SET FREE THE ASCENDANCY POINTS (or rework the lab) [New ascension methods/lab rework ideas]

"

Nishrek wrote:

Not being vocal about lab =/= liking current lab design
This is a fact.
This one fact is enough to invalidate your argument about "silent majority"
Also about "infinite amount of data" (BTW check what word "infinite" means next time)
It is funny that in Beta forum they are requesting answers which should be easily collected by the automated data gathering (also much precise, and on a bigger sample, because not 100% of beta players answers in this thread). Unless of course, they do not collect so much RAW data, or not consider it so relevant as you claim...
It was already explained that GGG can't get relevant data from their telemetry, and they can't use it to decide how big part of the community actually finds lab interesting.
And this is a fact.

Sure, I would easily guess 95% of people don't give two shits either way, and just engage with the lab the same way they engage with the rest of the content, especially because it's so similar to the rest of the content that they don't really see the lab as different. Then again I'm mostly here because what Turtledove is doing is an affront to data analysis.

"

Not sure where fruz is getting his number but his assumption is dumb as hell. In my guild every time I build lab 3/4 of my guild complains about it and voice their discontent with it. Did all of them voice their frustration on the forum? No, just 1 because of the assumption that GGG does not care about feedback, so do yourself a favor and stop embarrassing yourself by trying to embellish numbers.

Instead of random calculations and assumptions and other yada yada bullshit, let's look at the fact at hand. Numbers of people that voice their discontent for the forum versus number of people that love it. The whole idea that ''GGG have the internal numbers'' is a joke. No, they do not have the numbers if they ask people on the beta to provide their feedback on lab.

As long as you do proper data corrections based on preexisting market feedback data that's fine, and positive non solicited non invasive feedback is much much much less likely than negative feedback of the same category.

Yo turtle are you going to play 3.0 when it comes out?

"

Fruz wrote:

"

Turtledove wrote:

[...]

I even gave you an example ....


People that don't like something are more likely to voice their opinion here
-> among the posters here (<=>the players using the forum), the proportion of "lab hater" is (much) higher than among the player that do not voice their opinion.

Let's simplify it even more.

If you have 100 players, 20 dislike the lab, 55 are neutral, 19 like it, 6 really like it ( example here ).
Let's say ... 11 "dislikers" are voicing their opinion, the neutrals can't be bothered, 2 "likers" voice their opinion, and 2 that really like it voice their opinion.


We have 15 people voicing their opinion, including 11 that dislike it.
Let's take your fallacious reasoning, dividing 11 by 0.15 = we have now an estimation 73 players that dislike the lab ( when in fact .... it's 20 ).

"

Turtledove" wrote:

Did you forgot to put on your thinking cap?

(and you where the one attacking me on my English, earlier in this thread ?)

Here's the equation being discussed.
914/0.5% = 182,800 estimated players have a problem with labyrinth out of the active player base

The 0.5% was my hypothetical guess as to the percentage of people out of the active player base that have posted in the forums about ANYTHING in the past 16 months.

The 914 figure is the current number of names in the list. Over 330 threads discussing labyrinth problems with over 910 posters in support
___________________________________

Fruz, your above statements are completely irrelevant to the above equation. Let me try a different way of explaining the equation.

If we continue to use the hypothetical 0.5% of the active player base has posted in the forums on anything in the past 16 months. It would mean that each poster to the forum represents 200 players in the active player base (i.e., 1 / 0.5% = 200). If you multiply 914x200=182,800 which is the same as above by directly dividing by 0.5%. So in theory there would be 182,800 active players in the currently active player base that would generally agree with the people on the list.

I never attacked you for your English. I was just pointing out that it appeared to me that your personality occasionally changed and you became significantly more combative at times. These posts seemed to accompany a lower level of literacy than in your normally highly literate posts. At those times I've learned it's probably best to not bother trying to communicate.

"

j33bus wrote:

You haven't accounted for background and other major statistical biases in your data taking method, most of all that your list doesn't actually include people who are in your list whose problems have been solved, or people who have a small problem with the lab but generally like it. Also of course you're not dealing with possible alt account and other trolls, or the fact that people are more likely to post negative feedback so the flat extrapolation isn't correct. Did you forget to put on your bayesian thinking cap to do stats right? Because without a whole lot more work your list can ONLY be interpreted as these specific people had someone negative to say about the lab at the time, but don't necessarily have anything bad to say now.

Well a quick google search on the subject actually tells me that many studies say that one complaint is worth only about 25 people regardless of population stats.

Thank you for at least seemingly having a basic understanding of the simple equation and directly addressing the issues!! Very refreshing!!!

I don't see great relevance in your argument about people complaining being more likely to complain about labyrinth. The reason is that looking at the list over the past three hundred or so new people being added, I think that there's only been one (maybe two) people that had posted about labyrinth as their first posts to the forum. The rest had posts on the forum about other subjects prior to them voicing their opinion about labyrinth.

