426 réponses à ce sujet

[IanPolaris]

Cstaf wrote...

I have not seen the video. I have a habit of not watching an video of someone trying to make statistical inference who is not a statistician. It usually never ends well.

  Point well taken.  I asked because whatever one thinks of noob (and I have my doubts about his harping on IT in particular myself), he does IMO do a decent job at sourcing his data.

But to make it clear; it is not enough if you have 10, 100 or 1000 polls that says one thing to claim that to be the truth. You need to extract the bias of the aggregate of the polls prior to make an inference. And it is not certain after clearing the data of bias that you can draw a statistical conclussion of it given the relatively high standard error of the estimate this method produces.

Of course, you have to dig into the mechanics of the polls themselves (or the internals I've often heard it referred to) and your caution about error is well taken.  It's why I've tried to be very careful with my language.  I feel that the data is enough to make a reasonable inference that more people (a year later) don't like the enders vs those that do.  That's as far as I'm prepared to go.

I do wish a rigorous poll had been done.

-Polaris

[Indy_S]

iDeevil wrote...

She didn't have to go out on a limb and make an objective claim she wasn't prepared to back up.

She also didn't have to back it up, despite what you may think.

I disagree with this statement, not from a technical standpoint but from an ethical one. If you are willing to dismiss an argument when you have a position of authority, you back up your statements. Even if it requires bending the truth, that is preferable to a dismissal without just cause.

[iDeevil]

Cstaf wrote...
I have not seen the video. I have a habit of not watching an video of someone trying to make statistical inference who is not a statistician. It usually never ends well.
But to make it clear; it is not enough if you have 10, 100 or 1000 polls that says one thing to claim that to be the truth. You need to extract the bias of the aggregate of the polls prior to make an inference. And it is not certain after clearing the data of bias that you can draw a statistical conclussion of it given the relatively high standard error of the estimate this method produces.

So basically what you are saying is that it can produce a possibility of what the population estimates?
And/Or are you saying that standing alone, even with bias extracted, that these polls cannot be evidentary of any one definitive result?

Or did I read that wrong?

[clarkusdarkus]

Well my stance on it is this, I vote on all polls i come across on the web regarding ME3 as a whole and its ending, i always vote its crap...

Now out of my 5 friends who also dispise the game and its ending, I am the only one registered on here voting on polls and surveys around the web, My friends didnt buy the dlc as they gave it a few months and the EC did nothing for us all. They simply moved on and will not buy the next ME game as we was also dissapointed lately with bioware as a whole not just ME3..

Point is there are those who hate/like ME3 and its ending who just simply dont vote on such trivial things as polls and are the ones who either dont put the disk back in the console hence why ME3 wasnt even in the top 20 games of 2012 on xboxlive....people vote with there money more than they do on polls.

On a side note it wouldn't hurt if bioware released they're proof of the " minority " as we all reloaded the auto save to see the three endings so you cant base what ending was more popular than the other as we all reloaded just to see the ones we didnt choose first, So basically we all chose all 3 endings as a result.

Modifié par clarkusdarkus, 03 mars 2013 - 02:10 .

[iDeevil]

Indy_S wrote...

iDeevil wrote...

She didn't have to go out on a limb and make an objective claim she wasn't prepared to back up.

She also didn't have to back it up, despite what you may think.

I disagree with this statement, not from a technical standpoint but from an ethical one. If you are willing to dismiss an argument when you have a position of authority, you back up your statements. Even if it requires bending the truth, that is preferable to a dismissal without just cause.

You're making more of a moral argument than an ethical one, but this probably isn't the place to go there .

Seriously though, she doesn't have to back up anything she says.  Just like we don't have to believe her.  Choices/consequences etc etc.

[IanPolaris]

iDeevil wrote...

Cstaf wrote...
I have not seen the video. I have a habit of not watching an video of someone trying to make statistical inference who is not a statistician. It usually never ends well.
But to make it clear; it is not enough if you have 10, 100 or 1000 polls that says one thing to claim that to be the truth. You need to extract the bias of the aggregate of the polls prior to make an inference. And it is not certain after clearing the data of bias that you can draw a statistical conclussion of it given the relatively high standard error of the estimate this method produces.

So basically what you are saying is that it can produce a possibility of what the population estimates?
And/Or are you saying that standing alone, even with bias extracted, that these polls cannot be evidentary of any one definitive result?

Or did I read that wrong?

The way I am reading it is that  a large number of otherwise biased/self selected polls can be used to get useful data, but it's a difficult process and subject to some fairly large error bars.  Furthermore it's not enough to say that all such polls say the same thing, a priori.  Rather one must examine each poll and find the overall convergence (and it converges slowly if at all).  If I am getting the lesson wrong, please correct me.

