Let's discuss Learn By Doing for hopefully the last time.

[zztong]

I can see how AI came to that conclusion. with A18 & A19 Did you happen to factor in Covid and the majority of the world went into lock down allowing an uptick in new players through new sales. Due to everyone being stuck inside?
Did you also factor in that with low users/sales the variables on a graph will show a higher degree of variation?
I have no problem with you are taking AI references but when you do make sure you make sure you advise others.

I didn't use AI. The source of that data is lorenzostanco . com and the News posts by TFP to get the versions which I added to the graph by hand.

As I pointed out that I have no insight as to how they collect that data. Thus, I have no idea if COVID figures into that data or how. Could COVID have influenced people's ratings? I suppose so. Again, the data comes from lorenzostanco . com which is a black box to me. I cited that limitation in my post. If lorenzostanco . com is inaccurate, or I made a mistake in gathering release dates from the News posts, then my post falls apart.

This is me being honest about my conjecture.

I dont think you understand the the statement.

I believe I do. You said unless somebody can refute you, it must be true. By that logic, if you said you looked through your telescope and saw a walrus in orbit around Pluto, then I can't refute it because I wasn't there, so it must be true.

I have played down the cards you haven't. Making your argument invalid and your analogy correct.

But, if I correctly recall your post about your methods, you haven't played your cards, just described them. You cited methods along the lines of running a series of AI prompts that you did not show. You used four different AIs but did not name them. AI got its data from somewhere but you didn't list the sources. We cannot view the results. I cannot repeat your methods to see if I get the same result because didn't share enough detail. I can't evaluate the quality of the data coming out of your AI prompts, nor have any way to get AI to describe the sources of the data it used because that would all be locked up in your conversation with each of those AIs.

Now if I am incorrectly remembering, or overlooking your post with more detailed methods, then you'll have my apology. We are all human. And I'd love to be wrong as that would mean there's meaningful, useful data.

Lets not forget that those older versions of the game were far more janky than the current versions. Very buggy and very ugly things that I think drive negative reviews far more strongly than whatever systems might be in place. I think no matter your stance on LBD, no one can argue that the game hasn't become more polished over the years.

Agreed. As I mentioned, we only know relative positivity of the whole product, with limits on who could and did report. I acknowledged that LBD cannot be teased out of the data. I only stated that later versions without LBD were of higher relative positivity and then speculated that perhaps that was a sign that LBD wasn't missed. But as you say, the quality of the game's improvement could be the preponderance of new/changed features and the majority could still have been missing LBD.

 

[Riamus]

Since actual data is more useful than conjecture, I'll just add that I don't like LBD. I might consider a hybrid option where it isn't used at all with any form of crafting, and where anything related to weapons requires killing something and not just swinging a weapon around, and where you don't have to jump hundreds of times just to have a high parkour. But if any of those things are there, then I don't want it.

Now, that is one known data point. That is hardly a view of what everyone else thinks. Some will agree with me and some will not. That's true about anything. It is not possible to make everyone happy. Trying to do so just muddles a game and leaves everyone upset. Why people seem to think that every game should be an exact fit to what they want, I'm not sure. If you don't like a game, there are plenty of others to choose from that will be more to your liking. I move on to other games all the time when I don't like a game.

 

[Charger265]

I didn't use AI. The source of that data is lorenzostanco . com

Sorry for coming off so combative i took as "it was your own words". My mistake.
The data obtained from that site is AI generated. This was some of the early responses i received in my data set before cleaning it up.
By the way nice work on the graphs

I believe I do. You said unless somebody can refute you, it must be true. By that logic, if you said you looked through your telescope and saw a walrus in orbit around Pluto, then I can't refute it because I wasn't there, so it must be true.

I never said that.
All i have presented is data that provides weight to my logical conclusion that there is a marginal majority that are unhappy with the game in some form.
Along with that i also provided my investigation methods and applied variables and external factors to the model to form a logical conclusion. Why by all account is far more than i have seen anyone do regarding this matter to understand the depth of the rift within the community.