Yes, there have been a number of fixes to labyrinth, reduced trials and 60% ES as good examples. That would lower the numbers somewhat, agreed. On the other hand, there are likely far more posters that have posted on the forum but are not on the list at this time. This argument is supported by the fact that the list is still growing and by the fact that checking other sub-areas of the forum besides Feedback and Suggestions seems to show a much higher percentage of first time labyrinth complainers.

I don't understand your assertion about 25 people, in that there are national studies where they interview just a few thousand people and come to estimated percentages for the whole country.

You've used a purely frequentist estimator and everyone who ever does statistical analysis knows that that's not the way to do it.

Common market research says about 4% of people with an issue complain. So you only get to multiply by 25 not 200, and of course as stated before you 910 number conflates people who want a minor improvement to the lab, people who are raging because they died, and people who legitimately hate it. Your stats only get you at most 25,000 people, but more likely something like 15,000, when done well.

https://www.linkedin.com/pulse/20130604134550-284615-15-statistics-that-should-change-the-business-world-but-haven-t

The adjustment is just common bayesian unfolding to known trends in customer feedback. It comes from the fact the people were not interviewed here, just an open option for people to choose whether or not to give feedback, which means that you do not get a representative population sample.

Simply put a survey was not conducted, and so you can't extrapolate the population in the simple way you did.

"

j33bus wrote:

You've used a purely frequentist estimator and everyone who ever does statistical analysis knows that that's not the way to do it.

Common market research says about 4% of people with an issue complain. So you only get to multiply by 25 not 200, and of course as stated before you 910 number conflates people who want a minor improvement to the lab, people who are raging because they died, and people who legitimately hate it. Your stats only get you at most 25,000 people, but more likely something like 15,000, when done well.

https://www.linkedin.com/pulse/20130604134550-284615-15-statistics-that-should-change-the-business-world-but-haven-t

Nope, you're comparing apples and oranges. The 0.5% figure is not people complaining. That is the total number of active players that have posted to the forums. It is our total sample size. The assumption built into the equation is that 914 of our sample size dislike something about labyrinth and everyone else in our sample size is a labyrinth lover or a labyrinth I-don't-care.

"

Turtledove wrote:

"

j33bus wrote:

You've used a purely frequentist estimator and everyone who ever does statistical analysis knows that that's not the way to do it.

Common market research says about 4% of people with an issue complain. So you only get to multiply by 25 not 200, and of course as stated before you 910 number conflates people who want a minor improvement to the lab, people who are raging because they died, and people who legitimately hate it. Your stats only get you at most 25,000 people, but more likely something like 15,000, when done well.

https://www.linkedin.com/pulse/20130604134550-284615-15-statistics-that-should-change-the-business-world-but-haven-t

Nope, you're comparing apples and oranges. The 0.5% figure is not people complaining. That is the total number of active players that have posted to the forums. It is our total sample size. The assumption built into the equation is that 914 of our sample size dislike something about labyrinth and everyone else in our sample size is a labyrinth lover or a labyrinth I-don't-care.

I understand that, but what I'm saying is that it's not actually your sample size, you don't actually have a sample size in that sense because your didn't conduct a survey, you just tabulated volunteered feedback, and volunteered feedback is known to not get you a representative population sample, so lots of market research has been done to get you to that 1:25 number.

"

j33bus wrote:

I understand that, but what I'm saying is that it's not actually your sample size, you don't actually have a sample size in that sense because your didn't conduct a survey, you just tabulated volunteered feedback, and volunteered feedback is known to not get you a representative population sample, so lots of market research has been done to get you to that 1:25 number.

It's the same thing as a survey. (I'll come back to this point later.) Think of it this way. We have a survey of 0.5% of the player base. The survey responses are all posts to the forum since labyrinth release. The number of complainers has been tabulated at 914. (Sure I missed alot, especially in other areas outside Feedback and Suggestions, as mentioned.)

Returning to "same thing as a survey", here's areas where I believe the posters to the forum are different from the whole player base or a purely random sampling of the player base.

- Casual players are much less likely to post on the forum than more dedicated more frequent players.

- Forum posters are more likely more skilled at playing PoE than the more casual players that don't post.

- Forum posters are more likely to donate/buy points from GGG.

- What else?

edit: I hope I don't have to mention again that the 0.5% figure is a hypothetical figure, presumably GGG knows or could know what the true value of this number would be.

"

Turtledove wrote:

"

j33bus wrote:

I understand that, but what I'm saying is that it's not actually your sample size, you don't actually have a sample size in that sense because your didn't conduct a survey, you just tabulated volunteered feedback, and volunteered feedback is known to not get you a representative population sample, so lots of market research has been done to get you to that 1:25 number.

