Still no balance a year later - with facts

"

PleiadesBlackstar wrote:

"

Legatus1982 wrote:

"

And now that you edited your post I have to add. I don't have a strong position. From my statistical background I am presuming your claim is likely wrong. There are too many variables that you didn't or maybe cannot due to time and energy constraints account for. There is reasonable chance your claim is correct and I am not denying that. I am simply referring to the basics of statistics that your data is lacking despite claiming your argument is the only interpretation.

In your own post you demonstrated the problem. There is reasonable chance my claim is correct, you said yourself. But you also presume it's not. Why? What have you seen to make you think otherwise? It is not a reasonable position to think that with nothing showing you otherwise. You are allowed to be skeptical and desire more data, but denying the claim with no evidence otherwise? Not a reasonable position in the least. Even without conclusive data, all the data we DO have shows I'm right. Therefore you have no ground to stand on saying I'm wrong, even if you think I might not be right.

And yet look what people are saying. It's not normal unless it's obfuscation.

1)There is no account for the fact that was pointed to earlier that the "class" is fairly interchangeable. Your argument shouldn't be solely shadows but any builds that use the shadow area with this interchangeability.

2) There is no account for differences in players skills. The people who are getting into the top 40 as shadows may possibly be more reckless or new to being the top and dying as a result.

3) Other variables are not considered (people falling asleep trying to rush to the top for example).

These issues would effect any class chosen not just shadow. The point I am making is that you need to eliminate these variables to claim that it is the shadow nodes in particular that are the issue especially 1. These are my reasons for presuming incorrect opinion. Yes the percentages show shadows dying more often, does that mean that the shadow class aka the nodes and area is under-performing? You have not given reasonable evidence to suppose the second answer is yes. Only that shadows have died more. Shadows dying more =/= the shadow's nodes are weaker than any other class unless you control for the above variables and do the analyses.

Now you're the one sounding like a new guy to the world of statistics. Do you assume that all shadow players are less skilled than Witch players? All shadows fall asleep while playing? That isn't how stats work and as someone about to get a degree I'm certain you know that. You are most definitely now making excuses.

Basically, the counterpoint for your argument is that I could say the same for witches, and you're still discounting my argument without any evidence to the contrary and all evidence backing me. Not a reasonable position even if you think I might not have enough data.

And we've already been over the starting point/class point, it doesn't hold up.

"

Legatus1982 wrote:

"

PleiadesBlackstar wrote:

1)There is no account for the fact that was pointed to earlier that the "class" is fairly interchangeable. Your argument shouldn't be solely shadows but any builds that use the shadow area with this interchangeability.

2) There is no account for differences in players skills. The people who are getting into the top 40 as shadows may possibly be more reckless or new to being the top and dying as a result.

3) Other variables are not considered (people falling asleep trying to rush to the top for example).

These issues would effect any class chosen not just shadow. The point I am making is that you need to eliminate these variables to claim that it is the shadow nodes in particular that are the issue especially 1. These are my reasons for presuming incorrect opinion. Yes the percentages show shadows dying more often, does that mean that the shadow class aka the nodes and area is under-performing? You have not given reasonable evidence to suppose the second answer is yes. Only that shadows have died more. Shadows dying more =/= the shadow's nodes are weaker than any other class unless you control for the above variables and do the analyses.

Now you're the one sounding like a new guy to the world of statistics. Do you assume that all shadow players are less skilled than Witch players? All shadows fall asleep while playing? That isn't how stats work and as someone about to get a degree I'm certain you know that. You are most definitely now making excuses.

Basically, the counterpoint for your argument is that I could say the same for witches, and you're still discounting my argument without any evidence to the contrary and all evidence backing me. Not a reasonable position even if you think I might not have enough data.

And we've already been over the starting point/class point, it doesn't hold up.

Answered you in the bold. And even despite these factors which as soon as I posted I realized and will easily admit the last 2 points are not very likely. People marathoning and falling asleep is common enough, I don't think a factor like that can be debated. Are all of these people shadows or even playing POE? No but it happens and more often then just the popular streamers passing out on stream. Does it account for all deaths regardless of class. Again no.