In my discussion with others i have have implied that based on their stance to refute my conclusion they would have to apply the same rule/argument they are using against my conclusion. Which would infer and double negative. Pointing out exactly why the pretense of "unless somebody can refute you, it must be true" is a silly argument to present.

 

[Soulmonster]

You have said you have no way of knowing that im wrong or right.

I have no way of knowing whether the majority of gamers are disappointed with the direction of the game, but I know you are wrong in believing they must be based on the opinions of a possibly non-representative subset of gamers.

 

[Charger265]

But, if I correctly recall your post about your methods, you haven't played your cards, just described them. You cited methods along the lines of running a series of AI prompts that you did not show. You used four different AIs but did not name them. AI got its data from somewhere but you didn't list the sources. We cannot view the results. I cannot repeat your methods to see if I get the same result because didn't share enough detail. I can't evaluate the quality of the data coming out of your AI prompts, nor have any way to get AI to describe the sources of the data it used because that would all be locked up in your conversation with each of those AIs.

C

Post in thread 'Time to reflect'

Wednesday at 6:50 PM

Ok to remove all biased aspect i thought i would let Ai do an in depth analysis. used my API to push it out to Grok/chatgpt/gemini/deepseek
then asked it to combine and summarize the conclusion.

Community Backlash and Steam Rating DeclineOverview of Backlash:

The 7 Days to Die community has been vocal about their dissatisfaction with the game’s direction post-1.0 release (July 2024) and subsequent updates (V2.0, V2.1, V2.2). Key points of contention include:

Slow Updates and Content Focus:
Players criticize the game for receiving only one major update per year, often focusing on new...

• Charger265

• 
• 

"used my API to push it out to Grok/chatgpt/gemini/deepseek"

Method:

C

Post in thread 'Time to reflect'

Friday at 10:05 PM

AI is a useful tool. However, it is severely limited by what data is available, and even being careful with your prompts can't fix that. Take something that gets a lot of people talking about it who are incorrect in what they are saying. AI will see all those comments and little against them and state it as fact. There are ways to help mitigate this, such as telling AI to ignore all data from forums, but it is still always a problem. And this is especially true with opinions. AI will assume that the loudest voices are the most popular, just like people do. It will then state it...

• Charger265

To go into more detail i used google/yandex/bing and duckduckgo which is basically the same as google to identify sites for scraping data. I also specifically used steam/reddit/X and Facebook for the body of the content via a phthon script. To be honest im surprised my IP wasnt flagged for bot traffic.

I also did advise that if i was to share the data base there is no good way to post it here because just the instruction set used for the api prompts would fill 10+ forum pages

I even went as far as saying its not definitive proof and by no means am i a statistical expert. I dont know how i can be any more transparent without doing a full data dump which the likely hood of anyone here being able to understanding with any certain degree, it very slim.

 

[Soulmonster]

All i have presented is data that provides weight to my logical conclusion that there is a marginal majority that are unhappy with the game in some form.

Now you are starting to move the goal posts. Your original claim wasn't about the game "in some form", but specifically the direction of the game:

Because this is what you wrote: "However statically the majority are unhappy with the game direction"

It is also interesting to note that you now have included the qualifier "marginal majority" instead of just "majority", suggesting that you realize you are on weak grounds but are too proud to admit the possibility of being wrong.

 

[zztong]

Sorry for coming off so combative i took as "it was your own words". My mistake.
The data obtained from that site is AI generated. This was some of the early responses i received in my data set before cleaning it up.
By the way nice work on the graphs

No worries. Emotions run high when we all have a passion for the game. Let me buy you a beer. (Figuratively, unless we meet at some gathering.)

Thanks for pointing out that lorenzostanco . com comes from AI. That's disappointing, as I had hoped it were more reliable. His change log doesn't mention AI, and the data seems to predate the LLMs, so I had hoped for some kind of direct extract.

Steam filters changelog

www.lorenzostanco.com

I never said that.
All i have presented is data that provides weight to my logical conclusion that there is a marginal majority that are unhappy with the game in some form.
Along with that i also provided my investigation methods and applied variables and external factors to the model to form a logical conclusion. Why by all account is far more than i have seen anyone do regarding this matter to understand the depth of the rift within the community.

