slagg

It is currently July 21st, 2017, 3:27 am

All times are UTC - 8 hours [ DST ]


Post new topic Reply to topic  Page 1 of 3
 [ 59 posts ]  1, 2, 3  

Mew.Bear


Author Message
 Post subject: Mew.Bear
PostPosted: June 14th, 2011, 5:46 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Mew as of recent 4.2 snapshots now includes fairly comprehensive support for modeling Feral Tanking.

Reviews from the press:
"i'd never use it because bear is basically put my fingers over 1-5 and press all at once but it seems to be very good" - Floofles (Outland EU)

Tangedyn Formulation:
Image

This is a straight forward copy of Tangedyn's Mitigation Spreadsheet, adapted to work with the Mew UI. It attempts to give a reasonable estimate regarding survivability, and most people should be familiar with it by now.

Notable differences between Tangedyn's Spreadsheet and the Mew implementation:
  • Mew supports trinkets.

Simulator (Bear):
Image

This is a Simulator. Currently it is oriented towards TPS/DPS calculations, though it can be used to look at various other things if you are willing to sift through the output (In particular, Savage Defense/Vengeance/Survivability).

Things modeled:
  • Vengeance
  • Incoming damage (White Swing, Yellow Swing, and Magic)
  • Savage Defense
  • Rage (Optional, Unlimited Rage is also supported)
  • Cooldowns (Disabled by default, requires editing script)
  • Interrupt Duty

Things not modeled:
  • AoE
  • Tanking Trinkets
  • Magic Resistance

Notes:
  • Both models are incredibly sensitive to the amount of incoming damage. In particular, if the values specified are too low, the Simulator will not build realistic amounts of Vengeance, and your projected threat will suck (The screenshot illustrates this with 100k damage every 2 sec, Vengeance is comming out to 67% of the cap.).
  • The Simulator has quite a large amount of variance. The 95% confidence interval for TPS is ~4x that of the Cat model. When generating stat comparisons, use sufficient iterations, and pay attention to the error bounds.

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 6:30 pm 
Offline
Cub

Joined: June 10th, 2011, 10:24 am
Posts: 6
Having followed the development of Mew.bear from the beginning, I can fully endorse this product. It can tell you which of your 1-5 rotational faceroll buttons is contributing what to your TPS as well as how much mitigation and avoidance you'll have with any particular set of gear. Of course this is rather pointless because the only thing you should care about is the DPS numbers, but luckily Mew.bear supports this too!

Whether you're failbear or dumbbear, Mew has something for you!


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 7:31 pm 
Offline
Adult

Joined: June 14th, 2011, 4:40 pm
Posts: 72
What program do I run mew as? It gave me a list, not sure what select... I'm such a computer noob =\

_________________
http://us.battle.net/wow/en/character/lothar/joux/advanced


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 7:39 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Joux wrote:
What program do I run mew as? It gave me a list, not sure what select... I'm such a computer noob =\


Do you have Java installed on your system? If you do, Mew.jar should be setup to just open as an application. If not, go to http://www.oracle.com/technetwork/java/ ... index.html and download the JRE.

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 7:42 pm 
Offline
Adult

Joined: June 14th, 2011, 4:40 pm
Posts: 72
Alright, downloading now. Thanks!

I've got it set to open with java, but when I open it in a black box comes up and then closes. It's says Exception in thread java.lang and that's all I can get from it.

_________________
http://us.battle.net/wow/en/character/lothar/joux/advanced


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 7:54 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Joux wrote:
Alright, downloading now. Thanks!

I've got it set to open with java, but when I open it in a black box comes up and then closes. It's says Exception in thread java.lang and that's all I can get from it.


Ugh. That shouldn't happen since I defined Main-Class properly (And it's worked for my usual group of beta testers).

Just to make sure, you're using Mew-20110613.zip off my google code page, and you extracted the archive right?