Next, I will address skill. Not all reckless people will play shadow. Will reckless people play shadow. Definitely. It will happen. Will people continue to be reckless as they reach the top spots, less likely but as another post suggested, people being grossly over leveled can easily become reckless again and die to a low level strongbox or something along those lines. Again, I am not saying these things happen 100% of the time though you seem to suggest that I am.

So now you still have not addressed interchangeability of classes.
http://www.pathofexile.com/passive-skill-tree/AAAAAgAAOyi-vAOWguRuPZARYEOKr4MJSshvnm6qnKTKSvGztfKHE8M6S1c1kncHm7XA49i9GYpTNSLq_gotgw-r6Na3PpcGbWwWv-sUBbXGrsSiNj0mlY2_1CM8Lbvj44SdquywYeJRRzBxxPaX_v66DkhirGyMjAtVS-dUMHzdqCftfr19dV3yEZa5fIw2JP1_K419S3j8S6yq035w1ZAb1ooj9jpCbBZyqVnzAF6J03BWTZKHGfcymG_AGhmOcFKaE-oYkn2QVe2DKgvb1DBU5c8ILvPqPT7cIz-b

This is my passive tree for my Scion currently. Look how many points it would take to start as a Ranger. If we pretend other classes could cut through Scion then 1. But we can't so 5. And those evasion nodes might be better for me in the long run because I am currently having mana issues because of overleveling clarity filling my mana and attacking too fast. In a similar vein, most popular or active builds typically use other non-Scion starting nodes.
http://www.pathofexile.com/passive-skill-tree/AAAAAgYBAQkBbwJxBAcEswceDF8NfA48DkgRLxFQEZYUcRXXFf0WvxslHKcc3B1PHwIkiygqKPorCiy_LR8u_zW5NsU8LT0PQYdF9EcGSVFLrkyzTdhQMFBQUdlTpVW1VmNXyVxrXfJfP19qY0Nkr2ebaHRtGW9XcFJw1Xy7fr1_xoFvgseESIauidOKOIt6jHaTJ5SglSCVLpf-mZqbap48ns2fy6C0og-io6LZpn-nK6cwsJKw2LGztAy0xbVItve3PrjZwQTBxcHzw23EFcT2xq7ZW9pi2-fdSOL35FHpAuq662PtQfDV8uHzBvfB-ej7qv-T

And older posts here have suggested the same thing. So your "evidence" showing that shadows are unbalanced and death-prone to other classes, namely witch and ranger is rather unimportant. You would need to find and generalize every (presumably popular) passive tree stats in terms of survivability and compare them to other builds in the same way that never come close to using shadow passives (marauder and such). Then you need to show statistical significance. If it is significant then what likely would have to happen is a re-balance of passives that are not performing adequately, not just the shadow area, and this is always occurring in the meta. The shadows that are dead, no matter how large you make the population, will only mean those who are forced to start in that area by their play choice are dead, not that their character was under-performing. If somebody plays the exact same tree with a witch or ranger of the now dead shadow, and that witch or ranger is alive, it completely nullifies that shadows death as an under-performing. The death can be attributed to many other things (gear, lag, desync, player errors).

It makes no difference that I do not have "evidence" the issue is that yours is not actually available to use because you are ignoring all these possible conflicting variables. If Pavlov had rung a bell, slapped the table, shot a gun, blew up the next building over every time he brought the dog's food out, then tested the conditioning without the food exactly the same and he claimed that the bell was the only thing that caused the drooling, he would have never been associated with classical conditioning as he assumed that the one reason he wanted to be the cause was the cause without any testing. This is why people have ridiculed you. This is why people will not agree with you. This is why I am asking for you to open your mind to the many possible reasons as to why shadows are dead, not just your subjective opinion when looking at the amount of deaths.


I am done now. If you still can't acknowledge the realm of possibilities then... I dunno. I guess that is your problem and why you will be pissed every update there is no changes to "the shadow class." If any of this were suggested to be negligible I would be on your side but you provide nothing to convince the general audience of that.

