Ask Mr. Robot or Simulation craft?

[elfporn]

Nah it´s not... Opinion, it´s simcraft data + looking at what actual top players use, ie no mastery if they can help it.

simcraft data corroborates swol's results.

Following top players isn't always the best way to play. The prevailing opinion for a very long time was that mastery was junk, but we had a lot of it anyway because of gear(and many people were gearing for ass/sub after outlaw was nerfed in the first week) - no one has been raiding with 10% mastery to see if you do more dps with such imbalanced stats. Finally sims are in a decent place to simulate a large range of stats and gear with a fairly accurate model of outlaw (the most complicated spec to sim btw) and sims say your desired stats change significantly depending on what legendary/gear you have. It's that simple.

[Zoopercat]

@Wunju - did you not use a custom gearing strategy? The defaults are just for one setup, like guides are for one setup. I can run a custom setup for you to show you, if you give me your character name and realm. It sounds like you aren't familiar with how to use AMR properly, so I can walk you through a few things.

[Deleted]

@Wunju - did you not use a custom gearing strategy? The defaults are just for one setup, like guides are for one setup. I can run a custom setup for you to show you, if you give me your character name and realm. It sounds like you aren't familiar with how to use AMR properly, so I can walk you through a few things.

I´m familiar, it did not live up to expectations, I stopped using it, it´s simple enough. Maybe in some months when it´s more polished i´ll look into your tool again.

[elfporn]

I´m familiar, it did not live up to expectations, I stopped using it, it´s simple enough. Maybe in some months when it´s more polished i´ll look into your tool again.

What we're trying to tell you is that it's polished now. That results are synonymous with simcraft when both tools are used correctly.

The reasons for the odd outlaw results have been explained and verified by independent parties. Sub and assassination results have beeen extremely accurate.

What more do you want? Assistance with using the tool has been offered.

You also have not specified what did not live up to your expectations, what exactly is "wrong" with AMR? Are you expecting exact results as compared to simcraft? (Won't happen, two different sims using a different range of data points[gearsets])

FYI, you can modify the entire APL in AMR to suit your particular playstyle, as an example, i am slowly altering the sub APL for myself to more closely resemble how i play.

It's tough for zooper and swol to figure out if your'e just being hardheaded, don't understand the tool, or have found some bugs if you don't give specifics.

[ymirsson]

It did not support his confirmation bias, must be a faulty program.

[Zoopercat]

I should also make a post on the differences between the 2 programs. One isn't right over the other, they just have different assumptions in back-end things that probably only a dozen or so people even know about. That will help people understand a little bit more when things are different too, and some of those things can be configured to match. I'll work on that and hit you all up when it's ready.

[Deleted]

What we're trying to tell you is that it's polished now. That results are synonymous with simcraft when both tools are used correctly.

The reasons for the odd outlaw results have been explained and verified by independent parties. Sub and assassination results have beeen extremely accurate.

What more do you want? Assistance with using the tool has been offered.

You also have not specified what did not live up to your expectations, what exactly is "wrong" with AMR? Are you expecting exact results as compared to simcraft? (Won't happen, two different sims using a different range of data points[gearsets])

FYI, you can modify the entire APL in AMR to suit your particular playstyle, as an example, i am slowly altering the sub APL for myself to more closely resemble how i play.

It's tough for zooper and swol to figure out if your'e just being hardheaded, don't understand the tool, or have found some bugs if you don't give specifics.

How can assassination results be extremely accurate when not long ago I tried it and it gave MA relics literally zero value (among other discrepancies I found)? I mean hey you like it, you even seem invested in convincing others to use it is for some reason, like I said I find it inacurate, that does not make me "hard headed", kinda weird you´re borderline insulting someone for not using some software

[Zoopercat]

@Wunju - I think it all comes back to the fact that you probably need to run a custom gearing strategy. Months ago there are threads where I talked to people about the MA thing, and it was always that they didn't have custom stat weights.

To run custom ones, follow this tutorial (you need to do the first part, can't skip to the ML part until the first part is done). http://i.imgur.com/r2gpuBW.jpg Or, if you give me your character name and realm, I can run the weights for you, let my computer do the heavy lifting and then link you a set of weights that you can just save to your character.

Also, if you see anything that's inaccurate now, let me know specifics and I will look into it right away.

[elfporn]

How can assassination results be extremely accurate when not long ago I tried it and it gave MA relics literally zero value (among other discrepancies I found)? I mean hey you like it, you even seem invested in convincing others to use it is for some reason, like I said I find it inacurate, that does not make me "hard headed", kinda weird you´re borderline insulting someone for not using some software

I'm trying to get feedback out of you so that zooper can take it and make rogue sims even better.

We get a ton of attention from her and occasionally swol. I can post up a reply here saying something doesnt look right and she swings by, investigates, and implements change. It's tought to beat that.

Make the best of it and post up bugs/feedback when you see them.

[ymirsson]

How can assassination results be extremely accurate when not long ago I tried it and it gave MA relics literally zero value (among other discrepancies I found)? I mean hey you like it, you even seem invested in convincing others to use it is for some reason, like I said I find it inacurate, that does not make me "hard headed", kinda weird you´re borderline insulting someone for not using some software

It's borderline insulting to talk bad about a project several people poured lots of hours into without as much as giving even a smallish hint what you did to recieve your results.
Not to mention the fact that you never tire to tell people AMR is wrong.