I applaud the effort. We're quibbling over "the data", but I'm all for information and attempts to understand it.

In my discussion with others i have have implied that based on their stance to refute my conclusion they would have to apply the same rule/argument they are using against my conclusion. Which would infer and double negative. Pointing out exactly why the pretense of "unless somebody can refute you, it must be true" is a silly argument to present.

I'm not clear on the point you're making as from my perspective -- perhaps not the complete perspective -- I've seen people basically say "you can't apply your conclusion to the majority of people who play the game" as they are pointing out none of us has found a way to get a representative sample of everyone.

My perspective is this: I teach college classes to undergraduates and graduate students. I've had to advise, and in a couple of cases fail, graduate students who leaned heavily on AI using poor methods. Of the most blatant have been students who had AI write papers where the AI fabricated the sources for the attributions. The students presented everything as fact. It wasn't. They had no support.

On these forums, there was a discussion of PvP and a person used AI to show a trend. Using the same AI was used to repeat the observation it came to light the primary evidence was the very discussion we were having. The AI had so little data to go on, it speculated based on what it could see, which was only that discussion. In effect, a loop with the AI telling us what we had said, but portraying it as coming from a wider audience.

But overall, I don't know you, so I don't know your background. I may just think I understand your methods. I could be wrong about everything.

 

[HammerDano]

if only we had a community manager...someone to explain WHY they removed LBD...maybe they would sort thru and find the relevant conversations IF they exist...or they would have the SOURCE EXPERTS explain it. Would that make everyone happy? doubtful...but they would be informed, and a circular argument could end...or get pointed in a useful direction.

 

[theFlu]

there is a marginal majority that are unhappy

That's a claim that can't really be made by looking at public internet data. You only have two biased samples; people expressing their complaints and people who are expressing their happiness. No matter which way that combination skews, it does not reveal the opinion of the silents.

Marginal majority of the minority that's expressing their opinion publicly, sure. Any kind of majority of the overall player base, no. You can not extrapolate data beyond what you've measured. In this case "in between" the two ends you (can) have measured.

 

[Charger265]

I have no way of knowing whether the majority of gamers are disappointed with the direction of the game, but I know you are wrong in believing they must be based on the opinions of a possibly non-representative subset of gamers.

So im using your same argument you used on me.. Prove it.

Now you are starting to move the goal posts. Your original claim wasn't about the game "in some form", but specifically the direction of the game:

Because this is what you wrote: "However statically the majority are unhappy with the game direction"

It is also interesting to note that you now have included the qualifier "marginal majority" instead of just "majority", suggesting that you realize you are on weak grounds but are too proud to admit the possibility of being wrong.

I know its hard for you to follow but you are mixing up discussions to try and validate your argument. Im going to disengage from this discussion with you at this point. This is just combative behavior.

 

[zztong]

It is also interesting to note that you now have included the qualifier "marginal majority" instead of just "majority", suggesting that you realize you are on weak grounds but are too proud to admit the possibility of being wrong.

Careful. We're all human. If the message is evolving based on feedback then I think that's a good thing. We'll all be in a better position to understand the world as best we can.

 

[Charger265]

I'm not clear on the point you're making as from my perspective -- perhaps not the complete perspective -- I've seen people basically say "you can't apply your conclusion to the majority of people who play the game" as they are pointing out none of us has found a way to get a representative sample of everyone.

My perspective is this: I teach college classes to undergraduates and graduate students. I've had to advise, and in a couple of cases fail, graduate students who leaned heavily on AI using poor methods. Of the most blatant have been students who had AI write papers where the AI fabricated the sources for the attributions. The students presented everything as fact. It wasn't. They had no support.

On these forums, there was a discussion of PvP and a person used AI to show a trend. Using the same AI was used to repeat the observation it came to light the primary evidence was the very discussion we were having. The AI had so little data to go on, it speculated based on what it could see, which was only that discussion. In effect, a loop with the AI telling us what we had said, but portraying it as coming from a wider audience.

But overall, I don't know you, so I don't know your background. I may just think I understand your methods. I could be wrong about everything.

