Still no balance a year later - with facts

Heh, I feel for you OP. As you can see, I got really into this game a little over a year back and in less than a month I had my lvl 66 Marauder. Only unaware to the terrible facts I innocently used the dreaded Cyclone build/tactic. I quit playing shortly after reaching slightly high levels since the desync rendered my character completely useless and I'll never stand by a game that makes your personal playstyle unviable due to their shitty servers.

Like I said, that was about 16-18 months ago lol. I check the forums every semester or so, nothing's changed apparently :)

Give up, POE was a good memory for the overly-short time it lasted in my life.
Last edited by flynnzeke on Nov 27, 2014, 5:30:05 AM
Also the vicious nonsensical replies and biblical walls of text from the hardcore fanboys on these forums hurt my eyes x)
"
in less than a month I had my lvl 66 Marauder.


What a performance


"
Only unaware to the terrible facts I innocently used the dreaded Cyclone build/tactic. I quit playing shortly after reaching slightly high levels since the desync rendered my character completely useless and I'll never stand by a game that makes your personal playstyle unviable due to their shitty servers.


FYI world first Atziri killer in hardcore was a Cyclone build.


"

Give up, POE was a good memory for the overly-short time it lasted in my life.


Adios.
IGN : @Morgoth
"
Morgoth2356 wrote:
"
in less than a month I had my lvl 66 Marauder.


What a performance


"
Only unaware to the terrible facts I innocently used the dreaded Cyclone build/tactic. I quit playing shortly after reaching slightly high levels since the desync rendered my character completely useless and I'll never stand by a game that makes your personal playstyle unviable due to their shitty servers.


FYI world first Atziri killer in hardcore was a Cyclone build.


"

Give up, POE was a good memory for the overly-short time it lasted in my life.


Adios.


Thanks for proving my second comment right after I posted it, love =*
"
Morgoth2356 wrote:
"
in less than a month I had my lvl 66 Marauder.


What a performance


"
Only unaware to the terrible facts I innocently used the dreaded Cyclone build/tactic. I quit playing shortly after reaching slightly high levels since the desync rendered my character completely useless and I'll never stand by a game that makes your personal playstyle unviable due to their shitty servers.


FYI world first Atziri killer in hardcore was a Cyclone build.



The guy didn't actually USE cyclone against Atziri (he used spectral throw).
my evasion is so high i only insta rip sometimes
-----
Bug Fixes:
People were using cyclone for actual melee builds, so we nerfed it and made blade vortex. Also, we went ahead and made cyclone great for CoC casters while we were at it.
"
"
Xarog wrote:
"
mark1030 wrote:
The Shadow already got the revamp that the Duelist is getting (ease of getting out of the starting area). Name one build that is better as a Duelist than any other class. Name one build that is better as a Ranger than any other class. Name one build that's better as a Scion than any other class. Name one build that's better as a Templar than any other class. Name one build that's better as a Marauder than any other class. Nothing you said about a Shadow doesn't also hold true for all other classes with the exception of Witch because that start has more strong nodes than anywhere else. If a shadow start is underpowered, I'd like to see your list of builds proving that all the other classes are better.

You're missing the point. I'm not claiming that the shadow's start is over-powered or under-powered. I'm simply saying that you're doing the equivalent of cherry-picking the data set you want to work with in order to prove your point, but the cherry-picked data doesn't actually hold water. Considering the fact that you were bemoaning OP's behaviour regarding abuse of statistics, your position is somewhat hypocritical.

And it's exactly because I'm on the fence that I'd like to see a good argument for and against the idea that the shadow's starting point is in need of a buff. But so far, niether side has been very convincing.


I have been watching this thread every so often with some amusement because I do see issues with the OP's attitude on what numbers constitute as facts. But I also agree with you that this discussion got rather heated and now there needs to be a collective breath catching. What I would like to see now is OP actually collaborating thoughtfully with others on what would statistically define an under-performing class. If he cannot agree with a general consensus or actually citable sources, his own opinion on that matter means nothing (and so far I have found this an entertaining read because of this). At the same time, the other side was bashing OP too hard to make consensus feasible with OP.

