ravenshaun

Why is Engine ranked so low on Mr Robot???

Sign in to follow this  

12 posts in this topic

Lucky raid last night I got both the 915 Cradle of Anguish and the 920 Engine. Checked MR Robot and it has the Engine as way down the list, my understanding we it was better?....help please

https://worldofwarcraft.com/en-us/character/bloodhoof/kalypsom

 

 

Share this post


Link to post
Share on other sites

AMR is very unreliable esp when it comes to set bonuses and trinkets. I would say using your H WF setenial and engine are great esp since your engine has a socket.  you can always run your own sims too to see what works best for your set up. Side note: i looks like you are running the stomp build but have chosen BF over OwtP. Any reason for this? Stomp and OwtP go hand in hand. 

 

  • Like 1

Share this post


Link to post
Share on other sites

My guess is it's coded for you not grabbing the orbs or something dumb like that. Maybe AMR thinks it's a melee-only trinket since it procs from auto-attacks.

It makes no sense though. they have a 930 Engine below an 875 Eye of Guarm. It's ridiculous.

Use raidbots.com, it's better.

Edited by Khallid
  • Like 3

Share this post


Link to post
Share on other sites
21 hours ago, ravenshaun said:

Lucky raid last night I got both the 915 Cradle of Anguish and the 920 Engine. Checked MR Robot and it has the Engine as way down the list, my understanding we it was better?....help please

https://worldofwarcraft.com/en-us/character/bloodhoof/kalypsom

 

 

That's because it sucks.  Miss one orb, bad proc and it becomes garbage. 

Play around with the following links, keep in mind when you see someone with a blue dungeon trinket it is a high titanforge Mythic +15 trinket: 

https://www.warcraftlogs.com/statistics/13#class=Hunter&spec=BeastMastery&combatantinfo=Trinkets

https://www.warcraftlogs.com/statistics/13#class=Hunter&spec=BeastMastery&combatantinfo=Trinkets&difficulty=4

Play around with the % and each boss and difficulty. This is actual data vs. sim data. Also, the majority of those arcano crystals are 880 from sent farming. 

 

 

Share this post


Link to post
Share on other sites
7 hours ago, Banard said:

That's because it sucks.  Miss one orb, bad proc and it becomes garbage. 

Play around with the following links, keep in mind when you see someone with a blue dungeon trinket it is a high titanforge Mythic +15 trinket: 

https://www.warcraftlogs.com/statistics/13#class=Hunter&spec=BeastMastery&combatantinfo=Trinkets

https://www.warcraftlogs.com/statistics/13#class=Hunter&spec=BeastMastery&combatantinfo=Trinkets&difficulty=4

Play around with the % and each boss and difficulty. This is actual data vs. sim data. Also, the majority of those arcano crystals are 880 from sent farming. 

Come on, those statistics are mind-blowingly dumb. They are comparing different people using different gear altogether. Most of your dps doesn't come from just one piece of gear. You are analyzing one variable (dps output) which depends on 17 pieces of gear, not 1 (plus artifact power). This is the definition of a badly designed experiment that doesn't control for anything. All the results in those tables are meaningless garbage.

Kil'Jaeden's Burning Wish is at the bottom. Sure it's weak compared to other legendaries, but it's the strongest trinket in the game by a very substantial margin, and you would take it if you had no other legendaries. So it's at the bottom because people using this trinket have bad gear, not because the trinket itself is weaker than the other trinkets. It's a self-selecting group.

Hunger of the Pack is at the top because if you have a Mythic 15 hunger of the pack then it's a good choice that only already high geared characters can get. You don't see weak noobs with mythic 15 pieces of gear because they can't clear that level, and mythic 15 pieces are by definition of high item level. Another self-selecting group.

 

Straight-up statistics are only valid after you prove that your measuring value depends exlusively on the variables in question, and when the distribution behaves normally. Both things are being broken here. Self-selecting groups don't make a normal bell distribution, and measuring one variable out of 17 doesn't match the variable-dependency.

Edited by Khallid

Share this post


Link to post
Share on other sites
15 hours ago, Khallid said:

This is the definition of a badly designed experiment that doesn't control for anything. All the results in those tables are meaningless garbage.

 

Data is not garbage. Only those who cannot utilise it.  Those data tables are the first step in understanding if the sims are correctly siming trinkets.  I never said its the be all, ends all. There is always more analysis to do but you need to stop ignoring relevant information. 

For example, there enough evidence stating engine is overvalued in sims and the tomb off the last boss is undervalued. No shape or form can i get tomb siming that well. 

Looking at the trinket, its quite obvious its for the stomp build....but still no where can i get it to sim that high. And its doing really well across  enough parses to indicate something is on the go with the sims. 

Share this post


Link to post
Share on other sites
3 hours ago, Banard said:

Data is not garbage. Only those who cannot utilise it.  Those data tables are the first step in understanding if the sims are correctly siming trinkets.

You're not getting it, man. This is statistics 101. Aggregated data is useful when used properly. Averaging dps the way those charts do is just wrong. You can make a lot of deductions from looking at parses, but those are not it. You seem to think that the sheer number of parses make anything from them valid, but that's not how math works. You can prove anything with statistics if you do it wrong. If you've ever done a simple analysis of variance you would know that you can't make a correlation until you capture all variables and their interactions, and then all the analysis is invalid if the sample does not follow a normal distribution.