Im more than happy to have these discussion on this level. All the right questions are being asked and your perspective is included. Thanks for coming around.

I also did touch on this subject in previous posts
"AI hallucination"
AI does this all the time. The best method i have found to mitigate it is by the method i described have each step cross reference the data across different AI models. Its extremely time consuming wihtout using a database and API calling
I also did mention (self aware) "Yup i am just some random on a forum with no context to their background. So you have to be rightfully skeptical."

Again trying to be as transparent as i can. I do find it very strange that when one is willing to put the time and effort in to actually looking into the available data and applying sound methods to find common ground to address an issue and prompt meaningful changes that will most likely benefit everyone who plays 7DTD is met with such combativeness. I have to say from my years in the past of running forums the echo chambers are getting worse. Or maybe its just here as i try to avoid engaging for these exact reasons.

 

[Charger265]

That's a claim that can't really be made by looking at public internet data. You only have two biased samples; people expressing their complaints and people who are expressing their happiness. No matter which way that combination skews, it does not reveal the opinion of the silents.

Marginal majority of the minority that's expressing their opinion publicly, sure. Any kind of majority of the overall player base, no. You can not extrapolate data beyond what you've measured. In this case "in between" the two ends you (can) have measured.

At this point i urge you to educate yourself on the methods used and sources of data.

Using the words majority vs marginal majority they are the exact same thing.
In a mathematical context, majority generally refers to a number greater than half of the total, or more than 50%. Marginal majority specifically refers to the minimum required for a simple majority which is 50% in this case.

 

[zztong]

Im more than happy to have these discussion on this level. All the right questions are being asked and your perspective is included. Thanks for coming around.

I would say I still agree with Soulmonster. Your data is suspect and leaning on it doesn't support your observations when applied to all players.

Quoting from the other topic where you summarize your methods:

Ok to remove all biased aspect i thought i would let Ai do an in depth analysis. used my API to push it out to Grok/chatgpt/gemini/deepseek
then asked it to combine and summarize the conclusion.

Your source is AI. What is AI's source?

Or, more to the point, AI cannot instantly run a reliable survey of the all players. It can only work with what is available. In this case...

Community Backlash and Steam Rating DeclineOverview of Backlash:

... the "community." It isn't clear if that is Steam, Reddit, the Forums, and/or anything else. If it were all of those, that's still not a sample that allows for an accurate determination of everyone. It only describes those who have participated.

That is Soulmonster's point and they're right.

That said, your observations could be accurate when it comes to describing what AI did sample. We don't know. AI isn't great about showing its work, and...

"AI hallucination"
AI does this all the time. The best method i have found to mitigate it is by the method i described have each step cross reference the data across different AI models. Its extremely time consuming wihtout using a database and API calling
I also did mention (self aware) "Yup i am just some random on a forum with no context to their background. So you have to be rightfully skeptical."

But crowdsourcing LLMs doesn't add quality. The real data for all players doesn't exist. No LLM can change that.

Soulmonster isn't proposing an alternative interpretation. They're just pointing out the flaw in your statement.

If I were to characterize your results, based on what I think I understand, I would say "Asking multiple AIs to interpret the mood of an unspecified community of 7D2D players on various forums, the AI's interpreted the mood to be..."

And, I would site that the specific messages on those forums and the words chosen to characterize those moods were selected by the algorithms of those AIs and are not available for review.

 

[Riamus]

if only we had a community manager...someone to explain WHY they removed LBD...maybe they would sort thru and find the relevant conversations IF they exist...or they would have the SOURCE EXPERTS explain it. Would that make everyone happy? doubtful...but they would be informed, and a circular argument could end...or get pointed in a useful direction.

I'm sure there was something said back then. Faatal is good at giving details about things and I'm sure that happened when LBD was removed. It may be as simple as "we didn't like it", but you asked for the reason... the reason doesn't necessarily have to be a "good" one. But there also could be a lot of factors that went into that. I have no interest in trying to wade through posts from many years ago to find it, but that doesn't mean it doesn't exist. And they don't exactly need to keep rehashing their reasons for years just because there are new people playing the game.

