Methodology
On this page
On this page
Loading methodology page
SteamHardware.io separates official facts, reporting, community signals, reviews, and analysis so readers can see what is known, what is inferred, and what is still unknown.
The tracker uses anonymous reports submitted to SteamHardware.io. One report is one submission, not one verified person. Partial reports can still help when a chart needs only one date or milestone.
Charts use only the fields needed for that calculation. Cumulative funnels count a reported later milestone in earlier stages. Some totals also apply elapsed-time assumptions after a reported purchase, shipment, or invitation. Those assumptions do not change individual reports, Activity, timing results, or queue ranges.
Where follow-up is sparse, a separate marked extension may estimate missing updates. Some Steam Controller and Steam Machine totals use this extension. Machine estimates use Controller follow-up evidence and are labeled as estimates, not measured Machine rates. We hide the extension when the evidence is too weak.
Small or unusual samples can distort averages, so the tracker favors medians and typical ranges. If a state or region lacks enough reports, the page may fall back to a broader country result and labels the fallback. Very thin samples do not get a published range.
Tracker results show a trend, not a promise. They are not Valve data, sales totals, delivery guarantees, or counts of unique people.
Run Scores describe how a game runs on one device. The 1-to-5 scale uses the same anchors throughout the site: Broken, Poor, Playable, Good, and Excellent. It measures how the game worked, not whether someone liked it.
Steam Machine and Steam Deck results can include anonymous community reports, reviewed Community observations, and cited external tests. Frame uses a separate community-report model. Each numeric score included in an aggregate has equal weight, and the evidence counts stay visible. Missing settings stay missing. We do not turn an unanswered field into a hardware claim.
Community observations keep their exact source URL and identity private. Cited tests show a safe source link. Neither creates a contributor-owned community report or account.
Steam Machine, Steam Deck, and Steam Frame evidence stays separate. Frame standalone and streamed results also stay in their own matching groups. A score may appear with only one observation, so the visible count matters.
External evidence needs useful device-specific detail, such as an FPS target, graphics preset, or output resolution. Official compatibility badges remain a separate signal. If a source no longer meets the site's standard, it can be withdrawn from public results without deleting the audit history.
FPS, resolution, and graphics medians may come from different observations. They are not one recommended setup. A source-backed settings recipe instead keeps one tested setup with its measured result. We record only what that source states, give the recipe a high, medium, or limited confidence label, and leave missing details blank.
Run Scores are observations, not controlled benchmarks or performance guarantees. Game updates, software changes, display choices, and settings can all change the result.
You can read SteamHardware.io and contribute without an account.
The tracker and Run Scores do not ask for Steam credentials, Steam IDs, usernames, emails, order numbers, reservation identifiers, tracking numbers, street addresses, ZIP codes, screenshots, documents, account handles, payment data, or government identifiers.
Public reports show only the details needed to understand the result. Private edit links and tokens are not included in public pages or APIs. Run Scores may have an optional anonymous public profile, but hardware tracker reports are never linked to it.
There is no public download of individual tracker reports. Public summaries may be cited with attribution to SteamHardware.io. Do not use them to identify contributors, combine them with private account data, bypass public limits, or imply that the data came from Valve.
Read the full Privacy policy. If a public report exposes something private, contact SteamHardware.io.
If something looks wrong, contact SteamHardware.io with the page or public report link and the strongest source you have. Do not send a private update link, token, or personal information.
We update pages when the evidence changes or we find an error. A meaningful correction should say what changed and why. It should not hide uncertainty or imply Valve confirmed something it did not.
Editorial images may use original non-generative artwork, licensed or official assets, screenshots for source context, or existing source photos with the site's standard treatment. We do not use AI-generated artwork for article covers, heroes, thumbnails, or social images. Images must not invent test results, interfaces, product details, packaging, or availability.