Why are we limited to ignoring a maximum of 10 tags?

[Overwatch]

I have said it before, and I will repeat this for your benefit Steam.

Every inch of your store page is valuable.

Showing me games that I will absolutely, under no circumstance purchase, is money out of your pocket.

Limiting me to 10 tags in my ignore list means that 50% of the games you show me are something I just have to click through. I'm not going to buy them.

And you don't have infinite space. You have a few highlight sections. And they are mostly a waste of space.

If I WANT to go out of my way to find something to buy, then it's not a problem. But that's not how YOU want this to work. You want to "discover" some games to show me, in the hopes that I will buy something on a whim. And with your current system, you are wasting effort/bandwidth/store front space.

The silly thing is, you KNOW the tags I want to ignore. You have a list that suggests tags based on my frequently ignored games. And this list is 22 tags long. You already know I want to ignore all 22 of them.

The 10 I'm ignoring now is the set that includes the highest percentage of suggested games. I can't make it any more efficient for you. And the results are still only 50% effective.

Let us ignore more tags.

You will sell more games if you know what NOT to show us.

[R.4.V.3.N]

Ursprünglich geschrieben von ReamedBySteem:

The tag system is a f*cking mess, because ANY dillweed can add one, and will. How about NO user defined tags, because in the main, users are morons. Let the DEVS add ONE from a generic selection, e.g. adventure, action adventure, adventure stealth, adventure detective, point & click, action rpg, strategy rpg, jrpg, crpg, turn-based strategy, rts, fps, fps military for games like Ghost Recon/Hidden & Dangerous/Operation Flashpoint, arena fps, platformer, metroidvania, puzzle, racing sim, naval sim, flight sim, etc. Tags like 'pixel' and 2d and 3d are superfluous, use your f*cking eyes.

This I can agree with. There should only be main tag with their subcats.

IE: ARMA3 would file as Military Simulator only. CoD would be Military Shooter

ARMA3 Sub-Cats would be Realistic, Workshop, and SP/MP

CoD would be FPS Shooter, No Workshop, MP Only.


It's pretty simple really.

Main + 3 Sub-Catagories is all a game needs for being found. Rather than tons and tons of user generated tags.


Example: Titan Outpost = Main(Third Person RPG) + Sub-Cats(Story Rich + Sci Fi + Colony Manager).

Bam. Makes life easier.

[Tito Shivan]

Ursprünglich geschrieben von Professional ♥♥♥♥♥♥ 6.9:

If you have ten (10) slots, and add ten (10) more, then you'd expect to have 20 potential slots to utilize because that's just how it is....

Math is a universal, fact based language, not an opinion.

1 + 1 = 2

10 + 10 = 20

Maths are indeed awesome:

Do you want 20 tags to filter? You now have twice the chances of any of those tags being on a game you'd be interested in and have it ignored.

30 tags? Thrice that chances.

100 more tags? You're now ten times more likely to have a game you'd like filtered because they happen to have any of your chosen tags from that list.

The filtered tags are a scalpel, not a hammer.

Ursprünglich geschrieben von Solitary Doomsday Cultist:

A lot of the counter arguments also made more sense back when there were fewer tags to sort through. You need to re-balance it all if you decide to increase the quantity in a specific area.

That's why sincethe day the system was released I suggested ignored tags should work in combos.
Not filtering out games having 'anime' OR 'Visual novel' tags (Which is how it works now), but games having 'Anime' AND 'Visual novel' tags.

Ursprünglich geschrieben von ReamedBySteem:

Let the DEVS add ONE from a generic selection.

Just like they have made with genres for decades? Devs will also want to touch all the chords with tags to try attract as much people as possible.

You also need user input to fine-tune tags. Otherwise they end up being as generic and irrelevant as genres.

I agree it's not a perfect system, but Devs aren't beings of light who won't ultimately try to play the tag system in their favour. They've done it for years with genres.

Ursprünglich geschrieben von Professional ♥♥♥♥♥♥ 6.9:

Main + 3 Sub-Catagories is all a game needs for being found. Rather than tons and tons of user generated tags.

