426 réponses à ce sujet

[iDeevil]

IanPolaris wrote...

Yes I can.  This may sound rude, but take a class in combinatorial statistics and pass it.  When you do, you'll understand that these are NOT the same things.

Not only does it sound rude, you assume ones level of education is less than yours.  This assumption really does make an ass out of u, as the old saying goes (and unfortunately me by default).

That'a almost begging for a goosed poll.

-Polaris

And so does EVERY internet poll.

You can say whatever you wish, but the truth of the matter is you either have to dismiss both as false, or accept them both as true.

One does not outweigh the other when it comes to polling.  But it does prove that the questions that are asked can determine and change an answer.

[IanPolaris]

Indy_S wrote...

iDeevil wrote...

And again, it is not.  While there is no change to the endings in ME3, they are doing a hell of a lot of market research of wants/needs for DA3 and ME4.  They have learnt from it.

Now the only issue is that I am sure some fans are going to believe all their wants should be included into the game, despite the impracticality of it.

Why do I feel like it's a PR move to not show us any evidence of this? That's entirely contrary to what would appeal to me.

Given what 'Stan' said in another thread (Bioware has no obligation to show it's numbers) [that's my best recollection], I am almost certain Mertzen was caught in a PR move/blunder.

Moral of the story:  Don't go making objective numeric claims unless you are willing to have your bluff called.

-Polaris

[iDeevil]

Indy_S wrote...

iDeevil wrote...

And again, it is not.  While there is no change to the endings in ME3, they are doing a hell of a lot of market research of wants/needs for DA3 and ME4.  They have learnt from it.

Now the only issue is that I am sure some fans are going to believe all their wants should be included into the game, despite the impracticality of it.

Why do I feel like it's a PR move to not show us any evidence of this? That's entirely contrary to what would appeal to me.

The evidence is at this forum.  The evidence is in the call outs they've done in other places.  The evidences are in some fans being asked.

But yes, making a complete fan created game doesn't appeal to me.  But it's also imperative TOO ask fans so you have a temperature of what is and is not wanted/accepted and why.  It doesn't mean they'll use all of it, or any of it.  It just is a way to seat themselves a little better in the ride to finishing the game.

[Clayless]

IanPolaris wrote...

Robosexual wrote...

IanPolaris wrote...

As I explain above, yes I can.

In one case, we have multiple polls that sample thousands of people per survey done by very different organizations some of whom support EA and some that don't.

In the other case we have a bare handful with a few score respondants (at most) tightly controlled by a handful of professional publications that very much have a vested interest in making EA/ME3 look as favorable as possible. 

Very different situations.

-Polaris

It's tinfoil hat logic. The ground you have to dismiss it is the same ground you have to use the ones that agree with you.

No.  It's statistics.  It's a lot easier to skew a poll with a very small sample (especially if the people that do the counting have an agenda) than it is to skew a very large poll over multiple organizations that don't all have a set agenda.

-Polaris

No, this "agenda" isn't a statistic, it's a tinfoil hat wearing delusion.

[Indy_S]

Robosexual wrote...

No, this "agenda" isn't a statistic, it's a tinfoil hat wearing delusion.

I believe that's stalemate. Thankyou, it was a great game to watch, you two.

[IanPolaris]

iDeevil wrote...

IanPolaris wrote...

Yes I can.  This may sound rude, but take a class in combinatorial statistics and pass it.  When you do, you'll understand that these are NOT the same things.

Not only does it sound rude, you assume ones level of education is less than yours.  This assumption really does make an ass out of u, as the old saying goes (and unfortunately me by default).

When it comes to advanced mathematics, your education is vitually certainly less than mine.  It's also not the point.  The point here is you clearly don't have a grasp on combinatorial statistics and how they work.

That'a almost begging for a goosed poll.

-Polaris

And so does EVERY internet poll.

You can say whatever you wish, but the truth of the matter is you either have to dismiss both as false, or accept them both as true.

One does not outweigh the other when it comes to polling.  But it does prove that the questions that are asked can determine and change an answer.