It's pretty clear that I'm willing to accept possibility and you are not.

Of course, it is a possibility that shadows are being piloted more poorly or falling asleep more often.

But then, you have 0 evidence to support that view. My evidence is not rock solid, but it is unquestionably more than the opposition has.

So again, regardless of why the peons are giving me shit, what supports their view? Nothing. It's less evidence than I have thus I have the stronger case. Is it a lock? No, it's just stronger of a position than the opposition.

Thus, I can only assume with so many people telling me I'm wrong that it is a simple case of popular stupidity. They have no case currently, mine is not great but it's sure as hell better than theirs isn't it? So knowing that you're going to sit here with a straight face and tell me I'm wrong with all the evidence for me and against you? Not a reasonable position.

"

Legatus1982 wrote:

Thus, I can only assume with so many people telling me I'm wrong that it is a simple case of popular stupidity.

And that says it all.

"

Morgoth2356 wrote:

"

Legatus1982 wrote:

Thus, I can only assume with so many people telling me I'm wrong that it is a simple case of popular stupidity.

And that says it all.

I am more than willing to start taking your posts out of context as well, if that's where you want to go with this. It's pretty clear you have nothing anyways so there's no need to, but could be fun so why not.

"

Legatus1982 wrote:

It's pretty clear that I'm willing to accept possibility and you are not.

Of course, it is a possibility that shadows are being piloted more poorly or falling asleep more often.

But then, you have 0 evidence to support that view. My evidence is not rock solid, but it is unquestionably more than the opposition has.

So again, regardless of why the peons are giving me shit, what supports their view? Nothing. It's less evidence than I have thus I have the stronger case. Is it a lock? No, it's just stronger of a position than the opposition.

Thus, I can only assume with so many people telling me I'm wrong that it is a simple case of popular stupidity. They have no case currently, mine is not great but it's sure as hell better than theirs isn't it? So knowing that you're going to sit here with a straight face and tell me I'm wrong with all the evidence for me and against you? Not a reasonable position.

If you have done statistics your entire life then you know the null hypothesis is to assume the is no relationship between your variables. The position everybody opposing you has is the damn null hypothesis that there so no correlation between shadow deaths and passive tree performance. There is no evidence needed to support the null hypothesis to start the study. That is why it is a general statement to be assumed. If you are going to claim you have evidence to the contrary, you better put your statistical background where your mouth is and give evidence contrary to the null hypothesis. I.e. you need to show that all those variables I listed have no statistical impact on these tests and the very fact the shadows are dead is correlated with the weakness of the shadow area nodes. You have not, repeat, have not done that and refuse to because you are convinced your percentage data somehow fits these criteria. This is obvious to the point that the general audience has not been convinced because all these factors are present. You don't need a statistics background to assume these factors can conflict with your assumption. So far you claim you are open to these possibilities while also maintaining that you have proof of the only answer without any statistical significance to speak of. It is infuriating how much statistical knowledge you are failing to express even though it is supposedly important for your career.

"

PleiadesBlackstar wrote:

"

Legatus1982 wrote:

It's pretty clear that I'm willing to accept possibility and you are not.

Of course, it is a possibility that shadows are being piloted more poorly or falling asleep more often.

But then, you have 0 evidence to support that view. My evidence is not rock solid, but it is unquestionably more than the opposition has.

So again, regardless of why the peons are giving me shit, what supports their view? Nothing. It's less evidence than I have thus I have the stronger case. Is it a lock? No, it's just stronger of a position than the opposition.

Thus, I can only assume with so many people telling me I'm wrong that it is a simple case of popular stupidity. They have no case currently, mine is not great but it's sure as hell better than theirs isn't it? So knowing that you're going to sit here with a straight face and tell me I'm wrong with all the evidence for me and against you? Not a reasonable position.