Until you don't

What If I want a:
-Third person RPG (main)
-Story-rich
-Sci+fi
-colony manager

But with crafting? Or Zombies? or both?
At the end of the day you either end with genre-like broad categories or have a plethora of sub-characteristics to choose from to try not to end up with a huge list

[cinedine]

Ursprünglich geschrieben von Professional ♥♥♥♥♥♥ 6.9:

Example: Titan Outpost = Main(Third Person RPG) + Sub-Cats(Story Rich + Sci Fi + Colony Manager).

Bam. Makes life easier.

From this "description" I have absolutely no idea what the game is supposed to be. How can an RPG be a Colony Manager? Wouldn't that make it a strategy game? "Colony Manager" has only become a term last year. Before it was "City Builder". Or "Kingdom Builder". Would Anno or The Settler be a Colony Manager? I grew up when they were called economy simulations.
Anyway, also not exactly what I'd call a genre lauded for being "story rich". On the other hand most RPGs are, so actually kinda pointless for them.

If anything we need far more desciptive tags and less genres that can mean anything to anyone. Or assessments like "story rich" or "atmospheric" or "great soundtrack".

Or simply moderate tags instead of having them applied by a community that can't agree on anything. Lately I've seen the 100 Hidden games being categorized as "Creature Collector", which puts them into the same bucket as Pokemon or Ni No Kuni.

[Tito Shivan]

Ursprünglich geschrieben von cinedine:

Or simply moderate tags instead of having them applied by a community that can't agree on anything. Lately I've seen the 100 Hidden games being categorized as "Creature Collector", which puts them into the same bucket as Pokemon or Ni No Kuni.

There's some improvement that could easily be made in that regard.

There should be some sort of automated system which could trigger a review when a discordant tag is applied to a game.

Just like a 'Family friendly' game getting the 'psychological horror' or 'gore' tags. It's an statistical anomaly as you get tags that rarely mix and you get better focus on where to first deploy moderation manhours into reviewing tags.

Also games could have a 'tag blacklist' backend. So once Steam removes the 'gore' tag trolls added on a 'Family friendly' game it cannot be reapplied again or is simply shadowbanned (it can be added but in a non-functional manner).

[cinedine]

Ursprünglich geschrieben von Tito Shivan:

Also games could have a 'tag blacklist' backend. So once Steam removes the 'gore' tag trolls added on a 'Family friendly' game it cannot be reapplied again or is simply shadowbanned (it can be added but in a non-functional manner).

I rather have users doing it being banned from applying further tags, if not done already. Similar to those review trolls.

There was a recent thread about tags being misapplied being a reason to remove a game. Locked by Spawn I think. Valve changed something more or less recently so that the content creators can now apply far more tags than they used to. Leading to such grandious tagging like:
https://store.steampowered.com/app/1789680/Artemis/
With stuff like this, people have to get used to getting ... weird ... results in their discovery queues and no amount of "ignored" tags is going to solve anything.

---

I think a bit more knowledge at how recommendation systems work would be more helpful in such discussions. Each tag on ignore is simply reducing the relevancy score. So once you exhausted the most relevant games, you are desperately looking for a pattern of why it keep showing you ♥♥♥♥ you are not interested in and inevitably add more and more tags to the ignore list which isn't going to help.

It's also quite helpful to know when to stop using it. I.e. when you exhausted the most relevant options and start seeing more and more "because it's popular" and "just to see whether you like it" or generalisations like "because you played games tagged with Indie".

I can firmly recommend https://steampeek.hu They seem to have a better weighing of tags while ignoring popularity to make more useful recommendations coming from a specific game or a number of games you like.

[Tito Shivan]

Ursprünglich geschrieben von cinedine:

Ursprünglich geschrieben von Tito Shivan:

Also games could have a 'tag blacklist' backend. So once Steam removes the 'gore' tag trolls added on a 'Family friendly' game it cannot be reapplied again or is simply shadowbanned (it can be added but in a non-functional manner).

I rather have users doing it being banned from applying further tags, if not done already. Similar to those review trolls.

Both things are not mutually exclusive. They're different forms of applying friction to toxic users. On one side you ban users from keeping up with the trolling and also make pointless the actions of new trolls by ignoring the application of these tags.