Second, there needs to be actual statistics damnit! If you are all going to claim statistical significance then actually present your results properly. Give me some sexy sexy p <.05 or .01! Even without a definition of what under-performing is, the words relating to any form of statistics should never come up in your emotion-filled argument if you aren't even going to present some friggin' alpha levels you bums!

Psychology major rant over.


The numbers are all there already. This is why I don't understand where you are coming from. We've been through this part a hundred times. I'll re-iterate them for you:

"
EDIT: Let's add some facts as well. Class representation in Beyond was

Class "top-100"% ("top-15k"%)

Witch 43% (33.67%)
Ranger 22% (19.99%)
Scion 9% (9.25%)
Shadow 10% (12.98%)
Templar 9% (6.45%)
Marauder 5% (12.59%)
Duelist 2% (5.07%)


Any class here that has a lower first number and a higher second number is underperforming. IE, they are selected at a certain percentage (marauder for example, selected 12.59% of the time) and drastically drop off in higher brackets (in this case, top 100). This pattern is repeatable in every bracket I've tried. Feel free to find a "top-X" bracket where these classes don't exhibit these patterns.

In this case, witches are overperforming by roughly 10%, marauders underperforming roughly 5%, for a 15% perofrmance difference. For shadows, it's 10 and 3 for a roughly 13% difference. For duelists, a little more than shadow but still roughly 13%.

Also, in terms of mortality, 5 of the top 6 shadows were dead IIRC in the top 60 bracket. Witches? Out of 26 witches in the top 60 12 are dead.
This is the most extreme case, as you move further down the brackets it starts to even out in mortality, but at least in this bracket it's not even close. If you go further up, IE top 40 bracket, it's much more extreme in terms of witches out-surviving shadows.
Would like to see more data on the mortality issue but counting takes a long time and I'm tired of doing it. The data sets, at the very least, are exactly the same way they were back in March. Which my 3 links in the OP illustrate.

Again, there is no opinion here, these are ONLY data points..
my evasion is so high i only insta rip sometimes
-----
Bug Fixes:
People were using cyclone for actual melee builds, so we nerfed it and made blade vortex. Also, we went ahead and made cyclone great for CoC casters while we were at it.
Last edited by Legatus1982 on Nov 27, 2014, 9:58:55 AM
"
Legatus1982 wrote:


The guy didn't actually USE cyclone against Atziri (he used spectral throw).



I'm talking Invasion first kill (hc league that came with Atziri expansion). If you're talking about a permanent league first kill, it doesn't matter since all the gear that was already there trivialized the fight as shit. World first Atziri in Invasion by Havoc, with a Cyclone build.

https://www.youtube.com/watch?v=0AJdVgzVMLo

Welcome back btw.
IGN : @Morgoth
Last edited by Morgoth2356 on Nov 27, 2014, 10:14:11 AM
"
Legatus1982 wrote:
"


I have been watching this thread every so often with some amusement because I do see issues with the OP's attitude on what numbers constitute as facts. But I also agree with you that this discussion got rather heated and now there needs to be a collective breath catching. What I would like to see now is OP actually collaborating thoughtfully with others on what would statistically define an under-performing class. If he cannot agree with a general consensus or actually citable sources, his own opinion on that matter means nothing (and so far I have found this an entertaining read because of this). At the same time, the other side was bashing OP too hard to make consensus feasible with OP.

Second, there needs to be actual statistics damnit! If you are all going to claim statistical significance then actually present your results properly. Give me some sexy sexy p <.05 or .01! Even without a definition of what under-performing is, the words relating to any form of statistics should never come up in your emotion-filled argument if you aren't even going to present some friggin' alpha levels you bums!

Psychology major rant over.


The numbers are all there already. This is why I don't understand where you are coming from. We've been through this part a hundred times. I'll re-iterate them for you:

"
EDIT: Let's add some facts as well. Class representation in Beyond was

Class "top-100"% ("top-15k"%)

Witch 43% (33.67%)
Ranger 22% (19.99%)
Scion 9% (9.25%)
Shadow 10% (12.98%)
Templar 9% (6.45%)
Marauder 5% (12.59%)
Duelist 2% (5.07%)


Any class here that has a lower first number and a higher second number is underperforming. IE, they are selected at a certain percentage (marauder for example, selected 12.59% of the time) and drastically drop off in higher brackets (in this case, top 100). This pattern is repeatable in every bracket I've tried. Feel free to find a "top-X" bracket where these classes don't exhibit these patterns.

