Register

Mew Releases

Face-rippin fun.
Revered
Posts: 320
Joined: Tue Feb 08, 2011 5:51 pm

Re: Mew Releases

Postby ShmooDude » Thu Nov 03, 2011 7:11 pm

Konungr wrote:Well, its going to be hard to calculate it out, since the ICD will continue resetting each time a new stack, until 10, is applied. Since it is a 15% chance on attack to trigger, you have the ramp-up time, then as soon as the 10th Stack hits, 1070 Agi, it starts. As soon as it ends, 30 seconds later you can start gaining procs again. This is going to be one beast of a weapon. It's also going to play HELL on our refreshing of Rip prior to BitW phase.


Are you sure we're on the same page here? Based on the tooltip:
Equip: Your melee and ranged attacks have a chance to trigger Fury of the Beast, granting 107 Agility and 10% increased size every 1 sec. This effect stacks a maximum of 10 times and lasts 20 sec.

15% chance to proc. When it procs...
0-1 seconds 107 Agility & 10% size increase
1-2 seconds 214 Agility % 20% size increase
2-3 seconds 321 Agility & 30% size increase
...
8-9 seconds 963 Agility & 90% size increase
10-20 seconds 1070 Agility & 100% size increase

Clicking on the proc itself on wowhead lends to this interpretation.

Exalted
User avatar
Posts: 568
Joined: Sat May 28, 2011 3:43 pm

Re: Mew Releases

Postby Konungr » Thu Nov 03, 2011 8:11 pm

If that's the way it works, its going to be rubbish. No other Proc that I know of with stacks like this works like that.

Honored
Posts: 76
Joined: Sat Jun 04, 2011 4:21 am

Re: Mew Releases

Postby Yawning » Thu Nov 03, 2011 8:40 pm

Konungr wrote:If that's the way it works, its going to be rubbish. No other Proc that I know of with stacks like this works like that.


http://www.youtube.com/watch?v=FopyRHHlt3M

The mechanic has been done before. http://www.wowhead.com/item=50348 is a recent example.

The proc sims well (I think it's somewhat overpowered to begin with) but presumably the intent is to make these the last weapons you want in the expac so it's probably not a huge concern.

Also, the project's SVN repository, issue tracker and change logs are public for a reason. If there's functionality that's desperately needed that's not in SVN, either file it as a feature request in the issue tracker (that I will probably ignore), or implement it and send me a diff (that I will review and apply if it's actually any good).
"Understanding is a three-edged sword."

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

Posts: 4
Joined: Wed Dec 14, 2011 8:46 am

Re: Mew Releases

Postby Dohtem » Wed Dec 14, 2011 8:54 am

Is there a problem with Mew lately? It won't calculate relative stat values for me? (Used import function on an EU-char, also tried manually entering numbers)

Exalted
User avatar
Posts: 568
Joined: Sat May 28, 2011 3:43 pm

Re: Mew Releases

Postby Konungr » Wed Dec 14, 2011 1:11 pm

Dohtem wrote:Is there a problem with Mew lately? It won't calculate relative stat values for me? (Used import function on an EU-char, also tried manually entering numbers)


/facepalm

You can try reading around this thread or these forums for the answer, because it's been mentioned many times over, or just not use them, since you probably don't understand how to use them properly. In fact, based on your post in the Logs thread about WoW-Reforge, I can almost guarantee you are using them wrong.

Posts: 4
Joined: Wed Dec 14, 2011 8:46 am

Re: Mew Releases

Postby Dohtem » Wed Dec 14, 2011 1:41 pm

Dohtem wrote:Is there a problem with Mew lately? It won't calculate relative stat values for me? (Used import function on an EU-char, also tried manually entering numbers)


I got this q. answered in another thread. Sorry, but that tick really wasnt easy to find

Posts: 11
Joined: Sat Jun 04, 2011 2:23 am

Re: Mew Releases

Postby Odas » Thu Dec 22, 2011 2:36 pm

Is there a reason mew seems to be getting 80%+ critical chance for ravage?

Ravage:
DPS: 2364.6411
Nr. Ravages: 11.3343
Miss/Dodge/Crit/Hit: 1.37917%/0%/90.91246%/7.70837%
Average damage: 62817.63604


I got that from running the Mihir410 sample. I had similar results when using my own gear.

Honored
Posts: 76
Joined: Sat Jun 04, 2011 4:21 am

Re: Mew Releases

Postby Yawning » Thu Dec 22, 2011 2:42 pm

Odas wrote:Is there a reason mew seems to be getting 80%+ critical chance for ravage?

Ravage:
DPS: 2364.6411
Nr. Ravages: 11.3343
Miss/Dodge/Crit/Hit: 1.37917%/0%/90.91246%/7.70837%
Average damage: 62817.63604


I got that from running the Mihir410 sample. I had similar results when using my own gear.


Yeah it's a bug that's fixed in SVN (I just did it right now).

For future reference I only occasionally read these forums. I do pay attention to the project's bug tracker though.

Edit: It's fixed in Mew-20111222, go download it.
"Understanding is a three-edged sword."

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

Posts: 1
Joined: Sat Jan 07, 2012 7:59 pm

Re: Mew Releases

Postby Finwe » Sat Jan 07, 2012 8:02 pm

I'm not having any luck on finding the setup info to get the standalone version running on my computer. I just upgraded to windows 7. It worked fine on my old system. The .jar was associated with WinRAR on my new computer. I tried associating it with jave.exe, but that didn't get it to run either. Can you point me to a setup page or something like that?

Honored
Posts: 76
Joined: Sat Jun 04, 2011 4:21 am

Re: Mew Releases

Postby Yawning » Mon Jan 09, 2012 3:19 am

Finwe wrote:I'm not having any luck on finding the setup info to get the standalone version running on my computer. I just upgraded to windows 7. It worked fine on my old system. The .jar was associated with WinRAR on my new computer. I tried associating it with jave.exe, but that didn't get it to run either. Can you point me to a setup page or something like that?


...

File associations being screwed up isn't a Mew problem, but this comes up often enough.

http://www.softpedia.com/get/Others/Mis ... rfix.shtml

Download and run that and it will fix things.
"Understanding is a three-edged sword."

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

Exalted
User avatar
Posts: 568
Joined: Sat May 28, 2011 3:43 pm

Re: Mew Releases

Postby Konungr » Thu Feb 09, 2012 8:00 pm

Notdiggz wrote:So i was doing some sim work today and i am finding it hard to explain.

DPS (non-BitW): 51336.72936
DPS (BitW): 49359.01329

Am i just reading this backwards or is it saying my none execute dps is higher than execute dps? Now i assumed moving the Blood in the Water slider closer to 60% from 25% would change these results but at 40% i see even more of a change in the wrong direction.

DPS (non-BitW): 65487.49785
DPS (BitW): 50582.98977

If someone could explain whats going on would be appreciated.


I also hopped into Mew and ran a sim and got similar results.

Exalted
User avatar
Posts: 1563
Joined: Mon May 24, 2010 10:21 pm

Re: Mew Releases

Postby Tinderhoof » Thu Feb 09, 2012 8:25 pm

This is expected and correct.

Prior to our T13 2p bonus for a 5 minute fight only the last 1:15 seconds were being recorded as BitW phase. In that short window we would see a Berserk, 2 TF's, likely at least some of BL, and a Potion use. In the very small window DPS goes much higher.

Now we are on the flip side. For a 5 minute fight out of BitW is 2 minutes. In that 2 minutes we have a Berserk, 4 TF's and a Pot. With the full minute shorter we will see DPS go up.

At first glance it does look kind of odd, but it does make sense given the change.

Exalted
User avatar
Posts: 568
Joined: Sat May 28, 2011 3:43 pm

Re: Mew Releases

Postby Konungr » Thu Feb 09, 2012 9:23 pm

But it doesn't make sense that prior to BitW we are keeping up Rip, Rake, Mangle, SR, and using FB maybe once every 30-60 seconds if lucky, but during BitW we are keeping them up and using FB every 10-15 seconds. Why would DPS go down from being able to use a higher damaging ability more often without risk of losing uptime?

User avatar
Posts: 12
Joined: Sun Mar 06, 2011 2:29 pm

Re: Mew Releases

Postby feralCat » Thu Feb 09, 2012 10:53 pm

I have a question about Mew's Combat Log feature: Is it outputting a "real" format? I was hoping to toss the simulated log at an analyzer like world of logs to compare my reality with my simulation.

Honored
Posts: 76
Joined: Sat Jun 04, 2011 4:21 am

Re: Mew Releases

Postby Yawning » Mon Feb 13, 2012 5:04 pm

Konungr wrote:
Notdiggz wrote:So i was doing some sim work today and i am finding it hard to explain.

DPS (non-BitW): 51336.72936
DPS (BitW): 49359.01329

Am i just reading this backwards or is it saying my none execute dps is higher than execute dps? Now i assumed moving the Blood in the Water slider closer to 60% from 25% would change these results but at 40% i see even more of a change in the wrong direction.

DPS (non-BitW): 65487.49785
DPS (BitW): 50582.98977

If someone could explain whats going on would be appreciated.


I also hopped into Mew and ran a sim and got similar results.


Heroism is used on the pull.

feralCat wrote:I have a question about Mew's Combat Log feature: Is it outputting a "real" format? I was hoping to toss the simulated log at an analyzer like world of logs to compare my reality with my simulation.


Nope. I thought about doing so, but decided it wasn't worth it.
"Understanding is a three-edged sword."

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

Posts: 5
Joined: Fri Feb 24, 2012 9:04 am

Re: Mew Releases

Postby amilee » Fri Feb 24, 2012 9:10 am

I would like to test out some ideas for modification of the existing default cat strategy. The term "script" for the strategy files seems to imply the possibility of loading these on the fly rather than having to completely rebuild Mew. But I haven't found any way of activating my modified script file and haven't found any documentation on this. Can someone please shed some light on it :-)

Exalted
User avatar
Posts: 1145
Joined: Fri Jun 04, 2010 4:28 am

Re: Mew Releases

Postby Leafkiller » Fri Feb 24, 2012 7:17 pm

amilee wrote:I would like to test out some ideas for modification of the existing default cat strategy. The term "script" for the strategy files seems to imply the possibility of loading these on the fly rather than having to completely rebuild Mew. But I haven't found any way of activating my modified script file and haven't found any documentation on this. Can someone please shed some light on it :-)



"Custom Strategy: [Load]" on the "Model Parameters" tab.

Posts: 5
Joined: Fri Feb 24, 2012 9:04 am

Re: Mew Releases

Postby amilee » Fri Feb 24, 2012 8:05 pm

Leafkiller wrote:
amilee wrote:I would like to test out some ideas for modification of the existing default cat strategy. The term "script" for the strategy files seems to imply the possibility of loading these on the fly rather than having to completely rebuild Mew. But I haven't found any way of activating my modified script file and haven't found any documentation on this. Can someone please shed some light on it :-)


"Custom Strategy: [Load]" on the "Model Parameters" tab.


derp. don't know how I managed to overlook that button. thank you!

Posts: 5
Joined: Fri Feb 24, 2012 9:04 am

Re: Mew Releases

Postby amilee » Sat Feb 25, 2012 11:08 am

Working my way through the the default CatModelSimulator-script I stumbled upon this section:
Code: Select all
586:  if (!interruptSoon && isInBloodInTheWaterRange) {
[...]
593:       if (cps == 5 && isRipUp)
594:               return Action.BITE;
595:  }
The available energy isn't checked there. I would've assumed the condition should look something like this:
Code: Select all
if (cps == 5 && isRipUp && ((!isBerserkUp && energy >= 50) || (isBerserkUp && energy < 25))   
And yet it's absolutely 0 difference in dps when I modify it like that. Why is this so?

---

Apart from that, my ideas for strategy tweaking resulted in a ~150 dps gain so far (Mihir-Cat-410, default script 50768 dps, 50925 dps after). Is this relevant enough for you to be interested in the modifications I made? I'll post it then.

Exalted
User avatar
Posts: 1145
Joined: Fri Jun 04, 2010 4:28 am

Re: Mew Releases

Postby Leafkiller » Sat Feb 25, 2012 3:14 pm

amilee wrote:Working my way through the the default CatModelSimulator-script I stumbled upon this section:
Code: Select all
586:  if (!interruptSoon && isInBloodInTheWaterRange) {
[...]
593:       if (cps == 5 && isRipUp)
594:               return Action.BITE;
595:  }
The available energy isn't checked there. I would've assumed the condition should look something like this:
Code: Select all
if (cps == 5 && isRipUp && ((!isBerserkUp && energy >= 50) || (isBerserkUp && energy < 25))   
And yet it's absolutely 0 difference in dps when I modify it like that. Why is this so?.


Because it is not a dps up to wait for low energy to use your 5 point FBs during BiTW?

Many things we do in the rotation are tradeoffs, a little more of this at the expense of that. Putting that conditional in trades off some FBs for Shreds as you are wasting combo points. There may be other subtle side effects, which you can see if you analyze the damage breakdowns in the Mew output. It also, of course, varies depending on what gear someone has. Limiting the energy on FBs during BiTW phase is something several of us has tested and the fact that it is not in currently in the script tells you that non one yet has found it to be a dps increase.

Often, people come up with ideas for changes to the rotation that may or may not be dps improvements. So we test them, and keep those that are dps increases while discarding those that are not. The complexity of what is and is not being traded off is beyond simple napkin math, which is why Mew (and similar tools such as Simulationcraft) are such good tools to check out various ideas. Do not be surprised if you have to work at finding things that improves the rotation at this point as several people have been working on improving it throughout Cata. But don't let that discourage you, as I would be the last person in the world to claim that we have the best possible rotation now. Yo will notice that periodically I answer questions with hard data from Mew. This is often because someone has suggested something that has caught my attention, and rather than pontificate, I simulate, measure and report. You should figure that most of the obvious stuff has already been measured, at least for a Patchwerk style fight. Also figure that we have only done limited testing of more complex fight scenarios although it is possible to emulate very complex fights in Mew once you know how to use the script effectively (for example, I can't tell you the optimum strategy to use for adds on Yor'sahj is, although there are some very interesting suggestions in the Raid advice about how to mix in keeping your DoTs up with Swipe spam).

Revered
User avatar
Posts: 446
Joined: Wed Jun 29, 2011 4:49 am

Re: Mew Releases

Postby Stenhaldi » Sat Feb 25, 2012 4:08 pm

But it is a damage boost to wait for high energy (since ferocious bite converts energy to damage at a better rate than shred), which was the point of the post.

The berserk condition is a bad idea though - there's no reason to avoid ferocious bite during berserk. It's still at least as efficient as shred, but also burns more energy in order to more quickly reach the point where it becomes more efficient.

Exalted
User avatar
Posts: 1145
Joined: Fri Jun 04, 2010 4:28 am

Re: Mew Releases

Postby Leafkiller » Sun Feb 26, 2012 4:19 am

Stenhaldi wrote:But it is a damage boost to wait for high energy (since ferocious bite converts energy to damage at a better rate than shred), which was the point of the post.


Well, that is not what he tested, and I will reserve judgement on your assertion until I see some testing that validates or disproves it. To test that properly in the script you have to factor in when you are energy pooling ans when you are not. The script does no energy pooling during Berserk and TF. Perhaps you can be more specific as to what conditions you believe it is a dps improvement to pool energy with 5 combo points before hitting FB during the BiTW phase and we can measure it.

I have tested trying to energy pool for FB in the past and it has always been a dps loss, but I don't recall testing it with BiTW. I don't know if Mihir or Yawning tested that when they did their work for 4.3.

Posts: 5
Joined: Fri Feb 24, 2012 9:04 am

Re: Mew Releases

Postby amilee » Sun Feb 26, 2012 9:16 am

I hope my post didn't sound to you like I thought I found a glaring mistake in the script. I can quite well imagine how much time you and others have invested into it. I just didn't understand why my change (whether good or not) made no difference. This however I found out by now - I made the change in another script file than I used with mew ;)

Leafkiller wrote:
Stenhaldi wrote:But it is a damage boost to wait for high energy (since ferocious bite converts energy to damage at a better rate than shred), which was the point of the post.

Well, that is not what he tested

But it was my intention. I wanted to test pooling for FB outside of berserk and only doing low energy FB's during berserk, regardless of BitW.

I have a couple of questions regarding the script / sim engine:
- How often is getAction(..) being called with "High Resolution Timer" unchecked / checked?
- Does the sim engine simply do nothing if getAction(..) returns an action which can't be carried out due to lacking cp's or lacking energy or is there more to it?
- How would energy pooling work?

If there should be documentation on this kind of stuff please feel free to just point me to it instead of reciting it here.

Exalted
User avatar
Posts: 1145
Joined: Fri Jun 04, 2010 4:28 am

Re: Mew Releases

Postby Leafkiller » Sun Feb 26, 2012 2:38 pm

You can start here: http://code.google.com/p/mew-wow-druid-model/w/list

Also, you might want to post your questions here: https://groups.google.com/group/mew-wow

I don't know how frequently Yawning checks the posts here.

Honored
Posts: 76
Joined: Sat Jun 04, 2011 4:21 am

Re: Mew Releases

Postby Yawning » Tue Feb 28, 2012 5:02 am

amilee wrote:I hope my post didn't sound to you like I thought I found a glaring mistake in the script. I can quite well imagine how much time you and others have invested into it. I just didn't understand why my change (whether good or not) made no difference. This however I found out by now - I made the change in another script file than I used with mew ;)

I have a couple of questions regarding the script / sim engine:
- How often is getAction(..) being called with "High Resolution Timer" unchecked / checked?
- Does the sim engine simply do nothing if getAction(..) returns an action which can't be carried out due to lacking cp's or lacking energy or is there more to it?
- How would energy pooling work?

If there should be documentation on this kind of stuff please feel free to just point me to it instead of reciting it here.


This stuff isn't documented at all, though the code for it is fairly straight forward (the Simulation engine itself isn't all that big, though there are some quirks to the code and some outright ugliness in places). FWIW most of the answers are in yawning.mew.sim.SimulationEngine.

Anyway.

There's no set time when getAction(...) is called, because this allows the implementation to be considerably faster, yet still manage to query the strategy often enough to give accurate and useful information.

Things that prevent getAction(...) from being called:
* Rebirth/Tranquility casts. (No actions taken when you're forced into raid utility)
* scheduleIdleTime(double) from within the script. (No actions taken when the script says it will be idle for a while.)

Things that cause getAction(...) to be called:
* Time being 0.0 (Encounter start)
* Whenever a event is set to occur in the simulation space. (Auto attacks, DoT ticks, buff gains/fades, etc etc etc.)
* Whenever you exit a global cooldown.
* Every 1 ms whenever there's absolutely no events scheduled. (Not applicable for cats since the energy regen event will always be scheduled.).

Setting the High Resolution Timer option adds:
* Every 1 ms whenever you're not in a global cooldown.

The attentive readers would be going "Hey, doesn't that limit the frequency at which you can take actions off the GCD?", which is correct. I would look into revising the behavior if someone every presents a clear case for it but I don't particularly see any advantage in a button that essentially would be "make the simulation run really slowly".

The current High Resolution Timer option essentially dramatically increases simulation time granularity at moments when it's interesting to do so (you can take on GCD actions, but you haven't for some reason) and goes back to the default of "whenever something one would take actions off occurs" otherwise (Since Mew has cold robotic perfection, it will happily and perfectly hit a TF + Berserk + a pot + a trinket + a CP builder simultaneously, so off GCD actions don't particularly suffer from this either).

Whenever getAction(...) is called, the engine will attempt to execute the specified action (Depending on if the action taken is on the GCD or not, it will call getAction(...) repeatedly). If it fails for any reason, it will execute events scheduled to occur at this particular moment (potentially nothing, there's one rather overloaded event that has an empty execute handler that is scheduled to force getAction(...) calls), and then move on.

To implement energy pooling you would just return null from getAction(...) if you want to pool.

Leafkiller wrote:I don't know how frequently Yawning checks the posts here.


Once a week or so.
"Understanding is a three-edged sword."

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

Previous

Return to Kitty DPS

Who is online

Users browsing this forum: Ekthelion and 7 guests