Ursprünglich geschrieben von cinedine:

I think a bit more knowledge at how recommendation systems work would be more helpful in such discussions. Each tag on ignore is simply reducing the relevancy score. So once you exhausted the most relevant games, you are desperately looking for a pattern of why it keep showing you ♥♥♥♥ you are not interested in and inevitably add more and more tags to the ignore list which isn't going to help.

That's why I insist on providing positive feedback to these systems will always work better instead of negative one.

You're always going to have more stuff you don't like than stuff you do. It's easier if you point to the things you like than to everything else you don't

And that's why people make wishlists for gifts and purchases and not 'I-Don't-Want-This-Gifted-Ever-List'

Ursprünglich geschrieben von cinedine:

It's also quite helpful to know when to stop using it. I.e. when you exhausted the most relevant options and start seeing more and more "because it's popular" and "just to see whether you like it" or generalisations like "because you played games tagged with Indie".

It's a tool at the end of the day. When using tools one does need to be careful of not falling into the 'I have a hammer, everything is a nail' mindset.

[Start_Running]

OP. The reason is that Valve is pushing you to tell the system what you like!
Also due to the nature of games beyond a certain point you simply will see zero games. And I think 10 is pretty close to that point.

Look if you have that narrow of an interest zone (nothing wrong with that), maybe you should be using curators to find your games.

[crunchyfrog]

Ursprünglich geschrieben von Tito Shivan:

Ursprünglich geschrieben von Overwatch:

So...You believe that a suggestion engine that shows me products that are "similar" to 1000 other ones that I have manually ignored is working properly?

First read what I wrote in my post. The recommendation engine is meant to open the users views. To show them things that go outside of 'just the things I like'.

Second. The main problem of adding too many tags to the recommendation engine is the more you add, the (way) faster you run out of games to show (which may even interest you) just because it has one of the 20, 30 or 100 tags you've blocked.

If you start forbidding too many ingredients, you run out of things to cook. Even of things you like to eat.

Ursprünglich geschrieben von Overwatch:

The fact that I have to 'waste' 6ish blocked tags just to keep sports and fighting games out of my suggestions is a prime example. They are both popular genres, and I know I'm not ever going to consider playing them, even if they are free. So they are a waste of time for me, and a waste of money for Steam. (It costs money to show me, and they take up space where a game I might buy could be.)

I've suggested in the past that instead of just tags, the system could let users apply 'combos' of tags. That could allow for fin tunning the content you want to block

Ursprünglich geschrieben von Overwatch:

I see a LOT of suggestions to just search for games. But that is totally unfeasible. There are so many games available that searching is not a solution.

I've suggested the search for users who have a very narrow interest in games. And it's easy to fine tune the search to a handful of results when your interests can be narrowed enough

Ursprünglich geschrieben von Overwatch:

I do spend some time browsing major categories, and I have made some interesting finds. But that is not even close to a good replacement for a properly working suggestion system.

then you need to refine your search. Any major category is going to be full to the brim of content.

Ursprünglich geschrieben von Overwatch:

I already know that suggestion systems work poorly for me. That is why it is important to let users themselves have input. Ignoring tags is a good way to do it. It is low effort on Steam's part, and the 10 tags DO currently help.

But 10 is not enough.

But adding more isn't going to bring the results you expect from the system.

Exactly.

That'[s what I meant about my "6 degreses of Kevin Bacin" thing.

In just 6 steps EVERYBODY can link to Kevin Bacon. So you don't need more than 6 steps there.

The same applies here.

[Overwatch]

Ursprünglich geschrieben von Tito Shivan:

...The recommendation engine is meant to open the users views. To show them things that go outside of 'just the things I like'.

You still don't seem to get my point. So I'll give you a specific example of the problem that neatly shows both my annoyance and a waste on Steam's effort.

Before I switched my tags to my current list, I wasn't ignoring fighting games.
I was, however, ignoring ~15 publishers that exclusively made fighting games.
I had also manually ignored every major fighting game. And all the DLC for those games.

