It seems like this could be massively improved by not including play sessions that’re less than some baseline duration, only including data where people have more than say, 2 hours of total playtime, and throwing out outlier data.
It’s fairly impossible to get perfect data here, but Valve has historically been pretty good about iterative improvements.
It seems like this could be massively improved by not including play sessions that’re less than some baseline duration, only including data where people have more than say, 2 hours of total playtime, and throwing out outlier data.
It would still depend on what’s actually being rendered at that time, before I upgraded my GPU I would abuse Sky Boxes knowing I can increase my FPS by simply looking up or away from geometry.
Yeah, it wouldn’t be perfect, but I’d wager the majority of players wouldn’t bother doing that, so by removing outliers (maybe the top/bottom 10%), you’d probably mostly remove the effects of people doing non-standard things. And by requiring 2h (or some baseline amount) of playtime, you’d remove people who launched the game, spent most of their time on the menus, started the game once, decided it wasn’t for them and quit, as they’d likely have a higher average FPS. It’d be better, at least.
I don’t think it’s “more harm than good”, just need some changes
I guess they’ll soon see that it can be improved and they probably will.
Still, it’ll never be perfect and that’s why a website like SteamDeckHQ is great and should be supported.
If you want accurate numbers, then visit benchmarks. This is just a rough estimate and I think its doing more good than harm. Its not perfect, but I do not agree with the conclusion. If many people get a 60 fps, then at least you know it is possible. And most people do not fiddle long time with the settings. It’s just another tool next to the verified rating.
It’s in Beta right now, and Valve expects discussions. But even in the current form, I think this is better than not having this.
You know its possible but you lack the context. Where is it possible? Loading screens, menus and cut scenes or actual gameplay?
I think it’s especially bad that they show average FPS. Final Fantasy 10 has average 33 fps and is a perfect experience. Final Fantasy 16 has 42 FPS average and it’s terrible, even with some people hitting 90 FPS.
They just don’t average number, but have a graph. And they have it for 60Hz LCD and 90Hz OLED listed separately. We knew about this feature before it came out officially (could opt in for data collection). And we feared the same thing you say, but Valve did actually solve it by providing a graph. A game that is mostly stable will sit in one number mostly. Games that have a bit problematic FPS will be lower and all over the place. It’s still a rough estimate, and I don’t think a more detailed view is not needed nor possible, because people play with different configuration. This is not meant to be a perfect number, its just to give an idea how it performs.
It’s just a bit hidden, that’s the problem:
The graph helps but, like you said, it is yet another click away. Before you see the graph you’re already conditioned by the “42 FPS average” information. You see the FPS dipping in the graph but you’ll think that’s an outlier. The average is 42 FPS after all, that’s nice amirite.
Personally, I feel like the solution is that the
Steam Verifiedbadge should mean something. Solid 30 FPS should be the golden standard. If the game is frequently at 15-20 FPS then it should be “Playable”, not “Verified”. The current criteria where typing your name with a virtual keyboard means “Playable” but running at 15 FPS is “Verified” is not helpful.These charts are good for those games that are incorrectly marked as borked or don’t have a verified status. You can now see evidence that people are playing the game, and that’s nice.
it is yet another click away
I assume because it is in Beta stage. My hope is, that there is a setting in Steam to enable or disable this graph directly on the game page (in web browser and in the client).
Personally, I feel like the solution is that the Steam Verified badge should mean something. Solid 30 FPS should be the golden standard.
It actually means something. When the Steam Deck launched, an employee explained it means something like “the game should run like it would if this was on similar hardware playing on Windows”, from performance standpoint. And off course it should be an experience like a native console game with default settings working right away. Also the “playable” and “acceptable” tolerance is different for everyone.
And that’s the key point, it should behave like on consoles. I mean games on console are basically verified for that system, right? I mean any game you can purchase on Playstation 4 is verified for that system. It does not make any guarantees about performance or quality. And this is how I view the verification too. There are extreme cases, but that’s not what we talk about.
Some would argue, 40 FPS should be the golden standard on Deck in example, instead 30. There are games which have a hard time to stay at 30 FPS, like Baldur’s Gate 3, but are played a lot of the system.
It doesn’t matter what a Dev said a long time ago, it matters what they advertise. They advertise that a steam deck verified game “performs well on the Steam deck”. I know well is a broad term, but anything less than solid 30FPS should be immediately “Playable”.
Games like Final Fantasy 16 that frequently dip to 15 FPS cannot be marked as Verified. If they handed these badges correctly, there wouldn’t be a need for community based fps history or even protondb.
It doesn’t matter what a Dev said a long time ago
It matters, because that is an explanation to how it works. This (what I explained) explains it well. If you choose to ignore the logic, then all your argumentation makes no sense. Because you ignore what the verification actually means.
They advertise that a steam deck verified game “performs well on the Steam deck”.
Which is inline with what I said. It does not contradict to what the developer said.
I know well is a broad term, but anything less than solid 30FPS should be immediately “Playable”.
I strongly disagree. First it depends on what game. For a fighting game like Street Fighter or Tekken where all versions of the game MUST be 60 FPS to be on the same playing field, I can understand that 30 FPS leads to just a “Playable” rating. But if we talk about a strategy game like Civilization where FPS is not that important, or maybe a JRPG like Final Fantasy, a 30 FPS should not lead to just being Playable rating. That would be wrong.
Games like Final Fantasy 16 that frequently dip to 15 FPS cannot be marked as Verified.
Just because it sometimes goes below the mark does not disqualify the Verified rating. As explained in my previous reply. I think many (including you) misunderstand the Verification system with a Quality or Performance system. If the game however is constantly below 30 FPS and the FPS is crucial for the game, then I would understand it. But dipping here and there does not disqualify in my opinion. I think Valve handles the verification correctly.
It doesn’t matter because consumers aren’t going to hunt for Dev statements from years ago. They click the badge, the badge explains the criteria and that should be it.
I’m not saying all games with 30 fps should be verified. I’m saying all games below 30 fps should not be deck verified. That’s a very different statement.
You say I misunderstand the badge as a quality/performance metric but that is exactly what they say it is. I’m quoting them.
Some good points here. This makes more sense if the metrics are taken when based on the default graphics settings that are triggered when a game detects that it’s running on a SteamDeck.
And some game locations can absolutely skew the results. Including the menu FPS will also be misleading. Still, I think it will be useful to some degree.
I feel like there could be really good data from this framerate pool, but it needs to be sorted in some way.
For example, you could probably split it into “target” fps categories based on where that player’s majority fps is. Knowing that someone playing at 30fps target is seeing sustained drops below 25fps is more telling than knowing that the average fps is 37.