-Polaris

[GreyLycanTrope]

I'm sick of that video.

[iDeevil]

Indy_S wrote...

Remarkably, I don't want the endings changed. Never have. They're bad but they're what we have. It would be a lot more obvious that BioWare was listening to the fans if they were engaged in conversation with them. There is no discussion here. Fans yell, silence answers.

And that's what I want. I want to know why things happened like they did. I want to know why they just dropped Harbinger as a character. I want to know why they decided to give us two Pinnochio stories at once. I want to know why they thought that those endings would be excellent or even satisfying. But they won't tell me.

These are answers which are never really given.  Sometimes they are artistic choices, sometimes money.  Often though, and this is the same with movies and TV, they are choices handed down from the suits who want to see certain things.  Who have found something trends and wants to play on it etc.

But you'll never get a true honest answer to these questions and it's often better not to ask.

[Indy_S]

iDeevil wrote...

These are answers which are never really given.  Sometimes they are artistic choices, sometimes money.  Often though, and this is the same with movies and TV, they are choices handed down from the suits who want to see certain things.  Who have found something trends and wants to play on it etc.

But you'll never get a true honest answer to these questions and it's often better not to ask.

I disagree strongly with the last part. Honest answers might be too much to hope for and I have doubts every question would be answered but it is always worth it to ask.

[zyntifox]

iDeevil wrote...

Cstaf wrote...
I have not seen the video. I have a habit of not watching an video of someone trying to make statistical inference who is not a statistician. It usually never ends well.
But to make it clear; it is not enough if you have 10, 100 or 1000 polls that says one thing to claim that to be the truth. You need to extract the bias of the aggregate of the polls prior to make an inference. And it is not certain after clearing the data of bias that you can draw a statistical conclussion of it given the relatively high standard error of the estimate this method produces.

So basically what you are saying is that it can produce a possibility of what the population estimates?
And/Or are you saying that standing alone, even with bias extracted, that these polls cannot be evidentary of any one definitive result?

Or did I read that wrong?

I don't know how to do those fancy things where you seperates the qoutes so will just copy you two questions:

So basically what you are saying is that it can produce a possibility of what the population estimates?
You can sometimes, not always obviously. It depends on the distribution of the data and if they are on equal footings.

And/Or are you saying that standing alone, even with bias extracted,
that these polls cannot be evidentary of any one definitive result?
Semi-correct. This method produces relatively high standard errors of the estimate. What affects the precision of the estimates is the distribution of the data and the number of polls (which need to be on equal footings). And since this method produces higher standard errors of the estimate than a random sampled poll will it will not produce statistically significant results as often as a random sampled poll.

[zyntifox]

IanPolaris wrote...

iDeevil wrote...

Cstaf wrote...
I have not seen the video. I have a habit of not watching an video of someone trying to make statistical inference who is not a statistician. It usually never ends well.
But to make it clear; it is not enough if you have 10, 100 or 1000 polls that says one thing to claim that to be the truth. You need to extract the bias of the aggregate of the polls prior to make an inference. And it is not certain after clearing the data of bias that you can draw a statistical conclussion of it given the relatively high standard error of the estimate this method produces.

So basically what you are saying is that it can produce a possibility of what the population estimates?
And/Or are you saying that standing alone, even with bias extracted, that these polls cannot be evidentary of any one definitive result?

Or did I read that wrong?

The way I am reading it is that  a large number of otherwise biased/self selected polls can be used to get useful data, but it's a difficult process and subject to some fairly large error bars.  Furthermore it's not enough to say that all such polls say the same thing, a priori.  Rather one must examine each poll and find the overall convergence (and it converges slowly if at all).  If I am getting the lesson wrong, please correct me.

-Polaris

Very good.

[iDeevil]

Indy_S wrote...

iDeevil wrote...

These are answers which are never really given.  Sometimes they are artistic choices, sometimes money.  Often though, and this is the same with movies and TV, they are choices handed down from the suits who want to see certain things.  Who have found something trends and wants to play on it etc.

But you'll never get a true honest answer to these questions and it's often better not to ask.

I disagree strongly with the last part. Honest answers might be too much to hope for and I have doubts every question would be answered but it is always worth it to ask.

I agree with that sentiment 999,999 times from a million, just not in this case.  Save your breath, or your fingers, for ones that will elicit a truthful, or any response.  Plus, there are so many fingers in the pies of games it might be hard to track the why,

[ElitePinecone]

Dr_Extrem wrote...

tbh ... the current state of the forums could have been avoided, if the devs would have talked to the fanbase from the beginning.

this is partially their own fault.

So.... people being obnoxious, abusive, completely irrational and generally obsessive to the point of deep concern is the fault of those they're attacking?