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 7:57 pm 
Offline
Adult

Joined: June 14th, 2011, 4:40 pm
Posts: 72
Yep, It's now Mew, that's a JAR file. I probably screwed it up because I thought it was supposed to open with Internet Explorer, so I set it to open as that... I can change it, but I have no idea what to change it to.

I tried going into the java that I downloaded, but none of the applications there opened it either.

_________________
http://us.battle.net/wow/en/character/lothar/joux/advanced


Last edited by Joux on June 14th, 2011, 8:00 pm, edited 1 time in total.

Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 8:00 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Joux wrote:
Yep, It's now Mew, that's a JAR file. I probably screwed it up because I thought it was supposed to open with Internet Explorer, so I set it to open as that... I can change it, but I have no idea what to change it to.


Try this: http://johann.loefflmann.net/en/softwar ... index.html

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 8:04 pm 
Offline
Adult

Joined: June 14th, 2011, 4:40 pm
Posts: 72
Awesome, that did it. Thanks very much Yawning!

(P.S. One last question: Do I need to keep Jarfix on my computer to be able to double click jar files?)

_________________
http://us.battle.net/wow/en/character/lothar/joux/advanced


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 14th, 2011, 8:12 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Joux wrote:
Awesome, that did it. Thanks very much Yawning!

(P.S. One last question: Do I need to keep Jarfix on my computer to be able to double click jar files?)


Nope, all the tiny app does is re-associate .jar and .jnlp files with the JRE, you can get rid of it now that everything is setup properly.

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 15th, 2011, 11:43 pm 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
Hi Been trying this out and comparing the Tandgeyn method (Mew) with the spreadsheet.

I haven't been able to get the boss fail chance to line up and I am not sure why. Spreadsheet assumes food and 549 agi buff, + 5% stats. So I have selected these options in Mew.

I have not selected any flask or trinkets, and put the same agi and dodge ratings into both spreadsheet and Mew which I run in the Tangedyn Formulation.

I get for the example I am trying
Agi 3815 dodge 1166

I am using Night Elf in both.

Boss Fail Chance.
Spreadsheet 0.43701
Mew 0.4155985323127218

Which is a considerable difference.
Is this a bug, Or am I not using something correctly?


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 16th, 2011, 12:18 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Edit: Mew-20110619 has all of the issues mentioned here corrected.

Lev wrote:
Is this a bug, Or am I not using something correctly?


Yes. :lol:

Causes for the discrepancy with the publicly available Mew build:
  • There was a minor bug in my player dodge calculation (< 1% difference, fixed in r695).
  • There was a bug in how I was calculating avoidance penalty due to the boss being a higher level (~1.2% difference from reality, skimming his sheet he ignores the penalty all together, fixed in r698).
  • His spreadsheet assumes food (Cell B35 "=1.05*(B9 + 549 + 90)").

Comparing SVN Mew and his sheet, the two values match once you take into account the fact that Mew is penalizing avoidance by 0.012 (0.006 Miss, 0.006 Dodge) while he isn't.

As a side note, from what I can tell his sheet is also still modeling 4.1 while Mew models 4.2 so there's going to be other discrepancies as well. Things should be more consistent at 4.2 release, and if you're feeling adventurous you could always build a copy out of SVN (since I also fixed a minor bug in the threat model).

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 20th, 2011, 5:45 pm 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
So, I have been playing with the new version (Mew-20110619 ) and have been trying to tie down the armor numbers. The "Armor" Field in this is your total armor (including yellow bonus armor) and inclusive of the meta gem bonus. What then is the bonus armor. If I put in the raw bonus (490=240+250) or some meta effected number it 490*1.02?.

This is all a fairly small difference I am getting but slightly puzzling things are going on, eg for the same armor input numbers agile provides slightly more mitigation than austere. I would have thought it would be equal, or differing by perhaps the meta's adjustment to the bonus armor. The difference however seems smaller than 2% of bonus armor.