If you have done statistics your entire life then you know the null hypothesis is to assume the is no relationship between your variables. The position everybody opposing you has is the damn null hypothesis that there so no correlation between shadow deaths and passive tree performance. There is no evidence needed to support the null hypothesis to start the study. That is why it is a general statement to be assumed. If you are going to claim you have evidence to the contrary, you better put your statistical background where your mouth is and give evidence contrary to the null hypothesis. I.e. you need to show that all those variables I listed have no statistical impact on these tests and the very fact the shadows are dead is correlated with the weakness of the shadow area nodes. You have not, repeat, have not done that and refuse to because you are convinced your percentage data somehow fits these criteria. This is obvious to the point that the general audience has not been convinced because all these factors are present. You don't need a statistics background to assume these factors can conflict with your assumption. So far you claim you are open to these possibilities while also maintaining that you have proof of the only answer without any statistical significance to speak of. It is infuriating how much statistical knowledge you are failing to express even though it is supposedly important for your career.

Are you going to pay a major company millions to do a real study on POE players while they play all the classes with all the possible millions of build choices? No you're not? THEN YOU ARE WRONG CUZ NULL HYPOTHESIS HURR DURR

Nobody can do this, even ggg probably can't afford to so don't act like the impetus is on me to produce this data. are you even serious?

More importantly for the 5th time your null hypothesis is not an argument. There is no evidence leading to the null hypothesis at all and there is literally nothing that could possibly account for every variable even in a real study. You are being totally ridiculous.

The bottom line is I have supporting data and nobody else does, null hypothesis or not I have produced more data than 99% of the people who have posted on this forum and definitely more than anyone on this thread. You can't just dismiss the possibility on the argument of null hypothesis, with the data here it's not even reasonable to assume that position.

"

Legatus1982 wrote:

The expected value is that they perform at the level of their selection.

Hey! You finally came up with a hypothesis. Now you just need to take the next step and actually test your hypothesis.

I'd suggest for brevity (though perhaps less amusement) that you enlist the help of someone who understands the myriad factors that may affect exile selection beyond 'dur it die a lot i dont like' to inform your analysis of whether your hypothesis is correct.

I say exile selection because I think using the term 'class' is misleading at best.

"

davidnn5 wrote:

"

Legatus1982 wrote:

The expected value is that they perform at the level of their selection.

Hey! You finally came up with a hypothesis. Now you just need to take the next step and actually test your hypothesis.

I'd suggest for brevity (though perhaps less amusement) that you enlist the help of someone who understands the myriad factors that may affect exile selection beyond 'dur it die a lot i dont like' to inform your analysis of whether your hypothesis is correct.

I say exile selection because I think using the term 'class' is misleading at best.

Spoiler

And in all honesty I know I'm just pouring petrol on a fire, but this thread really has been kinda amusing and I don't honestly believe you're as pig-headed as you're trolling to appear ;)

I actually have a theory that you're attempting to prove Charan's signature from some time ago: if you happen to interact with an arsehole during your day, that sucks; but if everyone you meet on a given day is an arsehole, it's more likely you're the arsehole.

Class performance is not a hypothesis. Nobody has to test to find an expected value for heads or tails on a coinflip, you already know it's 50/50. in this case there are 7 choices so the expected value is that they account for 1/7 of the positions on the ladder. This is not a hypothesis, it's the facts.
In this case, the overall class distribution is a factor since the classes are not selected at a real rate of 1/7. That's what the 15k class distribution stat is for, so we know the real distribution relative to the bracket performance. It's all there in the data already given on this thread.

In terms of "the arsehole", the only "arsehole's" I deal with are pretty much on this forum. AKA the ever-obvious GGG fanboy who mostly participates in obfuscation tactics more than real debate, sense they typically immediately lose any attempt at actual debate.

This thread makes me sad. It's like ignorance incarnate.

All i get when i read this thread is like 8 people sitting around a round object with one dude saying "ITS A RED BALL", sadly for him, the ball has different colors depending on the position you are looking at it.

And instead of having a conversation so they can actually discover the different colors from different sides it's just,

RED BALL
RED BALL
RED BALL

over and over again.

Peace,

-Boem-



(Feel free to call me a fanboy simply because i disagree with your stance if that makes you feel any better, it will make some people laugh who know me. and i'm always up for spreading the joy.)