安裝 Steam
登入
|
語言
簡體中文
日本語（日文）
한국어（韓文）
ไทย（泰文）
Български（保加利亞文）
Čeština（捷克文）
Dansk（丹麥文）
Deutsch（德文）
English（英文）
Español - España（西班牙文 - 西班牙）
Español - Latinoamérica（西班牙文 - 拉丁美洲）
Ελληνικά（希臘文）
Français（法文）
Italiano（義大利文）
Bahasa Indonesia（印尼語）
Magyar（匈牙利文）
Nederlands（荷蘭文）
Norsk（挪威文）
Polski（波蘭文）
Português（葡萄牙文 - 葡萄牙）
Português - Brasil（葡萄牙文 - 巴西）
Română（羅馬尼亞文）
Русский（俄文）
Suomi（芬蘭文）
Svenska（瑞典文）
Türkçe（土耳其文）
tiếng Việt（越南文）
Українська（烏克蘭文）
回報翻譯問題
此主題已被鎖定 
討論區規矩與守則
檢舉此篇文章
Leads to drivel in gaming, e-books, youtube videos, websites, music and anything else that can be created in the medium.
Anyone can 'create' now and many do, some are scammers and others plain mediocre/bad.
You just have to filter it out somehow and find what is quality, doing research carefully as I have quite a few indie games that are very positive and I have no idea why.
The Internet has done many good and bad things, one of the worst offences is the pure volume of drivel that has been created by giving the opportunity to self publish anything you want.
And it won't ever stop, it will just keep growing exponentially.
There was a time when every search you did on Google would turn up a Yahoo answers, which basically consisted of someone's opinion rather than fact or non-answers.
I started totally ignoring those search returns and carried on past them.
At some point thankfully they got filtered out, if I do see them I still ignore them as they are so worthless.
"More data" doesn't automatically make it more accurate, particularly when that data still has a set of built-in biases depending on what's being sampled.
It's better to just let the user supply preference data rather than to try to rely solely on the machine to keep guessing it.
...I have no idea where you came up with that first thing.
Also, your suggestion of "play games I don't like" also misses the mark. I like TF2, I like 100% OJ. I like these two multiplayer PVP games.
This does not mean these two are my favorite games, however.
Nor does this not translate into me liking other multiplayer PVP games.
If your machine-learning algorithm is going to try to predict my behavior based on such a black-and-white dichotomy of "user likes X" vs. "user doesn't like X", then you're going to run into problems.
And if we're talking in platitudes and analogies, the picture they paint only illustrates one side of that person.
Pretty much this.
YouTube is also different from Steam in that YouTube videos are (mostly) free to watch, and free to stop watching. There is no investment of money into a purchase. Games are not like this. For games, it's all about getting to a purchase decision.
So the comparison doesn't even work well that well.
As for what YouTube knows, it knows only what I've already shown it. It doesn't know what I haven't yet shown it. Before the first time I watch a video of a guy having a bath with a polar bear, YouTube doesn't know whether I like it or not. Not to mention that YouTube still doesn't know even after I watch it. It can try to guess, based on whether I finished the video or clicked off of it quickly. But as long as I haven't upvoted/downvoted it, and haven't subscribed to the channel, there is no actual confirmation.
That confirmation comes in the form of...user-supplied preference data.
Let me refresh a few questions I asked earlier:
What if there was a great video game crash of 2021? What would cause it, and how (be specific)? What would it look like, and what would its ramifications be?
1. You can't review a product you haven't bought. If you're like any regular person, you'd probably only buy games you expect to like. This means that there's far less data on what you don't like than on what you might like.
2. You can't review products that aren't on Steam. Steam currently has no way to tell that I have an interest in metroidvania randomizers, for example, because all the randos I've played are not on Steam.
On the other hand, if Steam gives me ways to tell it what I like, that'd greatly improve its capability to show me other games I like. For example I could tell it that I like TF2 because it's multiplayer combined with co-op elements and a very silly aesthetic. I could tell it that I don't care for CS:GO because it's got a grittier aesthetic and also has more emphasis on competitiveness.
I'd like to tell Steam these things. There's no other way it can know this from a passive scan of my gameplay on Steam.
"More data" only works when I'm actually playing ALL the games and expressing opinions and feedback on ALL the things. But I'm not. I don't have the time to do that. There's no good comparisons for some things, to see what's statistical baseline/noise and what's not.
In short it wasn't a crash that had anything to do with a drop in demand (heck Demand remained constant) thats why nintendo was able to clean up the way we did because they released into a starved market. It wasn't a lack of supply either. It was simply retailers not wanting to sell.
That literally can't happen again. Especially since digital is such a thing now. Retaileers are a non-issue. Heck they'll probably notice right off after this cocvid thing that the lack of brick and mortar retail did next to didly squat to earnings.
You have to train machine learning on a proper set of data. You have to tell it which of its perceptions are accurate.
At the very least, user-supplied preference data is something the machine can use to confirm the patterns it tries to draw from the data.
And furthermore, user-supplied preference data can be used to inform the machine which patterns to prioritize.
Steam is not the alpha and the omega of all gaming. People don't just game on Steam. Steam at best sees a portion of the user's gaming activity. At best it can attempt to reverse-engineer the user's preferences based on that. And that's what you're expecting it to do.
It'd really help Steam for the user to give it some explicit clues as to what the user likes. Otherwise you're just asking the machine to presume that there's something that can be found by "reading between the lines" of the user's activity, and then try to make sense of that, whatever it may be.