Steam 'recognized' that I was not interested in fighting games. I know this because in the settings page where you can view/review/change the tags you are ignoring, it suggests tags to ignore based on your game history + your manually ignored games. "Fighting Games" was on the list of 5 tags it was suggesting I ignore. (And also the only one I wasn't ignoring at the time out of the 5...)

Yet, I still got a discovery queue where 4 of the 10 suggested games were new/popular fighting games that had just released.

That is a bad system.

There is a difference between suggesting things outside the ones you know someone likes, and suggesting things that someone has EXPLICITLY said they don't like.

At this point I don't even think you and I are talking about the same thing. I can't speak to your experience with these things, but, as I've mentioned, I have real-world developer experience with suggestion systems. (I have no idea if any of my code is still live, but I know that it was for quite a while.)

My requests are neither unreasonable nor un-doable. They are objectively a net gain for all parties.

[Overwatch]

Ursprünglich geschrieben von Start_Running:

OP. The reason is that Valve is pushing you to tell the system what you like!

There is literally no way to tell a system that you "like" a tag.

You can buy games.
You can play games.

Those are the only positive indicators.

Ursprünglich geschrieben von Start_Running:

...if you have that narrow of an interest zone (nothing wrong with that), maybe you should be using curators to find your games.

My interests aren't narrow. I play all SORTS of games.

I just have some game types that I know I'll never play. The list is pretty narrow actually.

The problem (as many of us have stated over and over) is that 10 tags isn't enough due to the way the tag system works. It takes 6 tags for me to block Sports and Fighting games. That leaves me 4 leftover tags, which I happen to be using to block super-low effort garbage or shovelware like asset flips and 'hidden item' games.

That's it. That's my whole ignore list right now.
I'd love to be able to ignore 2ish more groups of games, but I can't.


Imagine if a service like Spotify only let you ignore 10 specific sub-categories of music and you had to spend multiple tags to ignore Dark Metal, Black Metal, Death Core, and Grind Metal.

You'd probably be asking for more than 10 tags too...

[R.4.V.3.N]

Ursprünglich geschrieben von Overwatch:

Ursprünglich geschrieben von Start_Running:

OP. The reason is that Valve is pushing you to tell the system what you like!

There is literally no way to tell a system that you "like" a tag.

You can buy games.
You can play games.

Those are the only positive indicators.

Ursprünglich geschrieben von Start_Running:

...if you have that narrow of an interest zone (nothing wrong with that), maybe you should be using curators to find your games.

My interests aren't narrow. I play all SORTS of games.

I just have some game types that I know I'll never play. The list is pretty narrow actually.

The problem (as many of us have stated over and over) is that 10 tags isn't enough due to the way the tag system works. It takes 6 tags for me to block Sports and Fighting games. That leaves me 4 leftover tags, which I happen to be using to block super-low effort garbage or shovelware like asset flips and 'hidden item' games.

That's it. That's my whole ignore list right now.
I'd love to be able to ignore 2ish more groups of games, but I can't.


Imagine if a service like Spotify only let you ignore 10 specific sub-categories of music and you had to spend multiple tags to ignore Dark Metal, Black Metal, Death Core, and Grind Metal.

You'd probably be asking for more than 10 tags too...

This tho.. 6 tags for the sports/fighting, need 6-9 more for all the Anime/Hentai scam ♥♥♥♥, need more for all the Clicker/Idle games, need more for all the (we can go for a while on this.).

There's too many tags and sadly a ton of the low effort games just slap tons of tags on their page, on top of community made tags which...

It's the steam community, why did valve think the general public can tell the difference between Military Simulator and Colony Simulator??? Most people don't even realize what they are playing 90% of the time cause CoD taught them "Shiney Graphics and Trailer tho".

So not only do we need way more tag slots, (like 30), Steam also needs a major overhaul on tags in general, remove that "Community Tags" crap, and have a locked in system for games that removes games that devs wrongly tag or maybe even fines them for mis-tagging (Which imho counts as dishonest marketing).

[cinedine]

Ursprünglich geschrieben von Overwatch:

Yet, I still got a discovery queue where 4 of the 10 suggested games were new/popular fighting games that had just released.

That is a bad system.

Not really, it's just how the system works. You have reached a point where it is no longer useful.