Nice to know next time I want to blame a victim that their own behaviour justifies my lack of proportionality, restraint, civility or, heck, humanity. 

[iDeevil]

Cstaf wrote...

I don't know how to do those fancy things where you seperates the qoutes so will just copy you two questions:

So basically what you are saying is that it can produce a possibility of what the population estimates?
You can sometimes, not always obviously. It depends on the distribution of the data and if they are on equal footings.

And/Or are you saying that standing alone, even with bias extracted,
that these polls cannot be evidentary of any one definitive result?
Semi-correct. This method produces relatively high standard errors of the estimate. What affects the precision of the estimates is the distribution of the data and the number of polls (which need to be on equal footings). And since this method produces higher standard errors of the estimate than a random sampled poll will it will not produce statistically significant results as often as a random sampled poll.

Thanks, I get you now.  Just out of curisoity - on face value - do you think it'd be possible to determine what most of the fanbase felt/feels about the finale?

Please feel free not to answer that because I know it's inprecise.

[IanPolaris]

ElitePinecone wrote...

Dr_Extrem wrote...

tbh ... the current state of the forums could have been avoided, if the devs would have talked to the fanbase from the beginning.

this is partially their own fault.

So.... people being obnoxious, abusive, completely irrational and generally obsessive to the point of deep concern is the fault of those they're attacking?

Nice to know next time I want to blame a victim that their own behaviour justifies my lack of proportionality, restraint, civility or, heck, humanity. 

Actually other than the usual suspects, while the dislike for ME3's ending was widespread and the disbelief spread rapidly here and elsewehre, for the most part at least initially the overall reaction was disbelief not anger.  People wanted to hear from Bioware and wanted an explainataion.

It was Bioware that said, "Artistic Integrity, let our fans eat cake"  That's when the mood really turned ugly.

-Polaris

[iDeevil]

ElitePinecone wrote...

Dr_Extrem wrote...

tbh ... the current state of the forums could have been avoided, if the devs would have talked to the fanbase from the beginning.

this is partially their own fault.

So.... people being obnoxious, abusive, completely irrational and generally obsessive to the point of deep concern is the fault of those they're attacking?

Nice to know next time I want to blame a victim that their own behaviour justifies my lack of proportionality, restraint, civility or, heck, humanity. 

To be fair they did intially try and talk to the fan base.  They were yelled *at* in return. And in some cases threatened.

I'm amazed after some of the stuff that was said that the devs, and anyone in charge, threw us a bone with the EC.  They would have been well within their rights to back up their games and go home.

[Indy_S]

ElitePinecone wrote...

So.... people being obnoxious, abusive, completely irrational and generally obsessive to the point of deep concern is the fault of those they're attacking?

Nice to know next time I want to blame a victim that their own behaviour justifies my lack of proportionality, restraint, civility or, heck, humanity. 

I'm going to weigh in and say that you misinterpreted the Doctor's statement. Rather than let a wound fester, you provide constant relief to aid recovery.

[JamesFaith]

OdanUrr wrote...

Like I've said before, these numbers share the same problems. First, they're very small samples of the people who've played the game. Second, it is more likely that the people willing to answer these polls are the ones who are still angry about the endings. Think about it, people who liked or feel indifferent about the endings probably won't bother answering any polls. As a general rule, people who are angry about something are more vocal about that something than those who are not.

Yeap.

I have positive opinion on endings and I participated just in one poll. I simply have no need to cry to world my opinion.

[zyntifox]

iDeevil wrote...

Cstaf wrote...

I don't know how to do those fancy things where you seperates the qoutes so will just copy you two questions:

So basically what you are saying is that it can produce a possibility of what the population estimates?
You can sometimes, not always obviously. It depends on the distribution of the data and if they are on equal footings.

And/Or are you saying that standing alone, even with bias extracted,
that these polls cannot be evidentary of any one definitive result?
Semi-correct. This method produces relatively high standard errors of the estimate. What affects the precision of the estimates is the distribution of the data and the number of polls (which need to be on equal footings). And since this method produces higher standard errors of the estimate than a random sampled poll will it will not produce statistically significant results as often as a random sampled poll.

Thanks, I get you now.  Just out of curisoity - on face value - do you think it'd be possible to determine what most of the fanbase felt/feels about the finale?

Please feel free not to answer that because I know it's inprecise.

I haven't seen a lot of polls regarding the ending so i can't make a educated guess about the validity of the inference you can make from them. But you can definitely do statistical tests on the processed data to determine if the proportion of the costumers who are satisfied are larger or equal to 0.5.