No I don't.  Just because two things can be true does not mean that they are equally likely to be true.

Consider that while the internet polls that 'noob' quotes aren't scientific, he isn't selecting from any one source with any one agenda.  Many of his polls come from pro-ending sources.  

While it is possible for any one of these polls to be 'goosed', what you are suggesting is that about a score of different polls, done over different times, by different organizations, with different editorial positions on ME3 were ALL goosed exactly the same way, to give exactly the same results? 

The odds of that happening are literally astronomical (at least 5 sigmas away from mean given the numbers we have been shown).  I have a better chance of winning the lottery (or at least about as good).

OTOH, if you are talking about a poll of 100 that doesn't have controls (and whose results we don't see...just the final winner...and ME3 didn't win all of these btw...just was nominated), as little as 5-10 false/repeat data points can completely skew the poll (which is less than a single sigma).

That is why you can't equate the two.

-Polaris

[IanPolaris]

Robosexual wrote...

No, this "agenda" isn't a statistic, it's a tinfoil hat wearing delusion.

I never claimed it was.  What I said was that a poll of 100 (or even less) is a lot easier to goose than a poll of thousands esp if both are unscientific polls.

-Polaris

[iDeevil]

IanPolaris wrote...

When it comes to advanced mathematics, your education is vitually certainly less than mine.  It's also not the point.  The point here is you clearly don't have a grasp on combinatorial statistics and how they work.

And your grasp of psychology of the fan and it's relation to the internet is significantly less than mine.  Both of these are important to what we are discussing.  Yours is no more important than mine.

No I don't.  Just because two things can be true does not mean that they are equally likely to be true.

In this case yes they do.  Their methods, and the people they are appealing too are both the same.  The ability to 'goose' a poll in 1 place, it the same ability if one wishes to do it in multiple.  It's all dependent on want and wherewithall.  In the height of the Retaker movement, the passionate responses people had lead them to not only express theor anger here, but in every possible place.

This anger often leads to people acting out in ways to get heard, and one of them is in internet polls.  The larger the poll, the more validated the fan feels.  Anger and negative feelings almost always lead to a greater response then a neutral or happy one.

That is why you can't equate the two.

-Polaris

The above is why I do equate the 2.  The math here doesn't exist in a bubble simply because the internet allows for people to respond to a poll many times.  It's not like an election.

As I said, the fact that people both hate the ending and love the game show how different questions and their wordings can elicit different responses.

Modifié par iDeevil, 03 mars 2013 - 12:59 .

[ElitePinecone]

FlamingBoy wrote...

its called the sampling method, the 75k is a representative of a whole.

thats how stats works

FFS, no. That's not how stats work, and it's ridiculous and depressing that people here think so.

That 75k represents people obsessed enough about the game to bother responding to a poll. They're also self-selecting, in that people with a motivation to vote a certain way or have their voice heard.

The ONLY statistically useful method to be able to generalise to a whole is random sampling, or picking enough fans at random to be sure that mathematically the result gained was fairly valid. Since nobody has ever done that, and selecting random fans needs a) a list of everyone who bought/finished the game their contact information and c) a way to contact them to ask questions, it's probably never going to happen.

Quoting internet polls does your cause, whatever it is, no help. It's amateur, completely incorrect and reduces the credibility of any point you're trying to make. 

[Wulfram]

Self selecting polls are all pretty useless really.

edit: ^ What he said.

Modifié par Wulfram, 03 mars 2013 - 01:05 .

[IanPolaris]

iDeevil wrote...

IanPolaris wrote...

When it comes to advanced mathematics, your education is vitually certainly less than mine.  It's also not the point.  The point here is you clearly don't have a grasp on combinatorial statistics and how they work.

And your grasp of psychology of the fan and it's relation to the internet is significantly less than mine.  Both of these are important to what we are discussing.  Yours is no more important than mine.

In the main, it's not.  For this particular issue it is.  You can not compare with the same weight small polls from a single poll that is unscientific with a possible agenda with many polls (admittedly all unscientific) in the thousands over many different organizations.

