Valve Hands Devs FPS Data to Fix Shoddy Deck Games

"We're providing this data because while customers overwhelmingly agree with the Verified rating for titles (>95%), it can be valuable for developers to better understand the experiences of customers who disagree, especially in the context of a specific title or update." That line, buried in a Steamworks blog post published on April 23, is Valve quietly admitting what anyone who owns a Steam Deck already knows: the Verified badge doesn't always mean what it's supposed to mean.

The two new tools Valve is rolling out are straightforward. The first gives developers a chart showing the rolling 30-day average framerate for their game on Steam Deck, pulled from players who opted into anonymous performance data collection. Valve says it will expand this to include variance data, so devs can see whether their game holds a steady 30 fps or lurches between 20 and 40. The second tool surfaces results from player surveys. After someone plays a Verified game for at least 10 minutes on Steam Deck, Steam can ask them whether they agree with the Verified status. If they disagree, they pick from a list of reasons including performance and stability. Developers now get to see those responses.

Both features are currently limited to games that already carry the Verified badge, with Playable-rated titles coming later. I think the Playable tier is where this data will matter most, since those games are already flagged as having rough edges and developers would benefit from knowing exactly which edges are roughest.

The 95% Problem

Valve leading with that 95% agreement figure feels like a preemptive defence of the Verified system, and I'm not entirely buying it. A 95% satisfaction rate sounds great until you remember that the 5% who disagree are probably concentrated around a handful of high-profile games that have no business wearing the badge. We've all seen Verified titles that drop into the low twenties in busy scenes. A game that technically boots and accepts controller input is not the same as a game that runs well, and the Verified label has blurred that line for too long. Giving developers the actual framerate numbers, not just a pass/fail badge, is the right move. It shifts the conversation from "does this game meet Valve's minimum bar" to "is this game actually performing well for real players on real hardware."

The timing here matters too. Valve has the Steam Machine console and Steam Frame VR headset on the horizon, both running SteamOS. If the Verified system can't reliably tell players which games run properly on a single handheld, scaling that trust across multiple devices with different performance profiles is going to be a nightmare. Building the data pipeline now, before those products ship, is smart.

One concern raised by commenters on the Steamworks post is how the framerate data handles players running different graphics settings or using third-party tools like frame generation mods. If someone is playing at minimum settings and getting 25 fps while another player uses a performance mod to hit 50, the average becomes meaningless. Valve hasn't addressed this yet, and it's the sort of detail that could undermine the whole system if left unresolved.

What I'd really like to see next is Valve using this data to actually revoke Verified badges when performance drops below an acceptable floor. Right now, a game can earn Verified at launch, ship a patch that tanks framerate, and keep the badge indefinitely while developers may or may not get around to fixing it. The data tools are a good foundation, but they only work if there are consequences attached. Developers getting a chart that shows their game averaging 22 fps should trigger more than a polite notification.