[DISCUSSION] How Do You Know if Reviews are Fake? : Steam

[<™[-=> Cookiez <=-]™>]

If you look at the Steam reviews for certain games, some of the reviews are just 1 time deals.

The account has only 5 games, 4 free and only 1 purchased game which they reviewed.

All the review accounts have only 1 review in their entire existence.

Do you think it's possible that the reviews are fake?

[You, Yourself]

Originalmente postado por crunchyfrog:

Originalmente postado por Uncopyrightable Moonkey:

Here's an idea about how to solve this review reliability problem on Steam. Steam can only allow a review if the user has spent 2 hours into that game and as 2 hours also is the threshold of a refund, Steam can in the meantime start stating the number of refunds and also the number of purchases besides recent reviews and all reviews. I think the necessity to spend 2 hours in the game basically eradicates the possibility of fake reviews, the number of purchases and the number of refunds are also good indicators about the the game quality, the latter of which may count worse than a negative review.

An absolutely terrible idea.

I mean no offence to you by this, but there are strict laws about data protection and what identifiable or useful info you can turn out publicly. I can think of loads of reasons people would NOT want that data out there, and I doubt it would make any real difference anyway.

But worst of all, Valve would NEVER do this. Simply because in doing so, one could work out the threasholds to maximising refunds.

I don't get your idea, Steam is already collecting running times and I'm not proposing that they share anything personal, just collective numbers of their business, I bet numbers like that are already being shared and discussed in business settings. How would one maximise refunds and why is that bad ?

[<™[-=> Cookiez <=-]™>]

Originalmente postado por Uncopyrightable Moonkey:

Originalmente postado por crunchyfrog:

An absolutely terrible idea.

I mean no offence to you by this, but there are strict laws about data protection and what identifiable or useful info you can turn out publicly. I can think of loads of reasons people would NOT want that data out there, and I doubt it would make any real difference anyway.

But worst of all, Valve would NEVER do this. Simply because in doing so, one could work out the threasholds to maximising refunds.

I don't get your idea, Steam is already collecting running times and I'm not proposing that they share anything personal, just collective numbers of their business, I bet numbers like that are already being shared and discussed in business settings. How would one maximise refunds and why is that bad ?

So, you want the average amount of time played for each game?

I've mentioned this a long time before.

[crunchyfrog]

Originalmente postado por Uncopyrightable Moonkey:

Originalmente postado por crunchyfrog:

An absolutely terrible idea.

I mean no offence to you by this, but there are strict laws about data protection and what identifiable or useful info you can turn out publicly. I can think of loads of reasons people would NOT want that data out there, and I doubt it would make any real difference anyway.

But worst of all, Valve would NEVER do this. Simply because in doing so, one could work out the threasholds to maximising refunds.

I don't get your idea, Steam is already collecting running times and I'm not proposing that they share anything personal, just collective numbers of their business, I bet numbers like that are already being shared and discussed in business settings. How would one maximise refunds and why is that bad ?

You are so missing the point.

One of the common things that goes on Steam is fraud and exploitation. ANYTHING that has value, well people will try to either exploit, shortcut or otherwise cheese it.

We've seen it many times over sale periods (like the dreaded coal Christmas sale period where all the games in the pool were exhausted) to scams and so on.

The important thing you're missing here is something we see on these forums a lot. This simple question - how many refunds do I need to do before I get banned from refunding?

The answer is obvious. You're NEVER going to know because that would tell anyone that wants to exploit that HOW to exploit it. The calculation they used is likely based off number of games you buy, percentage you return and a lot more things. We can never know.

So let's take your proposal. That game times and refunds are listed on reviews. Can you not see the issue there? Any scammer would IMMEDIATELY correlate returns with how many games the person has in their library and work out how it works.

Valve would NEVER do that.

On top of that, I don't know of what possible use this would be. People can still lie and do all the other problematic things regardless of this info so it doesn't offer anything helpful.

[You, Yourself]

Originalmente postado por crunchyfrog:

Originalmente postado por Uncopyrightable Moonkey:

I don't get your idea, Steam is already collecting running times and I'm not proposing that they share anything personal, just collective numbers of their business, I bet numbers like that are already being shared and discussed in business settings. How would one maximise refunds and why is that bad ?

You are so missing the point.

One of the common things that goes on Steam is fraud and exploitation. ANYTHING that has value, well people will try to either exploit, shortcut or otherwise cheese it.

We've seen it many times over sale periods (like the dreaded coal Christmas sale period where all the games in the pool were exhausted) to scams and so on.

The important thing you're missing here is something we see on these forums a lot. This simple question - how many refunds do I need to do before I get banned from refunding?

The answer is obvious. You're NEVER going to know because that would tell anyone that wants to exploit that HOW to exploit it. The calculation they used is likely based off number of games you buy, percentage you return and a lot more things. We can never know.

So let's take your proposal. That game times and refunds are listed on reviews. Can you not see the issue there? Any scammer would IMMEDIATELY correlate returns with how many games the person has in their library and work out how it works.

Valve would NEVER do that.

On top of that, I don't know of what possible use this would be. People can still lie and do all the other problematic things regardless of this info so it doesn't offer anything helpful.

That could only reveal the average number of refunds compared to a given number of purchases, a number like 10 refunds per 100 purchases; and might only encourage refunds up to that ratio plus an expected deviation of refunds to purchases. On the other hand it shows the trustworthiness of Steam, encouraging more people to use Steam, and people to use Steam more.

[You, Yourself]

Originalmente postado por <™-=> Cookiez <=-™>:

Originalmente postado por Uncopyrightable Moonkey:

I don't get your idea, Steam is already collecting running times and I'm not proposing that they share anything personal, just collective numbers of their business, I bet numbers like that are already being shared and discussed in business settings. How would one maximise refunds and why is that bad ?

So, you want the average amount of time played for each game?

I've mentioned this a long time before.

That's a good idea, yes I would support its inclusion on Steam pages.

[crunchyfrog]

Originalmente postado por Uncopyrightable Moonkey:

Originalmente postado por crunchyfrog:

You are so missing the point.

One of the common things that goes on Steam is fraud and exploitation. ANYTHING that has value, well people will try to either exploit, shortcut or otherwise cheese it.

We've seen it many times over sale periods (like the dreaded coal Christmas sale period where all the games in the pool were exhausted) to scams and so on.

The important thing you're missing here is something we see on these forums a lot. This simple question - how many refunds do I need to do before I get banned from refunding?

The answer is obvious. You're NEVER going to know because that would tell anyone that wants to exploit that HOW to exploit it. The calculation they used is likely based off number of games you buy, percentage you return and a lot more things. We can never know.

So let's take your proposal. That game times and refunds are listed on reviews. Can you not see the issue there? Any scammer would IMMEDIATELY correlate returns with how many games the person has in their library and work out how it works.

Valve would NEVER do that.

On top of that, I don't know of what possible use this would be. People can still lie and do all the other problematic things regardless of this info so it doesn't offer anything helpful.

That could only reveal the average number of refunds compared to a given number of purchases, a number like 10 refunds per 100 purchases; and might only encourage refunds up to that ratio plus an expected deviation of refunds to purchases. On the other hand it shows the trustworthiness of Steam, encouraging more people to use Steam, and people to use Steam more.

Yes,a nd even that could be pored over to get MORE of an idea than we do now. That's the point.

If you don't think that will happen you haven't been paying attention on STeam to what has happened over the years.

Valve will NEVER go for that.

And again, you seem to like this idea, but there's no benefit in it being there. Can people still lie or make daft reviews with this metric being shown? Yup.

THen it's moot.

[<™[-=> Cookiez <=-]™>]

Originalmente postado por Uncopyrightable Moonkey:

Originalmente postado por crunchyfrog:

You are so missing the point.

One of the common things that goes on Steam is fraud and exploitation. ANYTHING that has value, well people will try to either exploit, shortcut or otherwise cheese it.

We've seen it many times over sale periods (like the dreaded coal Christmas sale period where all the games in the pool were exhausted) to scams and so on.

The important thing you're missing here is something we see on these forums a lot. This simple question - how many refunds do I need to do before I get banned from refunding?

The answer is obvious. You're NEVER going to know because that would tell anyone that wants to exploit that HOW to exploit it. The calculation they used is likely based off number of games you buy, percentage you return and a lot more things. We can never know.

So let's take your proposal. That game times and refunds are listed on reviews. Can you not see the issue there? Any scammer would IMMEDIATELY correlate returns with how many games the person has in their library and work out how it works.

Valve would NEVER do that.

On top of that, I don't know of what possible use this would be. People can still lie and do all the other problematic things regardless of this info so it doesn't offer anything helpful.

That could only reveal the average number of refunds compared to a given number of purchases, a number like 10 refunds per 100 purchases; and might only encourage refunds up to that ratio plus an expected deviation of refunds to purchases. On the other hand it shows the trustworthiness of Steam, encouraging more people to use Steam, and people to use Steam more.

Yes, revealing the number of refunds would be better.

That's why you need to weight the credibility of the reviewer as well.

There are a few factors you need to weigh.

Play time alone is insufficient as I mentioned that you can idle and gain hours.

[Sp1cy]

very good

[<™[-=> Cookiez <=-]™>]

Originalmente postado por d:

very good

Might as well mention that you can use achievements for weighing, length of review (assuming no lorem ipsum gibberish), signs of financial shilling.

These improve the quality of the review.

[Start_Running]

Originalmente postado por <™-=> Cookiez <=-™>:

Originalmente postado por d:

very good

Might as well mention that you can use achievements for weighing, length of review (assuming no lorem ipsum gibberish), signs of financial shilling.

These improve the quality of the review.

Again. What does it matter?
At somepoint you're going to have to judge the review based on its information content.

[crunchyfrog]

Originalmente postado por <™-=> Cookiez <=-™>:

Originalmente postado por Uncopyrightable Moonkey:

That could only reveal the average number of refunds compared to a given number of purchases, a number like 10 refunds per 100 purchases; and might only encourage refunds up to that ratio plus an expected deviation of refunds to purchases. On the other hand it shows the trustworthiness of Steam, encouraging more people to use Steam, and people to use Steam more.

Yes, revealing the number of refunds would be better.

That's why you need to weight the credibility of the reviewer as well.

There are a few factors you need to weigh.

Play time alone is insufficient as I mentioned that you can idle and gain hours.

Excep t that it'll NEVER happen for the precise reasons I've detailed.

If you think Valve are going to offer anything that will only serve to help people abuse the refund system, you are frankly out of your mind.

We have several examples of what happened on Steam when Valve overlooked the ability for people to exploit stuff. Several sales, for a start. Not least of which, the reason we cooldowns for trading and $5 "entrance" to the unlocked account is for these very reasons.

It's never going happen, and I don't see how it would help in the slightest like you think it will.

How can it possibly? Can people still lie, get things wrong and detail things in an incorrect manner with that data being there? Yup.

So it's a useless factor. But anyway can't happen.

[You, Yourself]

Originalmente postado por Start_Running:

Originalmente postado por I>U:

It's 2 big differences when someone wanted to share his own opinion - even if that turned to be a false one - and tried to rise a hype wave to tighten up people suspectable to it.

In other words, it's not as easy as you've just described.

Its actually easier. If one does not over complicate the matter.. Or to put it another way. Focus more on 'why?' rather than 'what?'.

Fake reviews is a thing on many platforms, so buying a 60 USD game is not as big of a deal as buying 400 USD monitor that mr. reviewer proclaimed to have 12bit matrix, while in reality it's 6bit+FBC for that ridiculous price. And you can only learn it by playing detective and help others by naming the false "facts".

That ones actually fairly easy. If the reviewer mentions a major feature that that the manufacturer themselves doesn't list...something is off and paid reviewers are usually given a script of things to mention so the reviews line up with the actual feature advertised.

The counter to this is to read both positive and negative reviews. If a game lacks negative reviews... well that's a warning flare for me.

So, no, OP: if you are interested in platform, do your research and - if you feel generous - share the results.

Research is of course a good thing. The more data sources the better and uits worth noting that this is one of the reasons I will NEVER Pre-Order a game, or purchase a game at launch. I prefer to wait a month.

And I consider heavy marketing pushes towards pre-orders and launch day purchases to be a giant red-flag. These are the time frames where you have the least available information.

You're being overly pessimistic about it. Steam's refund policy is not secret so the number of refunds may become public as well. Average running time and weighing reviews based on achievements or running time where achievements do not apply may also be useful. I think Steam would consider only one thing before making one of these decisions and that is whether the Steam users demand it or not.

[<™[-=> Cookiez <=-]™>]

Originalmente postado por Uncopyrightable Moonkey:

You're being overly pessimistic about it. Steam's refund policy is not secret so the number of refunds may become public as well. Average running time and weighing reviews based on achievements or running time where achievements do not apply may also be useful. I think Steam would consider only one thing before making one of these decisions and that is whether the Steam users demand it or not.

On a side note, Steam needs to redo their quotes with collapsible functions.

I don't want to keep deleting text when I reply, which will result in lost meanings.

Achievement weighting is correct but it's cheatable.

If I remember correctly, you can fake achievements.

The number of proper reviews should also be weighted as well. e.g. If someone just wrote 1 review, it's obviously not the same as someone who wrote 100 reviews. They could even be a shill account.

[JD777]

Human to fallible , Human to emotional , Human to obsolete , Human no 100% logic. Me trust algorithms and AI to review game because me no not how to think and it be much easy to have thinks for me. That is all brain hurts much .

[crunchyfrog]

Originalmente postado por Uncopyrightable Moonkey:

Originalmente postado por Start_Running:

Its actually easier. If one does not over complicate the matter.. Or to put it another way. Focus more on 'why?' rather than 'what?'.


That ones actually fairly easy. If the reviewer mentions a major feature that that the manufacturer themselves doesn't list...something is off and paid reviewers are usually given a script of things to mention so the reviews line up with the actual feature advertised.

The counter to this is to read both positive and negative reviews. If a game lacks negative reviews... well that's a warning flare for me.


Research is of course a good thing. The more data sources the better and uits worth noting that this is one of the reasons I will NEVER Pre-Order a game, or purchase a game at launch. I prefer to wait a month.

And I consider heavy marketing pushes towards pre-orders and launch day purchases to be a giant red-flag. These are the time frames where you have the least available information.

You're being overly pessimistic about it. Steam's refund policy is not secret so the number of refunds may become public as well. Average running time and weighing reviews based on achievements or running time where achievements do not apply may also be useful. I think Steam would consider only one thing before making one of these decisions and that is whether the Steam users demand it or not.

Really?

Then demonstrate what this forumla is for refunds.

You're not understanding that by providing MORE evidence than there is now, it would demonstrably hurt things. I made this very clear why and gave you examples of when it had happened before.

Valve are NOT going to do this. I will lay good money on that fact.

And if you think that Valve will JUST look at whether people want it versus encouraging fraud and more exploitative actions, then you should never ever own a company or be in a position of authrotiy because it simply does not work like that anywhere.