Can both be goosed?  Sure.  However the odds are nowhere close to equal.

No I don't.  Just because two things can be true does not mean that they are equally likely to be true.

In this case yes they do.  Their methods, and the people they are appealing too are both the same.  The ability to 'goose' a poll in 1 place, it the same ability if one wishes to do it in multiple.  It's all dependent on want and wherewithall.  In the height of the Retaker movement, the passionate responses people had lead them to not only express theor anger here, but in every possible place.

This anger often leads to people acting out in ways to get heard, and one of them is in internet polls.  The larger the poll, the more validated the fan feels.  Anger and negative feelings almost always lead to a greater response then a neutral or happy one.

The sheer logistics of the task make this a near impossibility.

That is why you can't equate the two.

-Polaris

The above is why I do equate the 2.  The math here doesn't exist in a bubble simply because the internet allows for people to respond to a poll many times.  It's not like an election.

As I said, the fact that people both hate the ending and love the game show how different questions and their wordings can elicit different responses.

The problem is you had many PRO-ending organiazations that also did surveys.  Given that you had multiple organizations collecting data with different biases, with different methods, but all come up with the same result lead to a very strong conclusion that the results are probably approximately valid.

As I have said many times before, Bioware could put this to bed immediately.  A SINGLE true scientific poll (and could be as few as 500 people if the sampling were done correctly) could answer the question.  Furthermore, this is data that one would think Bioware would want, and a single 500 person poll done correctly isn't all that expensive and there are many very good and reputable pollsters that could do this for very reasonable rates.

So the question becomes why hasn't bioware done this?  For all the faults of his data, at least noob presents data.

-Polaris

[iDeevil]

IanPolaris wrote...

In the main, it's not.  For this particular issue it is.  You can not compare with the same weight small polls from a single poll that is unscientific with a possible agenda with many polls (admittedly all unscientific) in the thousands over many different organizations.

Can both be goosed?  Sure.  However the odds are nowhere close to equal.

In this case I argue that they are not only close to equal, they are equal.

The sheer logistics of the task make this a near impossibility.

No it doesn't.  I have both witnessed it and heavily researched it.  It's not only possible, it does happen.  Pure math in this case doesn't account for the complete and utter ridiculousness of the interenet, and those on it.

The problem is you had many PRO-ending organiazations that also did surveys.  Given that you had multiple organizations collecting data with different biases, with different methods, but all come up with the same result lead to a very strong conclusion that the results are probably approximately valid.

And the fact that ME3 won more than one award this season also proves that the game was well liked.  Both of these can be true.

So the question becomes why hasn't bioware done this?  For all the faults of his data, at least noob presents data.

-Polaris

Because Bioware had no intention of changing the ending, either because they could not afford it or they liked the idea of it.

Also at this point it'll prove nothing other than some of us are still hanging around here a year later arguing the same or at best similar point coming up with the same result.

We really do define insanity.

[IanPolaris]

ElitePinecone wrote...

FlamingBoy wrote...

its called the sampling method, the 75k is a representative of a whole.

thats how stats works

FFS, no. That's not how stats work, and it's ridiculous and depressing that people here think so.

That 75k represents people obsessed enough about the game to bother responding to a poll. They're also self-selecting, in that people with a motivation to vote a certain way or have their voice heard.

This is very true which is why I agree that all the polls that noob has presented are unscientific.  That said, when you see multiple polls of this sort done by different organizations sampling different people and using different methedologies, but they all come up to the same result, the reasonable conclusion is that there is "probably" a real effect going on.

The logical next step would be to conduct a scientific survey as you discuss below.

The ONLY statistically useful method to be able to generalise to a whole is random sampling, or picking enough fans at random to be sure that mathematically the result gained was fairly valid. Since nobody has ever done that, and selecting random fans needs a) a list of everyone who bought/finished the game their contact information and c) a way to contact them to ask questions, it's probably never going to happen.

Actually you need to use controlled random sampling because otherwise you risk getting outlier samples.  This is why political polls (for example) will weight the raw random samples to census data for gender and ethnicity (for example).