I would point out that some of the people who have been around here a long time, including moderators, but also others, have given reasons for why certain things were changed. Does it matter if they are telling you the reasons instead of the devs doing so or a community manager doing so? If the answer is the same, why does it matter who said it? Why is it that even if an answer is given, people so often ignore it? Is it just because they don't like the answer? If that's the case, it doesn't matter who gives the answer because they still won't like the answer.

I'm not saying having a community manager is bad. But having one doesn't mean people will accept the reasons any more than they do now.

 

[Soulmonster]

So im using your same argument you used on me.. Prove it.

I know its hard for you to follow but you are mixing up discussions to try and validate your argument. Im going to disengage from this discussion with you at this point. This is just combative behavior.

You want me to prove that you haven't proven something? Give that some more thought. And if you don't get it, read up on "proving the null hypothesis".

And no, I am not mixing up discussions. Your quote ("However statically the majority are unhappy with the game direction") came n this very thread and is what prompted my involvement: https://community.thefunpimps.com/t...efully-the-last-time.44501/page-4#post-607969

I am sorry you find my behavior combative, I feel, on the other hand, that I have been quite patient with you despite your attempts at insulting my intelligence.

 

[Riamus]

Regarding the AI discussion, I already posted a LOT of stuff about its flaws and limitations in the other thread, so I won't get into that again. But I'll just remind people that, with enough data points, you can make what is considered a statistically accurate interpretation without interviewing every person. However, we don't have anywhere near enough data for that to be possible in this instance. Being clear when showing an AI's interpretation that this is using too little data to be statistically accurate and that the actual results could be entirely different is a good thing whenever you use AI without enough data. Using AI should always come with a number of disclaimers about its potential accuracy. You can say that based on what you were able to get for an interpretation from AI, you believe it is likely true that X, Y, or Z is true is one thing. Saying that based on the AI results, such and such is the truth is another. Now, I don't think the original posts with the AI said the latter without at least saying something like "I think" in there, but I don't feel like going and checking. Even so, it is good to be extra clear about the data not having enough information to be accurate when using AI.

Also note that it is extremely easy to manipulate AI results. And I'm not talking about using a prompt to manipulate it when you're trying to get an answer, though that's obviously possible. Instead, I am talking about others doing things to intentionally manipulate the results that AI can generate when different people are trying to look something up. This also applies to manipulating people, but it affects AI more because AI doesn't tell you the sources it uses if you're not specifically limiting it to specific sources. Take the example that ZZTong game about the AI being used for the PVP discussion. It had no other data and there were a lot of posts about it here, so it used our own posts as proof of what it said. By having so many posts about it in one place, AI treats those posts as being an accurate representation of the facts and presents it as fact. I almost wonder if Grandpa Minion is actually posting all his fake numbers in an effort to manipulate AI so it will use his fake numbers in its results, though I doubt that's the case. But it is certainly easy enough to do that to AI results. Post something enough times, especially in multiple locations, and AI will treat it as fact. It's harder to do with something that already has a lot of information available about it, but for things with minimal information, it's extremely easy to manipulate AI if you want to. I could go post a few hundred times in different places that this game is going to add the ability to fly to another planet and then when someone asks AI if there are plans to fly to another planet in this game, AI will say that there are plans to do so. It's that simple. Without being able to verify the sources, you can't be sure if the sources are real or fake.

 

[theFlu]

At this point i urge you to educate yourself on the methods used and sources of data.

"Educate yourself". I've heard that before, never in good faith. I read your description of your data:
Public data into an AI Blackbox contraption.

The "processed by AI" doesn't improve the quality of the input, and can't begin to correct for the silent.

Your claim covers:
"Happy and loud" "Happy and silent" "Unhappy and silent" "Unhappy and loud"

Your data coverage, to the best of the AI's ability:
"Happy and loud" " ???????????????????????????????????????????????????????????????????????????????????" "Unhappy and loud"

Not to scale, but I'd be surprised if your coverage was better than shown.
Your claim of a marginal majority, lands square in the land of the question marks.
AI can't unbake that cake.

 

[HammerDano]

I'm sure there was something said back then. Faatal is good at giving details about things and I'm sure that happened when LBD was removed. It may be as simple as "we didn't like it", but you asked for the reason... the reason doesn't necessarily have to be a "good" one. But there also could be a lot of factors that went into that. I have no interest in trying to wade through posts from many years ago to find it, but that doesn't mean it doesn't exist. And they don't exactly need to keep rehashing their reasons for years just because there are new people playing the game.

I would point out that some of the people who have been around here a long time, including moderators, but also others, have given reasons for why certain things were changed. Does it matter if they are telling you the reasons instead of the devs doing so or a community manager doing so? If the answer is the same, why does it matter who said it? Why is it that even if an answer is given, people so often ignore it? Is it just because they don't like the answer? If that's the case, it doesn't matter who gives the answer because they still won't like the answer.

I'm not saying having a community manager is bad. But having one doesn't mean people will accept the reasons any more than they do now.

If I ask a question to a dev or a moderator, I want them to answer. As you pointed out new people keep showing up, they don't know who is right or wrong. hell...even the long haulers can be accused of not knowing what they think they know. lol
that's why I think people prefer to get "official" answers.

I get why you wouldn't want to go back thru now to look at stuff from years ago...but if it did happen...if someone DID provide the collated data...it would be in a thread....that people could be pointed at. and there will still be people show up...not look to see if it is posted or asked already. annoying, but, that is people.
it's not reinventing the wheel every time...it's pointing them to data...that they can like or not, but, they can't say it was not explained.
TFP could still just explain their reasoning...tell us if it was too resource intensive? or just personal choice...i am sure there was reasoning put into the decision.

 

[Riamus]

If I ask a question to a dev or a moderator, I want them to answer. As you pointed out new people keep showing up, they don't know who is right or wrong. hell...even the long haulers can be accused of not knowing what they think they know. lol
that's why I think people prefer to get "official" answers.

I get why you wouldn't want to go back thru now to look at stuff from years ago...but if it did happen...if someone DID provide the collated data...it would be in a thread....that people could be pointed at. and there will still be people show up...not look to see if it is posted or asked already. annoying, but, that is people.
it's not reinventing the wheel every time...it's pointing them to data...that they can like or not, but, they can't say it was not explained.
TFP could still just explain their reasoning...tell us if it was too resource intensive? or just personal choice...i am sure there was reasoning put into the decision.

It isn't often that you'll find devs responding every time you ask a dev to respond to you in any game. Sometimes they will, but they aren't going to take time to respond every time someone says "I want a dev response." Especially when others have already provided the answers. The moderators generally know the right answers to a lot of topics that people cry for a dev to respond to. It isn't always the case, and they are human just like everyone else (even devs) and can make a mistake or remember something incorrectly. Even a dev could remember the original reasons incorrectly and give a wrong answer. It happens. But in general, you can get valid answers from moderators and others who have been around a long time.

As far as trying to keep track of every post about every change so you can point people to it... who would ever want to do that? You are talking about LBD, but that isn't the only change made to this game. A *lot* of things have changed during development that people complain about. Should someone try to keep track of links to a hundred different topics just to point someone to them if they want to know why something happened? That's kind of crazy. You might only care about LBD, but others care about other things. In order to point people to the topic they care about like you suggest, they'd have to keep track of all potential topics. Before you say just to keep track of certain topics, that helps you but doesn't help people with questions about other topics. Since no one is likely to want to track all topics, there's little reason to track any topics. It's easier to just give the reason and let someone who wants proof to find it themselves.

And as I said, I have no problem with them adding a Community Manager. But if they just give you the same answers that the moderators or other users (or faatal) have given, will you be any more happy? Answers don't change just because a more official person gives the answer. If I ask what kind of fuel is used in a space shuttle and a random person who knows the answer tells it to me, it doesn't change the answer just because the lead person designing the shuttle tells me the same answer even if they are a far more official person. A community manager will also only provide information that the devs choose to make public. You aren't necessarily going to get better communication even with one. You might see the posts that appear on certain social media sites start getting posted everywhere, including here, which would be nice. But you might not get any additional information just because there is a community manager.