[Zoopercat]

Actually, just to make sure @Wunju sees what I'm seeing. I use this program all day so sometimes I don't know if people know about the features and tricks that I do.

Here's the MA wiki page: You can see how we have every spell and item implemented (and even edit it). You can see how we have MA coded to reduce the CD.


Then, to make sure there isn't a bug somewhere, I ran a simulation with 3 MA equipped. Then, I scrolled down below the spell chart to the section that says 'log.' (Note, you have to select 'log' in the drop down in 'more options' when you start a sim to get this view). As you can see, MA is used about 72 seconds later.



One other possibility - maybe you ran a sim that was a short duration? In order to get 4 vendettas in over 3, the fight would have to be 224 seconds long, and then 20 more seconds to get the full benefit, for a total of 244 seconds. Maybe that's what you saw?

[howolowitz]

Wow this thread took off! Thanks for all the advice! I looked at all your suggestions and tried running the gearing strategy in askmrrobot but i keep getting an error. It runs for quite a while and then i get error: No solution to the regression could be found. And the client also gives an error. I've tried it 3 times now. Any ideas? My pc isnt going into sleep mode or anything.

[cface]

Well I'm convinced to give mrrobot a good long shot, thanks for this thread.

[Deleted]

Thx for all advices bro

[Deleted]

@Zoopercat no, i simmed it, used custom weights, it just gave ZERO value to MA relics. No idea if you fixed that in the meantime or not, it was a couple months ago.

[Deleted]

Wow this thread took off! Thanks for all the advice! I looked at all your suggestions and tried running the gearing strategy in askmrrobot but i keep getting an error. It runs for quite a while and then i get error: No solution to the regression could be found. And the client also gives an error. I've tried it 3 times now. Any ideas? My pc isnt going into sleep mode or anything.

I've been getting this one too, ran three times with the same error. Only seems to be since the updated version.
I'm just trying it on my laptop to see if it completes here.

- - - Updated - - -

For what it's worth, i'm really enjoying using AMR. Even *if* (As some people may think) it isn't 100% accurate compared to simulationcraft, the ease of use and extra features alone make it a far more valuable tool given a small margin of error. On that note however, after running custom gearing strats, I'm not finding a huge difference between Simcraft/AMR/My own meters anyway.

I'm not saying that Simulationcraft is not worth using - as players we need to use our own experience, intuition and a range of different tools to make decisions. If they corroborate, then we know we are on the right track. Relying on only one source (Even just Simulationcraft alone) is not ideal.

I think what annoys me most about the AMR vs Simcraft debate from what I've seen is that the defenders just state it's better but never seem to be able to expand on why. Ultimately it's probably just because someone else has told them it is.
The more we can do to feedback if/when we believe something is wrong, the better the tool becomes. Instead of saying Simulationcraft is better, state the reasons why so the devs can either investigate or explain more.

[Zoopercat]

@Zoopercat no, i simmed it, used custom weights, it just gave ZERO value to MA relics. No idea if you fixed that in the meantime or not, it was a couple months ago.

Hmm, I guess I'll chalk it up to misreading the report or something - since it's been working since legion-launch. Here's a thread where I address a similar concern back in November, so more than a couple months ago when you tested it. http://www.mmo-champion.com/threads/...5#post43512415

Hmm, do you have sims you can link me to show that - I don't see that in the sims I setup. I just did 3x 840 master assassin and compared it to 3x 840 rupture.

Master assassin: 340,408 DPS http://www.askmrrobot.com/wow/simula...3248354ada80fc
Rupture: 337,994 DPS http://www.askmrrobot.com/wow/simula...d167562e106e97

With different talents or artifact progress it could be different. So linking sims that you do will help

@howolowitz @demiise We have an update going out soon (a few hours I think) that we hope fixes that problem. In the meantime, if you give me your character name and realm, along with any special settings you might want (potions, fight length, etc), I can run them on my machine and send them over to you.

[Deleted]

Now thats service And another reason i'm happy to support you guys!

No problem, i'm not in any major rush. I have a strategy ran from a few weeks back. Just thought I'd test the updated version, so i'm happy to wait and run it once the new version is out.

[howolowitz]

Hey thats great service! Thanks alot. I wanted to run a gearing strat on my rogue. Snikrot-Ragnaros EU ilvl min 870 max 920. For the rest i dont have any special settings.
Thanks again!

[Deleted]

@Zoopercat

I am trying to simulate my shaman using askmrrobot - recently with the new enha golden trait and buffs on the legendary wrist a lot of players have switched to playing hothand with t19 4 set.

Any ideas why askmrrobot still showing BF better? Simulationcraft shows hothand as 14k higher

In 7.1.5 Boulderfist was our best build with t19.

Askmrrobot is still showing Boulderfist superior even with the new hothand + akainu+ eye of twisting nether+ hailstorm build which should be the best single target build ideally.
I mean if boulderfist is still superior fair enough but if it isnt it makes me wonder why everyone has changed to hothand if its not better with the said legendaries.