[Helios969]

I think a "majority" of fans are/were disastified with endings, however it's still not clear whether BW/EA has done any long term damage to itself.  I think the reality is that most VG players don't make the kind of emotional investment to care whether a game has a strong or weak ending...unfortunate from my standpoint as the story is the most important aspect for me.  Still, the game was hellafun as were the dlc's.  Despite many of the flaws I still think Bioware crafts the most interesting story content to drive their games...and I still haven't found another gaming company that even comes close to what I consider the perfect balance of RPG elements with combat mechanics that are accessible to even the most novice of players.

And they still seem to be making plenty of money. This may have been posted somewhere before, but I thought it relevant.


So far, EA has always benefited from the launch of DLCs for some of its famous franchises. In the recently concluded third quarter for instance, extra content and free-to-play were responsible for much of EA’s digital revenue growth. EA’s DLC and free-to-play micro transaction content shot up 50.0% year over year to $185.0 million in the quarter.

http://www.zacks.com...s-effect-3-dlcs

[IanPolaris]

Helios969 wrote...

I think a "majority" of fans are/were disastified with endings, however it's still not clear whether BW/EA has done any long term damage to itself.  I think the reality is that most VG players don't make the kind of emotional investment to care whether a game has a strong or weak ending...unfortunate from my standpoint as the story is the most important aspect for me.  Still, the game was hellafun as were the dlc's.  Despite many of the flaws I still think Bioware crafts the most interesting story content to drive their games...and I still haven't found another gaming company that even comes close to what I consider the perfect balance of RPG elements with combat mechanics that are accessible to even the most novice of players.

And they still seem to be making plenty of money. This may have been posted somewhere before, but I thought it relevant.


So far, EA has always benefited from the launch of DLCs for some of its famous franchises. In the recently concluded third quarter for instance, extra content and free-to-play were responsible for much of EA’s digital revenue growth. EA’s DLC and free-to-play micro transaction content shot up 50.0% year over year to $185.0 million in the quarter.

http://www.zacks.com...s-effect-3-dlcs

This last is why I think it's less than honest to compare ME3's sales to sales of other RPGs prior to it because ME3 is getting a lot of it's sales because of MP DLC or rather the microtransactiions that stem from it.  That puts that aspect of ME3 in a completely different category (MMO Shooter) and thus that revenue stream should be compared with pure SP RPGs (because it's different).

-Polaris

[LilyasAvalon]

I think people are more bitter than upset, stroking a festering wound to keep it filled with pus and swollen. Not that I entirely blame them, until AC3 came out, I was one of those people. Now I just shoot rabbits with arrows.

[AresKeith]

LilyasAvalon wrote...

I think people are more bitter than upset, stroking a festering wound to keep it filled with pus and swollen. Not that I entirely blame them, until AC3 came out, I was one of those people. Now I just shoot rabbits with arrows.

And Dead Space 3, ugh

[jijeebo]

RESPECT MY MAJORITAH!!1!


I swear I've seen this thread at least twice before... Y'all need to stop being so bitter and move on.

[Helios969]

IanPolaris wrote...

Helios969 wrote...

I think a "majority" of fans are/were disastified with endings, however it's still not clear whether BW/EA has done any long term damage to itself.  I think the reality is that most VG players don't make the kind of emotional investment to care whether a game has a strong or weak ending...unfortunate from my standpoint as the story is the most important aspect for me.  Still, the game was hellafun as were the dlc's.  Despite many of the flaws I still think Bioware crafts the most interesting story content to drive their games...and I still haven't found another gaming company that even comes close to what I consider the perfect balance of RPG elements with combat mechanics that are accessible to even the most novice of players.

And they still seem to be making plenty of money. This may have been posted somewhere before, but I thought it relevant.


So far, EA has always benefited from the launch of DLCs for some of its famous franchises. In the recently concluded third quarter for instance, extra content and free-to-play were responsible for much of EA’s digital revenue growth. EA’s DLC and free-to-play micro transaction content shot up 50.0% year over year to $185.0 million in the quarter.

http://www.zacks.com...s-effect-3-dlcs

This last is why I think it's less than honest to compare ME3's sales to sales of other RPGs prior to it because ME3 is getting a lot of it's sales because of MP DLC or rather the microtransactiions that stem from it.  That puts that aspect of ME3 in a completely different category (MMO Shooter) and thus that revenue stream should be compared with pure SP RPGs (because it's different).

-Polaris

I can't remember where but I thought I read somewhere that single player dlc sales are up too.  Regardless, I've seen lots of polls about the majority hate ME3 to the point they're not going to buy extra content or support future projects, but no real numbers to support the claims.  I guess we won't know for a couple years, though I'd maintain that any fanbase loss will be replaced by next gen gamers.  Somehow I don't think Bioware is going to fail anytime soon.

I'm not happy with the ending at all, but if I stopped playing Bioware games I'd essentially be giving up video games all together.  And I still think the good far outweighs the bad.