However, if you do this correctly, then a poll (certainly for the population that's completed ME3) of as little as 500 people would answer this question pretty much for good.  Such a poll wouldn't cost all that much, and given that Origins gives a very good "in" for independant data collection, it wouldn't be all that difficult for an independant research firm to do.

So why hasn't Bioware done this?

Quoting internet polls does your cause, whatever it is, no help. It's amateur, completely incorrect and reduces the credibility of any point you're trying to make. 

I agree up to a point.  However, when you have multiple internet polls from varying sources that all say the same thing, the chances that there is a real effect becomes a near certainty.  To quantify it, you would of course want to do a scientific survey.

-Polaris

[zyntifox]

I am probably the only statistician on these boards, unfortunately there aren't many of us. And i agree that there is no statistical foundation to use just one of these polls due to self-selection bias. However, and this is something i have been researching quite a lot, you can extract information out of several biased polls if the question in the polls are approximately on equal footing.

And given the sheer amount of polls regarding the ending to mass effect i could easily extract useful data from them that you can use to draw statistical conclusions of the "average" player's opinion of the ending.

[IanPolaris]

iDeevil wrote...

And the fact that ME3 won more than one award this season also proves that the game was well liked.  Both of these can be true.

If proves the game was well liked by the professional critics.  This is a point that no one contests.

So the question becomes why hasn't bioware done this?  For all the faults of his data, at least noob presents data.

-Polaris

Because Bioware had no intention of changing the ending, either because they could not afford it or they liked the idea of it.

Also at this point it'll prove nothing other than some of us are still hanging around here a year later arguing the same or at best similar point coming up with the same result.

We really do define insanity.

Not doing such a poll is insanity whether Bioware ever intended to change the endings or not.  Well constructed scientific polls like what I am suggesting would not only help answer if the fans liked or didn't like the endings but why.  This is valuable market research data into making sure that mistakes aren't repeated in the future.  We already know that Bioware does in fact do such research because early DA3 scientific surveys did get leaked.

More to the point here, a bioware employee should not be claiming something as an objective fact (even on Twitter) unless said employee is willing to prove it.

-Polaris

[carlosjulio]

So there's people actually defending the validity of these self-selecting (not randomly selected from the fanbase) polls. Unbelievable. I myself voted in every poll I could find online when the original ending came out saying I hated it, many others did the same in the hope that bioware would check these kind of online polls to justify making dlc to fix it. Let me repeat: It's the same people googling for these polls and voting over and over. You want to talk statistics? How about poll design? This is a joke, as the guy who makes this video seem to be.

Modifié par carlosjulio, 03 mars 2013 - 01:27 .

[Dr_Extrem]

Cstaf wrote...

I am probably the only statistician on these boards, unfortunately there aren't many of us. And i agree that there is no statistical foundation to use just one of these polls due to self-selection bias. However, and this is something i have been researching quite a lot, you can extract information out of several biased polls if the question in the polls are approximately on equal footing.

And given the sheer amount of polls regarding the ending to mass effect i could easily extract useful data from them that you can use to draw statistical conclusions of the "average" player's opinion of the ending.

so .. the sheer number of relativley equel polls, the number of participants and the absents of outliers, could result in a relativly "save" assumption, that the endings were not recieved well?

[IanPolaris]

carlosjulio wrote...

So there's people actually defending the validity of these self-selecting (not randomly selected from the fanbase) polls. Unbelievable. I myself voted in every poll I could find online when the original ending came out saying I hated it, many others did the same in the hope that bioware would check these kind of online poll to justify making dlc to fix it. Let me repeat: It's the same people googling for these polls and voting over and over. You want to talk statistics? How about poll design? This is a joke, as the guy who makes this video seem to be.

The key here is that we have at least 10 different polls, done differently, by different organizations, etc.  Any single poll of this sort is scientifically invalid.  No one questions that.  However, useful data and a first order approximation of a real effect of the larger population can be gleaned from analyzing multiple polls of this type.

It is as good as even one true scientific poll?

Not even close.

However, it's the best we have, and at least 'noob' is presenting data.  Bioware is not.