It's the same thing as a survey. (I'll come back to this point later.) Think of it this way. We have a survey of 0.5% of the player base. The survey responses are all posts to the forum since labyrinth release. The number of complainers has been tabulated at 914. (Sure I missed alot, especially in other areas outside Feedback and Suggestions, as mentioned.)

Returning to "same thing as a survey", here's areas where I believe the posters to the forum are different from the whole player base or a purely random sampling of the player base.

- Casual players are much less likely to post on the forum than more dedicated more frequent players.

- Forum posters are more likely more skilled at playing PoE than the more casual players that don't post.

- Forum posters are more likely to donate/buy points from GGG.

- What else?

edit: I hope I don't have to mention again that the 0.5% figure is a hypothetical figure, presumably GGG knows or could know what the true value of this number would be.

It's not a survey, you haven't distributed this among a random sample of the population, you waited for someone to come forward with a complaint, that's not a survey and cannot be treated as such. The 0.5% number is arbitrary, sure, but it doesn't matter in the slightest, because this isn't a survey, and you can't extrapolate to the population like that. This is the same as someone calling a company with a complaint, 100% of users have access to it, but you don't get an even distribution, and a simple extrapolation to total population size doesn't suffice for this kind of data analysis, because your inputs are decidedly non random. In order to do a survey you need a seek out a sample of people who are representative of the total population and not just tabulate people who actively seek you out. This is where the 1 complaint is 25 people number comes from, because market researches know that complaints aren't a representative sample of users, and so they actually did survey to figure out the amount of dissatisfied people actually complain.

To say this again, frequentist statistics do not get you to the right answer in the real world you need to be at least a little bayesian.

"

j33bus wrote:

"

Turtledove wrote:

"

j33bus wrote:

I understand that, but what I'm saying is that it's not actually your sample size, you don't actually have a sample size in that sense because your didn't conduct a survey, you just tabulated volunteered feedback, and volunteered feedback is known to not get you a representative population sample, so lots of market research has been done to get you to that 1:25 number.

It's the same thing as a survey. (I'll come back to this point later.) Think of it this way. We have a survey of 0.5% of the player base. The survey responses are all posts to the forum since labyrinth release. The number of complainers has been tabulated at 914. (Sure I missed alot, especially in other areas outside Feedback and Suggestions, as mentioned.)

Returning to "same thing as a survey", here's areas where I believe the posters to the forum are different from the whole player base or a purely random sampling of the player base.

- Casual players are much less likely to post on the forum than more dedicated more frequent players.

- Forum posters are more likely more skilled at playing PoE than the more casual players that don't post.

- Forum posters are more likely to donate/buy points from GGG.

- What else?

edit: I hope I don't have to mention again that the 0.5% figure is a hypothetical figure, presumably GGG knows or could know what the true value of this number would be.

It's not a survey, you haven't distributed this among a random sample of the population, you waited for someone to come forward with a complaint, that's not a survey and cannot be treated as such. The 0.5% number is arbitrary, sure, but it doesn't matter in the slightest, because this isn't a survey, and you can't extrapolate to the population like that. This is the same as someone calling a company with a complaint, 100% of users have access to it, but you don't get an even distribution, and a simple extrapolation to total population size doesn't suffice for this kind of data analysis, because your inputs are decidedly non random. In order to do a survey you need a seek out a sample of people who are representative of the total population and not just tabulate people who actively seek you out. This is where the 1 complaint is 25 people number comes from, because market researches know that complaints aren't a representative sample of users, and so they actually did survey to figure out the amount of dissatisfied people actually complain.

To say this again, frequentist statistics do not get you to the right answer in the real world you need to be at least a little bayesian.

You could take a stab at estimating a +/- but I'm not interested enough myself. I'm not claiming any highly accurate thing here, just a rough estimate. Sometimes all you can work with is what you got. Use what you got and make the best of it.

I listed all the things I could think of that would differentiate all posters in the forum in the past 16 months from a random sample survey. None of which seemed would be significantly problematic? Perhaps I missed some other things though?

How about further potential problems with our "sample" being representative of the full player base.

- People could create a GGG account and post about PoE without ever playing the game. (Highly unlikely in my opinion. edit: Second thought this one would probably skew the number more to the pro-lab side anyway.)

- People could create multiple alt-accounts and post their dislike for labyrinth. (Not apparently really a widespread problem based on the number of old accounts created prior to labyrinth being over 96% of the 914 names on the list.)

- Perhaps I'm just lying about everything and the list is totally fake? (Zaludos made the original list at 582 names. I've tried to be extremely transparent by posting details of almost everyone I've added to the list at the time when I add them. Some pro-lab posters have done some verification of the list on their own in the past and their posts are referenced in the OP of the list thread.)

- Other things?