Game is turning into a glass cannon only builds
" This is a very rudimentary understanding of statistics if someone is unable to, and actually ignore confounds. As seen by "80% of people only make it to maps based on steam achievements assertion." They arent practicing statistics at that point, they are just kind of making their own parameters on whatever head math they feel like. We are very off course here, but steam achievements aren't polls, they aren't surveys, they aren't a consensus. So, nothing you have put really applies to any claim you have made here. This is why its important, and honestly needs to be a requirement in my opinion, that people take a statistics course. As they hear the terms, like the ones listed by this user, but dont really understand several principles of data analysis as well as all the work that goes into it. Mash the clean
|
|
" You couldn't be more wrong, and this is exactly how modeling and forecasting is done at a professional level. Also I didn't make up the damn steam numbers. They are literally available to anyone that looks. Now you can say you "don't believe" them, but that is a far weaker analysis based on emotion to me. Unless of course you have refuting data with something else concrete, which I'm happy to entertain. Until then you either take the data for what it is, or your refute it based on feels. Either way I'm exhausted at this point saying the same stuff. Have a good night all. I will check in tomorrow if there are some other statistics produced. "Better to remain silent and be thought a fool than to speak out and remove all doubt." - Abraham Lincoln Last edited by DarthSki44#6905 on Sep 5, 2024, 9:33:55 PM
|
|
" You mean the same forecasting done at a professional level that in quantitative methods that relies heavily in historical data and struggles heavily with qualitative factors or forecasting done at a professional level with qualitative methods that relies heavily on human judgment or the forecasting done at a professional level that is often inaccurate as unknown variables occur that invalidade the assumptions of the original forecast with declining accuracy as the time horizon lengthens? The same forecast done at a professional level that made me come home wet today in fact.. Really dude? Steam shows "all time" data, not time stamped. You're the one that's believing at face value what steam is saying, and trying to convince everyone that's the hole truth. Not accounting bots, bans, double accounts, wush, nothing, everything there is real, ignore non-steam, ignoring forecasting variables... If that's what you truly are trying to argue here... Eat your 8 spiders, idc. Ruthless should be [Removed by Support]. Last edited by AdRonZh3Ro#4713 on Sep 5, 2024, 10:25:41 PM
|
|
" Modeling and forecasting is done by a team of people, so no, that's not how that works. And in the "professional" sphere, modeling and forecasting isn't just carried out by a statistician either. You are kind of just making things up now. Which you can look to your picture you posted above for my reaction to your comparison to steam achievement's and your assertions about the "professional" level of modeling and forecasting. Hope that helps! Mash the clean
|
|
|
Im just gonna drop here that when talking about human statistics, using a sample and generalizing the conclusion is inevitable
Saying "you can only make sample studies on non-human stuff" is absolute nonsense. Tons of countries have populations numbering in hundreds of millions, mankind stands at 7 billions. Its literally impossible to have available 100% of the data for anything worth studiyng, If you go by the logic of "you must possess all the data or its invalid" then theres not a single study about behavior, economics, or anything human-related that is valid - as in, not a single one, *literally* not a single one Thing is, people really are more uniform than it seems, taking samples is valid because if the sample is large enough, it becomes representative of the population. "Representative" does not mean it reflects every single person, or even most people(and in fact, studies involving large number of variables sometimes have an average that dont completely represent ANYONE), but it gives a good enough vision of the general profile. And thats good enough because scientists dont want to know about particular individuals, they want to know about the group as a whole And psichology and psychiatrics are both accepted branches of modern medicine, despite having ample "individual" variation. Medicine as a whole is build on top of statistical studies using samples and extending conclusions to mankind as a whole. You might be that one unlucky guy that gets side effects on a medication, that dont mean the study that said the medication was safe was bull As for the steam achieves stuff. To put in plain terms, they dont reflect current situation of the game because they are too global. Poe changed a lot over time and the achieves numbered represents EVERY account ever made on steam, including ones(likely the majority of them) from people that didnt touched the game this league and even people who were shooed away the game when it was young and plain, and never came back, wich means the data is polluted to talk about current situation of the game THAT being said, its not unreasonable to think most players indeed dont touch maps, steam charts do show a very large percentage play the league on the first few days and then never return, its a recurring thing and does make it fair to think the better part of players indeed dont touch maps or do but dont get very far(i would say its not bad to estimate that at least 80% dont reach yellows, but thats a number i pulled from my arse glancing at the charts, only GGG knows the exact proportion) |
|
" I take it you dont bellieve in psychology, but its a pretty much universally accepted science And by that logic of "no samples if human", you must not bellieve in medicine as a whole because taking samples and expanding the conclusion to the whole mankind is how the whole science works |
|
" Yes, i know its 16%, but where you took from that its a bad number? Where you got your ratio that represents "good" retention from? I already asked globaly: What is a good retention number for you? And now i must ask: Where you got that "good" number from? Its because 16% sounds low? Because if that is it, then... thats exactly the sort of "emotional" "living your own world" "antiscientifically" stuff you talked about |
|
" Impossible to have 100%? sure. Ignore 40% of the possible data and utilize heavily deprecated and outright inaccurate data (bots,bans,multi accs), which is what is actually being done here? Yes, invalid. " What are you even talking about? I think you've grossly misunderstood what i've meant about data from materials (qualitative and quantitative) and data from psychology (Its subject matter includes the behavior of humans and nonhumans, both conscious and unconscious phenomena, and mental processes such as thoughts, feelings, and motives) which is highly subjective and subject to bias. Human sampling as in taking each human as an individual sample which is obviously way too much data to be processed. Yes, i worded that poorly there, not my first language (nor second). Ruthless should be [Removed by Support]. Last edited by AdRonZh3Ro#4713 on Sep 6, 2024, 12:05:37 AM
|
|
|
What does player retention have to do with glass cannons. That is another thread.
PoE is free so more people try it and less people play it all the way through. This is expected. Not alarming. The metric to look at is probably those that complete Act 1 and then Act 10. The number is something like 46% for Act 1, and 15% for Act 10; of those that give it a thorough 'try', 33% (15/46 = .33 which is 33%) go all the way. This is overwhelmingly expected. The only reasonable improvement is to make the Act 1 completion something like 50% and the Act 10 completion something like 20%. This would be a retention of 40% or .4 if you only look at those that completed Act 1. You just have to account for botters on multiple accounts though. They may be inflating these numbers. This would mean a false "positivity". I suspect botters may account for 5% of this Act 1 playthrough -- multiplicative, meaning 2.3% of 46%, and since they go all the way to Act 10 it is 2.3% of 15%. The retention would be well over-inflated by just 5% of Act 1 completers (2.3% of the achievement) being bots --- they ALL go to Act 10. Going from 15% to 12.7% is a big drop. About a 15% drop, relatively. Not much else to discuss about retention it belongs in its own thread. Last edited by Ogre_Dylan#3868 on Sep 6, 2024, 1:05:43 AM
|
|
" They start from a reliable data point though and use double blind studies to confirm there is no bias. Methodology is important. The opposite of knowledge is not illiteracy, but the illusion of knowledge.
|
|