-Polaris

[IanPolaris]

Dr_Extrem wrote...

Cstaf wrote...

I am probably the only statistician on these boards, unfortunately there aren't many of us. And i agree that there is no statistical foundation to use just one of these polls due to self-selection bias. However, and this is something i have been researching quite a lot, you can extract information out of several biased polls if the question in the polls are approximately on equal footing.

And given the sheer amount of polls regarding the ending to mass effect i could easily extract useful data from them that you can use to draw statistical conclusions of the "average" player's opinion of the ending.

so .. the sheer number of relativley equel polls, the number of participants and the absents of outliers, could result in a relativly "save" assumption, that the endings were not recieved well?

That's what I'm taking away from this.

-Polaris

[iDeevil]

[quote]IanPolaris wrote...

If proves the game was well liked by the professional critics.  This is a point that no one contests.[/quote]

Given that it won fan awards also proves that it was liked better, in realtion to other games.  That in itself could be used as proof as well.


[quote]Not doing such a poll is insanity whether Bioware ever intended to change the endings or not.  Well constructed scientific polls like what I am suggesting would not only help answer if the fans liked or didn't like the endings but why.  This is valuable market research data into making sure that mistakes aren't repeated in the future.  We already know that Bioware does in fact do such research because early DA3 scientific surveys did get leaked.[/quote]

And we don't know they didn't do that.  Your ascertian they didn't based on 'we didn't hear' could be true.  But not knowing where, and from who and how they could sample means that folks may not have mentioned it because they aren't in these places to do so.

We're all making assumptions, where based on maths, science, psychology or simple opinion.  We just do not know anything.

More to the point here, a bioware employee should not be claiming something as an objective fact (even on Twitter) unless said employee is willing to prove it.

-Polaris[/quote]

Welcome to the job.  We've all had jobs where we have had to say untruths, half truths or placade people.  That's community leaders jobs.  Cut her, personally, a little slack.  She does what she is told and gives information the she to is told.  She doesn't make it up.

Really, at this point it doesn't even matter.  People hate the ending.  People like the ending.  People hate the game.  People like the game.  How people feel now about it is likely different than how people felt 6 months ago.

All of this pontificating proves that ****** on a bull have more purpose then this ever will.

[carlosjulio]

IanPolaris wrote...

carlosjulio wrote...

So there's people actually defending the validity of these self-selecting (not randomly selected from the fanbase) polls. Unbelievable. I myself voted in every poll I could find online when the original ending came out saying I hated it, many others did the same in the hope that bioware would check these kind of online poll to justify making dlc to fix it. Let me repeat: It's the same people googling for these polls and voting over and over. You want to talk statistics? How about poll design? This is a joke, as the guy who makes this video seem to be.

The key here is that we have at least 10 different polls, done differently, by different organizations, etc.  Any single poll of this sort is scientifically invalid.  No one questions that.  However, useful data and a first order approximation of a real effect of the larger population can be gleaned from analyzing multiple polls of this type.

It is as good as even one true scientific poll?

Not even close.

However, it's the best we have, and at least 'noob' is presenting data.  Bioware is not.

-Polaris

As I said, to get any useful data, you have to prove it's not the same people voting over and over in evey poll. They may be different organizations, but it's the same google.

Modifié par carlosjulio, 03 mars 2013 - 01:31 .

[ElitePinecone]

IanPolaris wrote...

So why hasn't Bioware done this?

Becaue they don't feel like indulging the fantastical delusions of a few hundred people on their internet forum? 

Or maybe because contacting people through Origin to ask about a year-old video game for a scientific poll is a) possibly illegal, incredibly creepy, and c) completely unnecessary. Why would Bioware even do that? They got the message last time, they certainly don't need numbers. So what if they can't prove their point to the miniscule proportion of people who are still hanging around on the BSN? What's it to them if you don't believe the numbers they quoted?

Quoting internet polls does your cause, whatever it is, no help. It's amateur, completely incorrect and reduces the credibility of any point you're trying to make. 

I agree up to a point.  However, when you have multiple internet polls from varying sources that all say the same thing, the chances that there is a real effect becomes a near certainty.  To quantify it, you would of course want to do a scientific survey.

-Polaris

No. Multiple polls, all unscientfic, don't somehow acquire legitimacy or validity just because there are more of them. Every single one has no sampling validity. Self-selection biases make them irrelevant for generalising about the trend in the wider population, and I won't even get started on the methodology of the questions. 

All that internet polls can tell you is the breakdown of responses by the people who voted for them. Generalising from that to the general population of Mass Effect players, or "fans" as some kind of homogenous hive-mind, is not only preposterous, it's dangerously deluded.

Was there a comparatively large number of people who didn't like the endings, and who voted that way on polls they chose to participate in? Yes, it seems as if there was, based on intenet polls where highly-motivated people chose to respond. 

What does that tell us about the playerbase as a whole? Nothing. A million other people (two million? three?) could've loved the game, or felt indifferent, or not completed it, and we have no idea if they did. To presume for them does your cause an enormous disservice, and it makes me uncomfortable to deeply dislike the ending too when people I'd normally agree with use this statistical nonsense to back up their claims. 

[iDeevil]

carlosjulio wrote...

So there's people actually defending the validity of these self-selecting (not randomly selected from the fanbase) polls. Unbelievable. I myself voted in every poll I could find online when the original ending came out saying I hated it, many others did the same in the hope that bioware would check these kind of online polls to justify making dlc to fix it. Let me repeat: It's the same people googling for these polls and voting over and over. You want to talk statistics? How about poll design? This is a joke, as the guy who makes this video seem to be.

Thank you for admitting to this.  I understand why you did it, and was glad in the end because the EC was nice in a number of ways.

[IanPolaris]

iDeevil wrote...

IanPolaris wrote...

If proves the game was well liked by the professional critics.  This is a point that no one contests.

Given that it won fan awards also proves that it was liked better, in realtion to other games.  That in itself could be used as proof as well.

It was nominated.  There is a difference.  Also such polls are very small, unsicentific, and secret, and done by organizations that have an agenda.  Colour me suspicious.

Not doing such a poll is insanity whether Bioware ever intended to change the endings or not.  Well constructed scientific polls like what I am suggesting would not only help answer if the fans liked or didn't like the endings but why.  This is valuable market research data into making sure that mistakes aren't repeated in the future.  We already know that Bioware does in fact do such research because early DA3 scientific surveys did get leaked.

And we don't know they didn't do that.  Your ascertian they didn't based on 'we didn't hear' could be true.  But not knowing where, and from who and how they could sample means that folks may not have mentioned it because they aren't in these places to do so.

We're all making assumptions, where based on maths, science, psychology or simple opinion.  We just do not know anything.

We have absolutely no evidence that they did, and Bioware has refused to release any numbers when even one (cheap) scientific survey could have put this to bed immediately.

More to the point here, a bioware employee should not be claiming something as an objective fact (even on Twitter) unless said employee is willing to prove it.

-Polaris

Welcome to the job.  We've all had jobs where we have had to say untruths, half truths or placade people.  That's community leaders jobs.  Cut her, personally, a little slack.  She does what she is told and gives information the she to is told.  She doesn't make it up.

No can do.  She could have said something to the effect, "We are proceding with the new DLC based on what we think is best and will be best received."  She did NOT have to make the specific and objective claim that she did.  I cut her no slack for this.

Really, at this point it doesn't even matter.  People hate the ending.  People like the ending.  People hate the game.  People like the game.  How people feel now about it is likely different than how people felt 6 months ago.

All of this pontificating proves that ****** on a bull have more purpose then this ever will.

This wouldn't have come up except it seems to me (and apparently a lot of other people) like Bioware (or at least a Bioware rep) was caught making stuff up and was caught red handed.

-Polaris

[Indy_S]

Once again, it does not matter whether or not it a 'majority' is at work here. It remains a significant part of the consumer base and BioWare should not dismiss it. (Not saying that they are, just that they look an awful lot like they are)