Every game is given a relevancy score.
If it shares tags with games you often play it will get a bonus.
If it is popular among friends or got good reviews from them, it will get a bonus.
If it is popular among the community, it will get a bonus.
If it is on sale, it will get a bonus.
If it shares tags on your ingore list, it will get malus (usually way more than any of the above bonui).
If it has bad reviews, it will get a malus.

The games are than sorted by their relevancy score.
The queue will never not show you games unless you have completely exhausted the store.

The last one is the main difference from other recommendation systems, like e.g. on Amazon if you look for "red t-shirt". There is a cut-off point. (Also Amazon is hilariously "bad" too, try putting a faceroll string into search.)
Over time you will run into games less and less relevant and the games with tags on your ignore list will keep showing up, because the threshhold is so low, that the malus isn't enough anymore.

Steam has a limited amount of games and new releases and the vast majorit of them are not of interest for the individual user. Rule of thumb "90 % of everything is ♥♥♥♥" applies, even if you have a rather broad interest. So it's a matter of how often you use the discovery queue until it becomes useless. From my experience it's once a week and later once every other week.

I assume one of the reasons Steam has changed the free cards for using the queue during sales to one instead of three is because many users have reached the point of the queue becoming less usefull.

[R.4.V.3.N]

Ursprünglich geschrieben von cinedine:

Ursprünglich geschrieben von Overwatch:

Yet, I still got a discovery queue where 4 of the 10 suggested games were new/popular fighting games that had just released.

That is a bad system.

Not really, it's just how the system works. You have reached a point where it is no longer useful.

Every game is given a relevancy score.
If it shares tags with games you often play it will get a bonus.
If it is popular among friends or got good reviews from them, it will get a bonus.
If it is popular among the community, it will get a bonus.
If it is on sale, it will get a bonus.
If it shares tags on your ingore list, it will get malus (usually way more than any of the above bonui).
If it has bad reviews, it will get a malus.

The games are than sorted by their relevancy score.
The queue will never not show you games unless you have completely exhausted the store.

The last one is the main difference from other recommendation systems, like e.g. on Amazon if you look for "red t-shirt". There is a cut-off point. (Also Amazon is hilariously "bad" too, try putting a faceroll string into search.)
Over time you will run into games less and less relevant and the games with tags on your ignore list will keep showing up, because the threshhold is so low, that the malus isn't enough anymore.

Steam has a limited amount of games and new releases and the vast majorit of them are not of interest for the individual user. Rule of thumb "90 % of everything is ♥♥♥♥" applies, even if you have a rather broad interest. So it's a matter of how often you use the discovery queue until it becomes useless. From my experience it's once a week and later once every other week.

I assume one of the reasons Steam has changed the free cards for using the queue during sales to one instead of three is because many users have reached the point of the queue becoming less usefull.

Either way tho it's still a massive pain for us looking for very very specfic games.

IE: I currently play Titan Outpost and Project Zomboid. Both of which are gems imho.

Problem is, if I search tags specific to them I still gotta look through 99% fluff and filler, and in the past 2 months of looking I haven't found anything that actually matches those two games in complexity, design, type, genre.

I mean FFS I have Anime and Dating sims Excluded and Steam still shows me that trash...

[Overwatch]

Ursprünglich geschrieben von cinedine:

Ursprünglich geschrieben von Overwatch:

...That is a bad system.

Not really, it's just how the system works.

What?

If a system fails to achieve it's purchase, it is bad. Period.

• Steam sell games.
  
• System put games in front of eyes.
  
• Does suggestion result in sale?
  
  • Yes - System is good.
    
  • No - System is bad.

[R.4.V.3.N]

Ursprünglich geschrieben von Overwatch:

Ursprünglich geschrieben von cinedine:

Not really, it's just how the system works.

What?

If a system fails to achieve it's purchase, it is bad. Period.

• Steam sell games.
  
• System put games in front of eyes.
  
• Does suggestion result in sale?
  
  • Yes - System is good.
    
  • No - System is bad.

In some peoples cases, the system just doesn't work at all, it's beyond bad.

If you're someone into very niche game types, you're basically only gonna purchase a game once every several months.