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 20th, 2011, 6:56 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Lev wrote:
So, I have been playing with the new version (Mew-20110619 ) and have been trying to tie down the armor numbers. The "Armor" Field in this is your total armor (including yellow bonus armor) and inclusive of the meta gem bonus. What then is the bonus armor. If I put in the raw bonus (490=240+250) or some meta effected number it 490*1.02?.

This is all a fairly small difference I am getting but slightly puzzling things are going on, eg for the same armor input numbers agile provides slightly more mitigation than austere. I would have thought it would be equal, or differing by perhaps the meta's adjustment to the bonus armor. The difference however seems smaller than 2% of bonus armor.


Bonus armor is raw bonus armor, so 490 usually, though if you had something that wasn't affected by Thick Hide (trinkets?) that also would go there. The difference is small because stats have a extremely tiny impact on the mitigation value (which is why when I calculate relative weights the output is normalized around Agi). I could display raw mitigation gain per point instead but unless people feel like looking at things in scientific notation, it would be somewhat hard to read.

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 21st, 2011, 6:40 am 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
Hi thanks for responding. I may not have been clear but wasn't talking about the RSV, but the armour dmg reduction.

I have a set up with 11937 Armour and Zero bonus, no armor buff.

With an Austere I get
Armor Dmg Reduction: 0.5893223567657447
With Agile I get
Armor Dmg Reduction: 0.5893418213154497
Agile matches Tangedyn's spreadsheet

I know its small and hardly worth worrying about, but I would have thought they would be the same.


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 21st, 2011, 3:22 pm 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
It's a rounding error introduced by the fact that I need to derive the pre-armor meta armor value. It's not something really worth fixing because if someone were to actually want to compare metagems, the would need to adjust the amount of armor passed into the sim in the first place (since character sheet armor will change).

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 21st, 2011, 4:08 pm 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
Ok thanks again for replying. If its just rounding then I agree it is too small to bother. It might be my background in validating derivatives pricing systems that makes me focus on small differences.


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 23rd, 2011, 8:27 pm 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
Just another question.

I take it that the SD uptime that Mew quotes in the simulation is directly comparible to the SD uptime that Tangdyen's spreadsheat calculates. ie it calculates the % of time when SD is up when the boss hits you not raw time SD is up (which is considerably lower)?


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 24th, 2011, 4:23 am 
Offline
Math Robot
User avatar

Joined: June 10th, 2011, 9:55 am
Posts: 352
Lev wrote:
Just another question.

I take it that the SD uptime that Mew quotes in the simulation is directly comparible to the SD uptime that Tangdyen's spreadsheat calculates. ie it calculates the % of time when SD is up when the boss hits you not raw time SD is up (which is considerably lower)?


Yep. The raw time SD is up is relatively meaningless, so I calculate it in the same way that he does. It's different from what his sheet predicts primarily because Pulverize uptime is not realistically anything near 100% with the full 9% crit benefit (It's lower than a few things on the priority list so you have global collisions on refresh, Berserk comes into play, and unless that component is disabled there's rage starvation to worry about).

Sort of glad someone apart from me is cross checking the models, since I want the tool to be accurate and useful.

_________________
"There's space for a paper airplane race in the eye of a hurricane."

Mew: http://code.google.com/p/mew-wow-druid-model/ (RIP)


Top
 Profile  
 
 Post subject: Re: Mew.Bear
PostPosted: June 27th, 2011, 4:38 pm 
Offline
Juvenile

Joined: June 15th, 2011, 9:27 pm
Posts: 34
Quote:
Sort of glad someone apart from me is cross checking the models, since I want the tool to be accurate and useful.


good, I hoped I wasn't annoying you with nitpicking questions. I'm not being completely thorough but just trying to make sure it makes sense and I understand any differences.


Top
 Profile  
 
Display posts from previous:  Sort by  

Post new topic Reply to topic  Page 1 of 3
 [ 59 posts ]  1, 2, 3  

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron