[Feral] Legion Discussion

[RayenDark]

I see TF just got massively nerfed in 7.1?

+60 down to +20 rage? O_o

TF hasn't been touched. 20 engery is now baseline for non feral specs. Feral gets a passive to make is 60 again.

[vokey]

I have a question. I'm currently running Predator - Soul of the Forest - Bloodtalons for normal dungeons while gearing up, and I'm not quite sure when to use Berserk and TF during my opener. Using Ashamane's Frenzy after Rake I don't find myself low enough on energy to use TF, and I was used to putting Berserk and TF up together in WoD (last time I played my feral was in early HFC).

Do you have any tips on a opener with that spec?

[Deleted]

I have a question. I'm currently running Predator - Soul of the Forest - Bloodtalons for normal dungeons while gearing up, and I'm not quite sure when to use Berserk and TF during my opener. Using Ashamane's Frenzy after Rake I don't find myself low enough on energy to use TF, and I was used to putting Berserk and TF up together in WoD (last time I played my feral was in early HFC).

Do you have any tips on a opener with that spec?

Rake > shred > shred > Rip > shred > shred > TF > Berserk > HT > AF > Rip


and we're back to AMR being about 15k off from simcraft, So they've made huge improvement from the last time i looked at its no longer a 100k difference between the two. But still enough off that i wouldn't use it until its improved.

[Yellowfive]

Hey some people asked me to take a look at the AMR simulator for feral, see if I can see what's different between the two. I'm the main developer of the simulator, and focus more on making the core engine work than the specific rotations and all that. So my approach tends to be more methodical, I try to isolate and reduce the number of variables.

The biggest difference between the two that I see is Ashamane's Frenzy damage -- everything else looks pretty much the same. But that difference is quite large. So I went in-game and got the artifact weapon on my wife's level 100 druid this afternoon, and ran a simple test on a level 100 target dummy. As far as I can tell, simc isn't even in the right ballpark for the bleed damage.

Here is a pruned log with the events:
http://pastebin.com/aWCcdQn5


Here's the AMR setup and result:
http://www.askmrrobot.com/wow/simula...8276b9248b8adb


Here's the simc setup:
http://pastebin.com/x1NiDyMr

When you run this in simc, go to the "Options" tab, "Globals", set Target Level to Max Player Level (this will make it same level as the player rather than 3 above, impacts the amount of armor and thus the damage of the direct hits). Also go to Buffs/Debuffs tab and uncheck everything on both panes. Also set Generate Debug to Log Only so you can see the specific events -- reporting for Ashamane's Frenzy in simc is pretty bad, doesn't separate out the direct and DoT hits so hard to tell what's happening without a log. Here's the resulting simc log for the first use:

http://pastebin.com/zYkVPvQt


For people who don't want to dig into that:
IN GAME:
direct hit damage x15: 1060 normal, 2163 crit
first tick damage: ~11325 normal, 23103 crit
last 4 ticks damage: ~14155 normal, 28877 crit

(Note in that log I got lucky and got a crit on every tick. I did do multiple tests though. Also note that sometimes the game does the first tick after 10 hits (as in this case), sometimes after 11 hits. Doesn't really matter, but changes the damage of the ticks -- total damage remains the same though.)

AMR:
direct hit damage x15: 1060 normal, 2162 crit
first tick damage: 11328 normal, ~23109 crit
last 4 ticks damage: 14160 normal, 28887 crit

SIMC:
direct hit damage x15: 1061 normal, 2163 crit
first tick damage: 15118 normal, 30840 crit
last 4 ticks damage: 22677 normal, 46260 crit


Maybe someone else will see something that I don't... but right now it seems to me that the bleed part is bugged in simc -- massively inflated damage. Someone should probably review and fix this ASAP. In the meantime I'd recommend using AMR -- ours matches the game exactly.

I know that our simulator is new... and we need to build up some trust with the community, but one of my major goals with the design of the new simulator was something that ANYBODY can verify and set up. You can very easily see what we are doing for this ability:

http://www.askmrrobot.com/wow/theory...l&version=live

On that page I can click through to every mechanic, clearly see that mastery buffs both parts (look at the "referenced by" section at the bottom), one part is reduced by armor and the other isn't, using a 20% AP coefficient, etc. It's all right on the page.


In simc here is all I see:

http://pastebin.com/4dAJNxj7
http://pastebin.com/SyVZd6Ms

Cool. Good luck figuring out if that is correct. And this is a huge problem -- even as a programmer, it would take me hours to figure out how to go about fixing this problem in simc. So all I can do is shrug and hope someone comes along and fixes it who knows the C++ inside-out.

I strongly urge you to give our new simulator a try -- the potential benefits to the TC community are huge. Hit me up on our discord if you EVER find anything not working or want guidance on how to modify things yourself, or just want to talk more about my vision for taking theorycrafting tools to the next level.

[Sar-]

Is there any chance the AMR website will work properly on mobile?

It's a mess atm when trying to run a sim or look at a char's gear options.

[vokey]

Rake > shred > shred > Rip > shred > shred > TF > Berserk > HT > AF > Rip


and we're back to AMR being about 15k off from simcraft, So they've made huge improvement from the last time i looked at its no longer a 100k difference between the two. But still enough off that i wouldn't use it until its improved.

Rake has usually worn out by the time I use AF in this scenario though, isn't that pretty bad?

[Swol]

Is there any chance the AMR website will work properly on mobile?

It's a mess atm when trying to run a sim or look at a char's gear options.

This is on our list to fix up. When we analyzed our traffic from WoD, we noticed that very few people were using the site on mobile devices, so we decided to delay the mobile form factor support until we get some other features finished and tightened up.

I have a fix for Safari browsers that I'm working on, which might help a little for people with iPhones. It's really hard to debug Safari when you don't own a Mac

[Sar-]

Aye, I tend to use the site more on my PC, but would use it a lot more if the mobile version were working

Either way, it's well worth the sub money each year

[-aiko-]

and we're back to AMR being about 15k off from simcraft, So they've made huge improvement from the last time i looked at its no longer a 100k difference between the two. But still enough off that i wouldn't use it until its improved.

In my experience, though, AMR simulations are closer to real-life results. I'm currently simming at around 330k with SimC but in practice that seems more and more like an unattainable, fantasy number. Whereas AMR results I've actually been able to manage.

[Tradu]

Hey some people asked me to take a look at the AMR simulator for feral, see if I can see what's different between the two. I'm the main developer of the simulator, and focus more on making the core engine work than the specific rotations and all that. So my approach tends to be more methodical, I try to isolate and reduce the number of variables.

The biggest difference between the two that I see is Ashamane's Frenzy damage -- everything else looks pretty much the same. But that difference is quite large. So I went in-game and got the artifact weapon on my wife's level 100 druid this afternoon, and ran a simple test on a level 100 target dummy. As far as I can tell, simc isn't even in the right ballpark for the bleed damage.

Here is a pruned log with the events:
http://pastebin.com/aWCcdQn5


Here's the AMR setup and result:
http://www.askmrrobot.com/wow/simula...8276b9248b8adb


Here's the simc setup:
http://pastebin.com/x1NiDyMr

When you run this in simc, go to the "Options" tab, "Globals", set Target Level to Max Player Level (this will make it same level as the player rather than 3 above, impacts the amount of armor and thus the damage of the direct hits). Also go to Buffs/Debuffs tab and uncheck everything on both panes. Also set Generate Debug to Log Only so you can see the specific events -- reporting for Ashamane's Frenzy in simc is pretty bad, doesn't separate out the direct and DoT hits so hard to tell what's happening without a log. Here's the resulting simc log for the first use:

http://pastebin.com/zYkVPvQt


For people who don't want to dig into that:
IN GAME:
direct hit damage x15: 1060 normal, 2163 crit
first tick damage: ~11325 normal, 23103 crit
last 4 ticks damage: ~14155 normal, 28877 crit

(Note in that log I got lucky and got a crit on every tick. I did do multiple tests though. Also note that sometimes the game does the first tick after 10 hits (as in this case), sometimes after 11 hits. Doesn't really matter, but changes the damage of the ticks -- total damage remains the same though.)

AMR:
direct hit damage x15: 1060 normal, 2162 crit
first tick damage: 11328 normal, ~23109 crit
last 4 ticks damage: 14160 normal, 28887 crit

SIMC:
direct hit damage x15: 1061 normal, 2163 crit
first tick damage: 15118 normal, 30840 crit
last 4 ticks damage: 22677 normal, 46260 crit


Maybe someone else will see something that I don't... but right now it seems to me that the bleed part is bugged in simc -- massively inflated damage. Someone should probably review and fix this ASAP. In the meantime I'd recommend using AMR -- ours matches the game exactly.

I know that our simulator is new... and we need to build up some trust with the community, but one of my major goals with the design of the new simulator was something that ANYBODY can verify and set up. You can very easily see what we are doing for this ability:

http://www.askmrrobot.com/wow/theory...l&version=live

On that page I can click through to every mechanic, clearly see that mastery buffs both parts (look at the "referenced by" section at the bottom), one part is reduced by armor and the other isn't, using a 20% AP coefficient, etc. It's all right on the page.


In simc here is all I see:

http://pastebin.com/4dAJNxj7
http://pastebin.com/SyVZd6Ms

Cool. Good luck figuring out if that is correct. And this is a huge problem -- even as a programmer, it would take me hours to figure out how to go about fixing this problem in simc. So all I can do is shrug and hope someone comes along and fixes it who knows the C++ inside-out.

I strongly urge you to give our new simulator a try -- the potential benefits to the TC community are huge. Hit me up on our discord if you EVER find anything not working or want guidance on how to modify things yourself, or just want to talk more about my vision for taking theorycrafting tools to the next level.

Report your "bugs" to the SimC devs directly, repeatedly publicly posting bugs about competitors while pushing your own simulator isn't helping build any "trust".

[Deleted]

Report your "bugs" to the SimC devs directly, repeatedly publicly posting bugs about competitors while pushing your own simulator isn't helping build any "trust".

When this one is proved to be not a bug again i'll fucking laugh

In my experience, though, AMR simulations are closer to real-life results. I'm currently simming at around 330k with SimC but in practice that seems more and more like an unattainable, fantasy number. Whereas AMR results I've actually been able to manage.

then frankly. Get better. I guarantee you're not playing perfectly

I know that our simulator is new... and we need to build up some trust with the community, but one of my major goals with the design of the new simulator was something that ANYBODY can verify and set up. You can very easily see what we are doing for this ability:

I guess thats why your code is open source

[Zoopercat]

Report your "bugs" to the SimC devs directly, repeatedly publicly posting bugs about competitors while pushing your own simulator isn't helping build any "trust".

I did report it and they are looking into it. When people claim AMR is bad without any evidence, we do tests to find out why it's different than SimC. The post a couple of days ago about rake was wrong (I corrected that, and it's because I couldn't get the relics to work right in SimC, my fault). So we looked again, bc the damage is still off and Ashamane's frenzy is doing too much damage in SimC. If people would stop blindly saying 'lol amr sucks' I wouldn't have to counter it like this. I hate doing it, but if people spread misinformation, we have to point it out.

I also feel it's helpful to point out how to do the test, so other people can help get involved with testing. Yellow clearly laid out how the test worked and provided profiles for people to follow along. Who knows, even a few people helping with bug testing could make a big impact and really help out the devs

@Guiltyas if you can do a test to prove it isn't a bug, share it. I'm sure the SimC devs would appreciate that. But based on our evidence, it is bugged. If you find something wrong with the test, do point it out.

In the meantime, we still haven't found bugs in AMR's feral module. And guiltyas hasn't found any either. If you find any, feel free to point it out, and we'll get them fixed in a day. All of the data is open source and fairly easy to verify.

The 2 simulators are very close in damage, so any blanket 'lol' statements don't hold up. It also is a likely indication that there are no major bugs in either simulator.

Until then, everyone can feel confident in the AMR simulator Cheers!

[Sar-]

Report your "bugs" to the SimC devs directly, repeatedly publicly posting bugs about competitors while pushing your own simulator isn't helping build any "trust".

Well when people blindly suck SimC's cock, and equally blindly state AMR sucks because it gives different results to SimC, what do you expect him to say, other than defend his work?

Personally, based on the evidence given, to me it seems AMR is the more accurate of the two sims.

Yet to see anyone posting evidence to the contrary.

[Deleted]

I did report it and they are looking into it. When people claim AMR is bad without any evidence, we do tests to find out why it's different than SimC. The post a couple of days ago about rake was wrong (I corrected that, and it's because I couldn't get the relics to work right in SimC, my fault). So we looked again, bc the damage is still off and Ashamane's bite is doing too much damage in SimC. If people would stop blindly saying 'lol amr sucks' I wouldn't have to counter it like this. I hate doing it, but if people spread misinformation, we have to point it out.

@Guiltyas if you can do a test to prove it isn't a bug, share it. I'm sure the SimC devs would appreciate that. But based on our evidence, it is bugged. If you find something wrong with the test, do point it out.

In the meantime, we still haven't found bugs in AMR's feral module. And guiltyas hasn't found any either. If you find any, feel free to point it out, and we'll get them fixed in a day. All of the data is open source and fairly easy to verify.

The 2 simulators are very close in damage, so any blanket 'lol' statements don't hold up. It also is a likely indication that there are no major bugs in either simulator.

I haven't tested it because frankly at this point i'd rather litterally shoot myself than use this piece of shit. You spent litterally hours attacking me in feral discord because I refused to blindly believe your bullshit about a rake bug and you refused to believe my testing showing me to be correct "because i wouldn't use the profile you linked" only after that for you to prove yourself incompetent.

why in the everliving fuck would i waste my time to go through another set of shitty fucking tests.

and you say you posted the bug? You posted a bug report that said "there are bugs test them" jesus christ if i got a bug report like that it'd be closed within 30 seconds.

And ashamane's bite copies rips damage, so it's demonstrably correct.

my mistake apparently you're talking about ashamanes frenzy not ashamanes bite. so you're at least not 100% retarded

[Zoopercat]

The reason you 'waste time' to go through tests is to ensure quality. Anytime someone thinks something is wrong with AMR or could be improved, we investigate. Sure, it might take an hour or two, but that's how we are able to ensure it is working properly. 90% of the time there is no bug, but that 10% of the time is what we are looking for so we can continuously improve the simulator. We don't wait for other people to test it (though they are welcome to), we do it ourselves. And I definitely don't stick my head in the sand.

If someone says the damage is low, I look into it. I ask questions to see how they setup the test, I try to reproduce it, and then either fix a bug, or explain how their testing profiles didn't match between the two simulators.

I've been emailing and discording with the SimC devs on the specifics, trying to help them track it down. So hang in there while they verify/fix it/or-can't-reproduce-it.

Yes, sorry, Ashamane's frenzy. It's late here, I typo'd

[Deleted]

because you have such a good track record now when comparing the two tools. JK spent hours with the woe is me attitude because you refused to believe that i could demonstrate that your previous bullshit was just that. bullshit. "we do this for the community out of the good of our hearts boo hoo" fuck off. you're in this to make add dollars, as your endless pushing of this garbage demonstrates.

[Yellowfive]

A follow-up for people: the simc guys confirmed that they never circled back and updated/tested Ashamane's Frenzy after a lot of accumulating DoTs from the Legion beta were fixed to behave correctly. So it is indeed doing way too much damage (I tested this back on the beta and talked with the simc guys about it, it was doing like 40% or so more than the tooltip/intended damage was on the beta, and we had no idea what kind of magic-math the game was using).

The difference is that we went back and re-tested after the game went live. This doesn't always happen in simc because... testing is boring, so open source hobby projects don't usually do much in the way of QA unless someone comes along and helps them out.

When you see a difference between simc and AMR like this, you should test it instead of just guessing which one is right! Then report what you find to the devs of both simulators -- we all benefit from two tools that independently verify one another. It makes huge headaches for both me and the simc devs if people are running around saying their shit is bugged. The guys who work on simc have jobs, and a lot of them spend a ton of time playing WoW. They don't have time to go back and test everything. You are doing them (and everyone who uses simc) a disservice by just assuming everything works perfectly.

[Deleted]

A follow-up for people: the simc guys confirmed that they never circled back and updated/tested Ashamane's Frenzy after a lot of accumulating DoTs from the Legion beta were fixed to behave correctly. So it is indeed doing way too much damage (I tested this back on the beta and talked with the simc guys about it, it was doing like 40% or so more than the tooltip/intended damage was on the beta, and we had no idea what kind of magic-math the game was using).

The difference is that we went back and re-tested after the game went live. This doesn't always happen in simc because... testing is boring, so open source hobby projects don't usually do much in the way of QA unless someone comes along and helps them out.

When you see a difference between simc and AMR like this, you should test it instead of just guessing which one is right! Then report what you find to the devs of both simulators -- we all benefit from two tools that independently verify one another. It makes huge headaches for both me and the simc devs if people are running around saying their shit is bugged. The guys who work on simc have jobs, and a lot of them spend a ton of time playing WoW. They don't have time to go back and test everything. You are doing them (and everyone who uses simc) a disservice by just assuming everything works perfectly.

like you tested rake bug before spamming that shite everywhere?

just fuck off allready. i'm bored

[Tellof]

like you tested rake bug before spamming that shite everywhere?

just fuck off allready. i'm bored

Why are you being such a prick? It's pretty obvious that these guys from both AMR and Simc are working way harder than than necessary to provide a service for the community - which is largely a thankless task. What do they owe you specifically, and why are you going out of your way to attack them any time they post anything? How are you actually contributing to make it better? Almost all of your whole 14 total posts are you yelling at people or being negative in some way. I'm amazed they're even responding to you at this point.

@Yellowfive & @Zoopercat I haven't used the AMR sim a lot because I've been too busy just playing the game, but I've used classic AMR a lot over the years and as a programmer I appreciate the work involved. You guys have a nice product and anyone who expects it to be perfect all of the time is clueless, frankly.

[Lucrece]

Why are you being such a prick? It's pretty obvious that these guys from both AMR and Simc are working way harder than than necessary to provide a service for the community - which is largely a thankless task. What do they owe you specifically, and why are you going out of your way to attack them any time they post anything? How are you actually contributing to make it better? Almost all of your whole 14 total posts are you yelling at people or being negative in some way. I'm amazed they're even responding to you at this point.

@Yellowfive & @Zoopercat I haven't used the AMR sim a lot because I've been too busy just playing the game, but I've used classic AMR a lot over the years and as a programmer I appreciate the work involved. You guys have a nice product and anyone who expects it to be perfect all of the time is clueless, frankly.

Just report him to the blue mods. The class forum mod doesn't seem available so just contact the super moderators, with his posting history consisting of a litany of insults to various people it shouldn't be hard to see this kind of vile behavior corrected or removed.