In this case, witches are overperforming by roughly 10%, marauders underperforming roughly 5%, for a 15% perofrmance difference. For shadows, it's 10 and 3 for a roughly 13% difference. For duelists, a little more than shadow but still roughly 13%.

Also, in terms of mortality, 5 of the top 6 shadows were dead IIRC in the top 60 bracket. Witches? Out of 26 witches in the top 60 12 are dead.
This is the most extreme case, as you move further down the brackets it starts to even out in mortality, but at least in this bracket it's not even close. If you go further up, IE top 40 bracket, it's much more extreme in terms of witches out-surviving shadows.
Would like to see more data on the mortality issue but counting takes a long time and I'm tired of doing it. The data sets, at the very least, are exactly the same way they were back in March. Which my 3 links in the OP illustrate.

Again, there is no opinion here, these are ONLY data points..


Statistical significance requires you meet an alpha level not percentages. You can claim your percentages mean something all you want but any statistician will ask you the same thing I just did. Yes your data indicates something might be up requiring further statistics but in this case you need to collaborate on what the definition is. If you want me to accept that these percentages are a definition of under-performance then you need something objective to cite. Don't just cite your own claims when these obviously aren't accepted. Back in my PSYC 366 Statistical Analysis II, we discussed a psychologist who was claiming his studies proved existence of supernatural cognitive capabilities. Yes his methods and statistics were sound but his reference for what considered the definition of supernatural cognition was based off his previous studies that nobody else had ever replicated or verified in any fashion.

In a similar matter you have two issues:
1) Stop spouting percentages, they mean absolutely nothing statistically until you give some alpha values. If you want to prove this get to work doing the statistical analyses properly by hand or buy SPSS or something so we can see your work. Until then, stop claiming that the percentages matter, because so far you have not proven they do.
2) Unless you can other people agreeing with your definition, you will continue to be ridiculed and criticized like the above psychologist was and nobody will take you seriously. And to have a proper consensus you need more people who actually sit down and debate the definition thoroughly. You can't just use "I agree with your definition without any statistical reasoning" people.

tl;dr for both of my posts: alpha levels or gtfoos (gtfo of statistics). And there is a legitimate reason why people are criticizing you so harshly.
"It's all clearer now
And I hear her now
And I'm nearer to
The Salvation Code"
"
"
Legatus1982 wrote:
"


I have been watching this thread every so often with some amusement because I do see issues with the OP's attitude on what numbers constitute as facts. But I also agree with you that this discussion got rather heated and now there needs to be a collective breath catching. What I would like to see now is OP actually collaborating thoughtfully with others on what would statistically define an under-performing class. If he cannot agree with a general consensus or actually citable sources, his own opinion on that matter means nothing (and so far I have found this an entertaining read because of this). At the same time, the other side was bashing OP too hard to make consensus feasible with OP.

Second, there needs to be actual statistics damnit! If you are all going to claim statistical significance then actually present your results properly. Give me some sexy sexy p <.05 or .01! Even without a definition of what under-performing is, the words relating to any form of statistics should never come up in your emotion-filled argument if you aren't even going to present some friggin' alpha levels you bums!

Psychology major rant over.


The numbers are all there already. This is why I don't understand where you are coming from. We've been through this part a hundred times. I'll re-iterate them for you:

"
EDIT: Let's add some facts as well. Class representation in Beyond was

Class "top-100"% ("top-15k"%)

Witch 43% (33.67%)
Ranger 22% (19.99%)
Scion 9% (9.25%)
Shadow 10% (12.98%)
Templar 9% (6.45%)
Marauder 5% (12.59%)
Duelist 2% (5.07%)


Any class here that has a lower first number and a higher second number is underperforming. IE, they are selected at a certain percentage (marauder for example, selected 12.59% of the time) and drastically drop off in higher brackets (in this case, top 100). This pattern is repeatable in every bracket I've tried. Feel free to find a "top-X" bracket where these classes don't exhibit these patterns.

