Valve is now showing Steam Deck average FPS and framerate distribution graphs on Steam game pages, giving handheld buyers better performance context while leaving important caveats around settings, VSync, samples, and stutter.
Image: IGDB
Store links: Windosill: Steam Deck Edition on Steam
Valve turns Steam Deck performance into public store data
Valve has started showing Steam Deck framerate graphs and average FPS data directly on Steam game pages, a concrete upgrade to a compatibility system that has often told players whether a game runs without clearly showing how well it runs. According to Valve’s Steam announcement, the new data appears for games that have been tested for Deck compatibility, including Verified, Playable, and Unsupported titles, when enough opted-in player data exists.
The immediate tension is useful: Valve is making real-world Steam Deck game performance visible at the point of purchase, but the numbers are not controlled benchmarks. They come from customers who opted in to share anonymized framerate data, which means they reflect actual play across different habits and in-game settings rather than one standardized test route.
That distinction is the center of the feature. Steam Deck average FPS is now visible under the Steam Deck Compatibility section by selecting “Learn more,” as Valve described and as GamingOnLinux, Kotaku, VideoCardz, PC Guide, TheClick, and Rock Paper Shotgun all reported from the rollout. Where available, the store page shows an average framerate and an expandable histogram that breaks down how often reported samples landed in different FPS ranges. The same information is also available in the Steam Library for users on Steam Client Beta, according to Valve’s post as quoted by GamingOnLinux and VideoCardz.
For handheld buyers, this is the most meaningful Steam Deck store change since the Verified program became the main shorthand for Deck support. The badge still matters, but Valve is now attaching performance evidence to it. A green check no longer has to carry the whole argument.
The Verified badge was never a full performance promise
Valve’s existing Steam Deck Compatibility labels solve one problem and leave another exposed. Verified, Playable, and Unsupported are compatibility categories, not a precise performance scale. Valve says its test criteria “may not match every customer’s bar for acceptable gameplay,” a line echoed across the reports because it explains the design gap the new feature is trying to close.
That gap has been obvious to Deck owners for years. A Verified game can be technically comfortable to launch, read, control, and navigate while still landing below a player’s preferred framerate. An Unsupported game can sometimes run well enough for someone who accepts workarounds or a lower target. Rock Paper Shotgun points to the familiar mismatch where a game barely holding 30 FPS and another racing far higher can both sit under positive-looking compatibility language.
The new Valve Steam Deck performance data changes the buying calculation from a binary yes-or-no check to a probability read. If a strategy game, RPG, or roguelike shows a strong average and a tight distribution, the risk profile looks different than a Verified game with a wide spread and visible low-end samples. That matters most in genres where feel depends on stable input and timing. In a turn-based tactics game, 30 FPS may be acceptable if the frame pacing is consistent. In an action game, the same average can feel poor if the histogram suggests frequent drops.
Kotaku reported examples from live store data, including Final Fantasy VII Rebirth showing an average of 37 FPS on Steam Deck OLED and Hades II showing 88 FPS on the same hardware at the time of its article. Those figures are useful because they illustrate range, not because they settle the performance story forever. One title sits in the zone where expectations, settings, and tolerance matter. The other appears to land much closer to the OLED screen’s high-refresh ceiling. Both are far clearer signals than the same compatibility badge alone.
How to read the histogram without overvaluing the average
The average is the headline number, but the histogram is the sharper tool. TheClick explains that the graph shows the percentage of reported samples in different FPS ranges. That makes it valuable because an average can hide the behavior players actually feel. A game averaging 45 FPS could be spending much of its time near that target, or it could be bouncing between smoother stretches and low dips that make input and camera movement feel uneven.
Valve’s implementation is therefore closer to a crowd-sourced performance distribution than a traditional review benchmark. It can show whether a game regularly clusters near 30, 45, 60, 90, or lower ranges, depending on the model and the game. It cannot replace a reviewer’s fixed-route test with stated settings, nor does it publish the kind of 1 percent low or frame-time analysis PC performance readers often use to diagnose stutter. TheClick specifically cautions that the data is not a standardized 1 percent low test, so it cannot tell you precisely how long every hitch lasts.
For buyers, the practical read is simple but important. Start with the compatibility rating, then open the framerate panel, then treat the histogram as the truth serum for the average. A high average with a noticeable low-end tail should prompt caution, especially for fast games. A modest average with a tight cluster may be perfectly playable if you are comfortable with that target. In other words, the graph helps players define “good enough” for themselves, which is exactly the weakness Valve says it is trying to address.
This is where the feature becomes strategically interesting for Steam as a storefront. Performance used to be outsourced to YouTube benchmarks, ProtonDB-style community notes, forum posts, and specialist sites. Those still matter, but Valve is now moving a slice of that decision support into the store flow. That can reduce bad purchases, but it also raises the standard for how carefully players interpret a single number.
The numbers are real-world data, not lab results
The strongest caution is that Valve’s public framerate data is drawn from opted-in customers, not identical benchmark conditions. VideoCardz reports that Valve only includes results from Steam Deck systems using their default framerate limits, which are 90 FPS for the OLED model and 60 FPS for the LCD model. That is a helpful filter because it removes some manual cap variation, but it does not standardize in-game graphics settings.
TheClick makes the same point directly: one player using low settings may contribute alongside another choosing higher image quality. That mixed input is both the strength and weakness of the dataset. It reflects ordinary play rather than artificial consistency, but it also means a game’s Steam Deck average FPS is not a guarantee for your personal setup. A player who enables ray tracing where available, increases texture settings, installs mods, or values battery life enough to reduce power draw may see a different result.