Do you even understand what I mean when I say this is not a normal distribution? When a distribution is not normal, the average and the mean value are different. The common mathematical approach literally doesn't work. The variance is not constant across the range. I'm talking in hard mathematical terms, this is not my opinion. These charts are averaging dps without considering anything, which is invalid. Invalid doesn't mean that I can't figure it out, it means the averaging operation is wrong. It's akin to adding apples and oranges and saying that the average is a peach. That's what's wrong with the charts you posted.

I'm assuming you refer to the "Tome of Unraveling Sanity" when you said "Tomb". There are very obvious reasons why it gets good parses. It's because of the self-selecting manner of how it drops and who gets it. It's loot from the final boss, which on average is only killed by the better portion of the playerbase. But it's also higher item level by default, because Kil'Jaeden's loot is higher item level than the rest of the raid at the same difficulty, so it's not being compared to the rest of the trinkets on a level field. But the real issue here is that you're believing the parse averages to mean that one trinket is better than the other. I'm telling you that mathematical operation is invalid. Sims are much better because they compare your output with your gear, not an average of gear from different people at different ilvls.

The fact that the numbers are averaged all across, ignoring all other factors, can give you any result. Have you ever heard about randomized trials? A good randomized trial has to control for all the independent variables and keep them steady. The players using those trinkets simmed them and selected them as their best choices, so the sample is not random and averages are not valid.

Titanforging makes the problem worse because a lot of mythic players are using titanforged Entwined Elemental Foci from Nighthold. It can easily beat better trinkets by virtue of its high item level. The high variability in item level introduces another factor that the averages don't catch, and there's no good way to control for it in aggregate.

If you look at top parses, for example, it almost doesn't matter which trinket you're using. People get good parses with just about any trinket combination. And this makes sense because one piece of gear doesn't determine your whole dps, plus there's no reason to expect everyone to have the perfect loot, they just use the best they got, even if there is a theoretical better trinket they don't have. There are even Bloodthirsty instincts being used by some players, probably highly titanforged versions of it, but on average BTI is out because it's unreasonable to get a really high ilvl one.

  • Like 1

Share this post


Link to post
Share on other sites
3 hours ago, Khallid said:

You're not getting it, man. 

Man your just not getting it. 

I'll repeat "I never said its the be all, ends all. There is always more analysis to do but you need to stop ignoring relevant information." 

I find it hilarious you think you can educate me on statistical analysis.   Again look at the data, there is a lot you can take from it in comparison to sims.  Its not the end all be all but there is quite a bit of information that is relevant.  

I am also not going to do the analysis here for free on a website that makes $$. I just highlighted to the OP another source of information to compare along with the sims. 

And to keep in mind that miss one orb, one bad proc, and the engine becomes garbage compared to a arcano. 

Also, https://www.warcraftlogs.com/statistics/13#class=Hunter&spec=BeastMastery&combatantinfo=Legendaries&aggregate=amount

This seems to match the sims.  As i stated above and as other has stated, the sims dont necessarily sim trinkets correctly. Its best to compare what your seeing in sims to live data.  

 

 

Edited by Banard

Share this post


Link to post
Share on other sites
26 minutes ago, Banard said:

I find it hilarious you think you can educate me on statistical analysis.   Again look at the data, there is a lot you can take from it in comparison to sims.  Its not the end all be all but there is quite a bit of information that is relevant.

Averages are not data, how do you not understand that? You need to make an invalid mathematical operation to turn the data into those tables. I'm not arguing against data, I'm arguing against using data incorrectly.

Share this post


Link to post
Share on other sites

Guys, again with what Blainie said; keep the discussions civil. 

To the OP, Engine is much the same as Entwined Elemental Foci was - it's a feast or famine trinket, if you get good procs and pickup all four orbs it's good, but if you're unlucky and get 1-2 procs on a 3-4 minute fight it's trash. For BM, generally stat sticks will perform better. Or items that have a consistent proc rate / usage, e.g. BTI or Tome of Unraveling Sanity.

Share this post


Link to post
Share on other sites

to sisco

"Side note: i looks like you are running the stomp build but have chosen BF over OwtP. Any reason for this? Stomp and OwtP go hand in hand."

yes there is a reason, do to my personal play style and after about 6 hours of raiding and target dummy wall bashing, this is what seems to give me the highest consistent dps. I change it some for AOE fights. so its a more so works for me thing :) . I tried the 4/2 stomp build after weeks of trying to get shoulders and it was lack luster.

I'm still hoping to get the ring soon that will allow me both OWTP and BF, but in the mean time, its figuring out these trinkets.

Thanks for the advice!

 

 

Share this post


Link to post
Share on other sites
On 7/19/2017 at 0:16 PM, ravenshaun said:

Lucky raid last night I got both the 915 Cradle of Anguish and the 920 Engine. Checked MR Robot and it has the Engine as way down the list, my understanding we it was better?....help please

As has already been said, AMR is garbage... use www.raidbots.com or download simcraft...

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this  

  • Recently Browsing   0 members

    No registered users viewing this page.