In this case, witches are overperforming by roughly 10%, marauders underperforming roughly 5%, for a 15% perofrmance difference. For shadows, it's 10 and 3 for a roughly 13% difference. For duelists, a little more than shadow but still roughly 13%.

Also, in terms of mortality, 5 of the top 6 shadows were dead IIRC in the top 60 bracket. Witches? Out of 26 witches in the top 60 12 are dead.
This is the most extreme case, as you move further down the brackets it starts to even out in mortality, but at least in this bracket it's not even close. If you go further up, IE top 40 bracket, it's much more extreme in terms of witches out-surviving shadows.
Would like to see more data on the mortality issue but counting takes a long time and I'm tired of doing it. The data sets, at the very least, are exactly the same way they were back in March. Which my 3 links in the OP illustrate.

Again, there is no opinion here, these are ONLY data points..


Statistical significance requires you meet an alpha level not percentages. You can claim your percentages mean something all you want but any statistician will ask you the same thing I just did. Yes your data indicates something might be up requiring further statistics but in this case you need to collaborate on what the definition is. If you want me to accept that these percentages are a definition of under-performance then you need something objective to cite. Don't just cite your own claims when these obviously aren't accepted. Back in my PSYC 366 Statistical Analysis II, we discussed a psychologist who was claiming his studies proved existence of supernatural cognitive capabilities. Yes his methods and statistics were sound but his reference for what considered the definition of supernatural cognition was based off his previous studies that nobody else had ever replicated or verified in any fashion.

In a similar matter you have two issues:
1) Stop spouting percentages, they mean absolutely nothing statistically until you give some alpha values. If you want to prove this get to work doing the statistical analyses properly by hand or buy SPSS or something so we can see your work. Until then, stop claiming that the percentages matter, because so far you have not proven they do.
2) Unless you can other people agreeing with your definition, you will continue to be ridiculed and criticized like the above psychologist was and nobody will take you seriously. And to have a proper consensus you need more people who actually sit down and debate the definition thoroughly. You can't just use "I agree with your definition without any statistical reasoning" people.

tl;dr for both of my posts: alpha levels or gtfoos (gtfo of statistics). And there is a legitimate reason why people are criticizing you so harshly.


This would require the OP to be objective about the situation, which his initial post already demonstrates he is not.

This threads feels more like a gangbang on GGG (at least an attempt) rather then a discussion worth spending time on.

I commend both your post's sir.

Instead of digging for the truth, this thread falls into the pitfall of a yes i am right vs no you are wrong mess.

My only regret after reading all these pages is that Charan didn't spend 15 minutes on a response and Xavderion held himself back in his responses.

+1 for mark's post and for PleiadesBlackstar.

Peace,

-Boem-

(op you can bash my opinion if you so desire, but understand it's just an opinion so i wont feel compelled to change it, cheers)
Freedom is not worth having if it does not include the freedom to make mistakes
It actually doesn't require objectiveness on the OP's part. Just a willingness to make a definition on more than just his own subjective opinion. There are plenty studies done with subjective opinions behind them. The difference is that actual statistics are hard to show these biases if the procedure is approved upon by objective others and has a large enough sample.

In this case, OP takes the numbers in the ladder turns them into percentages which yes is data and a fact. The moment OP claims this is under-performance he is making it an opinion in a statistical process and he isn't willing to admit this and then the community is bashing him for it. I'll admit, this is also where I was getting my amusement from. After a while though, the statistics student in me couldn't stay quiet and was becoming insulting to the field of study.

Edit: I forgot to mention that the percentages can then be interpreted in any sort of manner. From my point of view, it is similar to some others voiced statements on here. Many witch and ranger builds can also be run on shadows because the base build uses skill points from 2 or more starting points. It is difficult to claim the shadow area is under-performing on the skill tree when many people are taking witch and ranger and ending with the same builds and these two classes are being separated from shadows in the data.

A more appropriate way to look at this is claim with areas of the passive tree may be leading to builds failing and dying. There are also many other variables besides the skill tree though.
"It's all clearer now
And I hear her now
And I'm nearer to
The Salvation Code"
Last edited by PleiadesBlackstar on Nov 27, 2014, 1:28:25 PM

Report Forum Post

Report Account:

Report Type

Additional Info