There is also a sample threshold. Valve says a game needs a minimum amount of anonymous data before performance stats are displayed, according to PC Guide and VideoCardz. That means larger player-base games are more likely to show data early, while niche releases, brand-new launches, and older titles with limited Deck activity may remain blank even if they run well. Absence of a graph should not be read as proof of poor Steam Deck game performance. It may only mean Valve does not yet have enough public data to show.
Then there is the OLED and LCD split. Reports from Rock Paper Shotgun and PC Guide note that Valve shows data for both Steam Deck LCD and OLED models. That is necessary because the OLED model has a 90Hz display and small hardware differences, while the original LCD model uses a 60Hz display. Still, the presence of two graphs can tempt players to overread the gap. Rock Paper Shotgun’s example with Mortal Shell II is a useful warning: the site observed LCD results mostly around 30 FPS, matching its experience, while OLED results appeared around 45 FPS, a spread the author argued was larger than the OLED’s modest APU advantage would normally explain.
VSync can make clean-looking graphs misleading
Valve itself warns that VSync can skew the displayed results, according to Rock Paper Shotgun and TheClick. That is the most technical caveat, but it may be the one most likely to confuse buyers because the graph can look authoritative even when the frame delivery behind it is messier.
TheClick describes an OLED-specific pattern where a 90Hz screen and VSync can produce readings clustered around 90 FPS and 45 FPS. Valve’s example, as reported by TheClick, uses a repeating 90, 90, 45 sequence that averages to 75 FPS. A game with an internal 60 FPS cap can still bounce between those displayed rates. The lesson is that a 90/45 split does not necessarily mean half the play session feels like smooth 90 FPS and half feels like sluggish 45 FPS. It may be an artifact of how frames are presented and measured.
For anyone used to competitive settings menus, this is familiar territory. Frame caps, refresh rates, VSync behavior, and engine-level limiters can make the same nominal FPS feel very different. On a handheld, the situation is even more layered because players often tune performance around battery life, fan noise, and heat. Valve’s data currently tells you framerate distribution. It does not tell you battery drain, wattage, thermals, fan behavior, shader compilation behavior, or whether the player changed resolution scaling.
That does not make the feature weak. It makes it a scouting report rather than a match result. If the histogram says a game usually lives near your preferred target, it deserves a closer look. If it shows a spread that suggests instability, you should search for settings guidance or wait for more reports. Handheld buying has always been about tradeoffs. Valve has given players a better map, but the terrain still matters.
Desktop expansion could be bigger, and harder
Valve is already presenting this Steam Deck rollout as a first step. In its announcement, quoted by Rock Paper Shotgun and VideoCardz, Valve said it sees the update as “a stepping stone toward providing framerate data for more devices than just Steam Deck.” The company is prompting Steam desktop client users to opt in to anonymized framerate reporting through the Steam Client Beta, with the opt-in planned for all customers later.
That ambition is significant because Steam Deck is the easiest version of the problem. Even with LCD and OLED differences, Valve is dealing with a relatively fixed hardware family. Desktop PCs are a different economy entirely: CPUs, GPUs, memory, drivers, storage, display refresh rates, and user settings vary wildly. VideoCardz notes that presenting comparable desktop results would require additional hardware information. Without that context, an average desktop FPS number would be nearly useless.
If Valve solves that problem carefully, Steam could eventually become a much better place to judge performance before purchase across handhelds, SteamOS devices, and conventional PCs. If it solves it too broadly, the data risks becoming noisy and overtrusted. That balance will define whether Valve Steam Deck performance data becomes a durable buying tool or another store widget players learn to discount.
There is also a hardware strategy angle. GamingOnLinux argues that starting with a fixed system like Steam Deck makes sense for ironing out the system before moving elsewhere, and PC Guide reports that desktop support is coming through the opt-in path. Valve has also been expanding SteamOS ambitions beyond the original Deck ecosystem, according to the reporting cited by Rock Paper Shotgun. Public framerate telemetry gives that ecosystem a language for performance that storefront badges alone cannot provide.
How handheld buyers should use the new Steam Deck FPS data
The safest approach is to treat Steam Deck framerate graphs as a decision aid, not a verdict. If you are buying a game primarily for handheld play, check the Steam Deck Compatibility panel, expand “Learn more,” compare the LCD or OLED data that matches your device, and read the histogram before trusting the average. If the game has no graph, assume only that Valve lacks enough eligible data to show one publicly.
For demanding new releases, wait for additional testing if the histogram looks broad, if the average lands below your comfort zone, or if the game is likely to receive performance patches. For lighter games and genres that tolerate lower framerates, the new data may be enough to buy with confidence, especially when the distribution is tight. For action-heavy games, do not ignore low-end bars simply because the average looks acceptable.
The feature is good news for buyers because it moves Steam Deck verified framerate discussion out of guesswork and into visible evidence. It is also early enough that caution is warranted. Valve’s data is opt-in, settings are mixed, VSync can distort patterns, and the graph does not explain power use or frame pacing in full.
Still, this is a meaningful shift in Steam’s handheld meta. Valve has taken performance, one of the most important variables in portable PC purchasing, and placed it beside the buy button. The players who benefit most will be the ones who read the graph as a range of outcomes rather than a promise.