:23 PM

Originally Posted by Sideshow

Hyperspeed Accelerators are way off in this version - the hover tip appears correct (56 haste rating), but it's ranked like half of what Minor Haste is (10 haste rating)

I recently tried adding the effect of haste lowering spell gcd of Exo/Cons into the Rotation Simulator. Overall I don't like how it works out, because of how it creates many weird edge cases which are the issues everyone is seeing now. I am going to revert it for the next build.

Endoscient · 04/03/09, 4:35 PM

I made a bunch of changes since 2.2b6 last night, so I am going to release a new version of Rawr.Retribution. You can download it off of Filefront. To get this to work, download and install Rawr 2.2.0b6 like normal, then download the new Retribution dll, and replace the old one in the Data folder with it.

Changes

All specials are now correctly modeled using a 2-roll system. It will result in a large increase to hit/expertise values, this is not a bug.
Rotation Simulator bug fixes and improvements.
- Removed spell gcd (uses 1.5 for spells now) because it was creating many weird edge cases with haste.
- Added a new option called "Wait", amount of time the simulator will wait for an ability of higher priority.
Fix for base crit chance.

chinoquezada · 04/05/09, 12:24 PM

* moved to correct forum thread.

Endoscient · 04/05/09, 12:41 PM

Originally Posted by Sideshow

Okay that makes sense, I wonder why I get a huge difference though regarding upgrades between spreadsheet/rawr then. In the spreadsheet almost every slot says break T7 4 piece for a new item (slightly in some cases, but still an increase). But in RAWR every slot says keep T7 4 piece. I know the spreadsheet has the updated STR/STAM on the new gear, I'll see if that's all it takes to push it over in RAWR.

RAWR also said to drop betrayer for jawbone for a ~30 dps increase (spreadsheet did not). I know expertise is worth a little more in 3.1, and I'm a dwarf, and I believe weapon speed weights changed to favor slower even more right? But still shocking to see Jawbone up on top of Betrayer. Going to look into that some more before I shell 1000g for berserker on jawbone going into Ulduar..

Last comment, one thing I love about exemplar/bellators sheet is it's so easy for me to go twink trinket modeling or libram modeling to what I think they should be. Like I'd have no idea how to put in the proc for Bandit's Insignia in RAWR, or to update the libram changes in 3.1. It seems like RAWR was upgraded to include a lot of actual proc effects from wowhead that I guess you can tap into in your c# model right?

I think the two spreadsheets are still using a 1-roll system for special attacks. Switching over to 2-roll drastically increases the value of hit/expertise, and as a Dwarf Jawbone gives you -3% chance to dodge. For me as a Dwarf it shows Jawbone 10 dps under Betrayer. Which is why I said in the first post, if you ever think something is wrong, you need to post your character file for me to take a look at it.

How Rawr rates items, it shows the value of changing that one slot. So if you are just going slot by slot to see if there is an upgrade worth breaking 4T7 over, it probably won't show you one since 1 slot won't be able to make up the difference. You need to check the total dps of not using any T7 vs the dps of using 4pc.

The only librams that were changed were [Deadly Gladiator's Libram of Fortitude] was changed to 10 seconds, and [Venture Co. Libram of Retribution] changed to 81 Divine Storm damage, right? The new duration of Glad Libram is properly modeled in the latest version. The Venture Co. Libram, while it isn't currently parsed from Wowhead/Armory correctly, you can manually edit it to add in the Divine Storm Damage stat (under Equipment Procs).

I also changed the Greatness Proc to a non stacking stat, so when the next version gets released if you have one in one slot, the other slot will value the proc as 0.

chinoquezada, that is not an error specific to the Retribution model. So please report it in the main Rawr thread, Rawr Model Development.