v2.0 · STEAM
Documentation
Learn how to run the Aim Control Test, read the Aim Score, set DPI and sensitivity, and compare a retest. A practical guide to FPS Performance Lab.
Training lets you repeat a specific action. Diagnostics controls the measurement conditions and helps explain which part of aim control needs work. FPS Performance Lab connects these roles, but a training-round score is not a replacement for a complete Aim Control Test.
The practical workflow is measurement, a comparable series, analysis, training and retest. Warm-up prepares you for a session; it is not a diagnostic assessment. A settings recommendation is a hypothesis to verify manually, not a promise of ideal sensitivity.
The application runs in its own window. It does not automate aiming, read game memory or operate as an overlay. It does not increase frame rates. Results are neither a medical diagnosis nor a guarantee of performance in a particular game.
The current version is 2.0 (Nov 15, 2026): a 3D training arena built with Unreal Engine 5, a match summary with clips, a new Dashboard and a training plan with a retest – described on the updates page. The module descriptions below come from version 1.11 and are being updated for 2.0. The app has been on Steam since October 15, 2026. The interface, diagnosis, in-app help and PDF report are available in 17 languages; by default the app uses your Steam language.
Names since 1.9.0: Aim Score (formerly ControlIndex), the pillars Flick, Stopping, Tracking and Consistency, and the modules Technique Analysis (formerly Aim Technique Coach) and Vision Training.
UpdatesCheck the Windows requirements, connect your mouse and set its actual DPI in the device software. Save your setup in the application: mouse, DPI, sensitivity, target profile and session context. Player-declared values are not an automatic sensor reading.
Choose a readable theme, target contrast and comfortable motion mode. A short warm-up familiarises you with the task. Avoid changing several parameters during one comparable series. Take a break when fatigued and consider that context when interpreting results.
Your first complete test establishes a baseline. After a change to the input pipeline, protocol or important conditions, do not automatically combine the old and new history into one trend.
The dashboard combines your latest control score, measurement status, four pillars, priority, plan, trend and recent activity. The headline number summarises a profile, so read it alongside data quality and the distribution of pillar scores.
View summary and View guidance expand the player-facing conclusion. Open plan leads to the recommended action. The full recommendation explains settings and the limits of the conclusion. Hardware guidance also uses declared inputs; it is not a laboratory measurement of the mouse alone.
The chart is meaningful when sessions are comparable. Insufficient history or a warning about changed conditions is a reason to collect more data, not to infer a trend.
The protocol covers preparation and warm-up, reaction, rapid target acquisition, stopping on target, smooth tracking and response to direction changes. In version 2.0 the test runs in the 3D arena. Analysis requires completion of the required stages and enough valid trials.
Acquisition describes reaching the target, endpoint control describes stopping the movement, continuous control describes following a moving target, and consistency describes repeatability across trials. These are different skills: high click accuracy does not automatically mean high scores in every pillar.
Reaction times describe a task performed by a player under specific conditions. They are not an isolated measurement of mouse, monitor or computer latency.
Windows Raw Input records mouse movement and buttons while the application window is active. In the background only while you run a match recording, a 3D arena round or the in-game turn measurement yourself – and only until it ends. The Raw Input field in a profile stores a declaration about the game configuration; it does not switch the measurement input path. Control Optimizer is an available experimental module whose findings require manual verification.
The Aim Score (0–100) combines four pillars: Flick (25%), Stopping (30%), Tracking (30%) and Consistency (15%). Reaction time is reported separately and does not change the Aim Score.
The site's archived example is 57 × 0.25 + 77 × 0.30 + 73 × 0.30 + 84 × 0.15 = 71.85, rounded to 72. This explains a specific report dated 14 July 2026; it is not the user's latest result or a population benchmark.
The same overall number can come from different pillar profiles. Exercise selection also depends on repeated patterns, conditions and interpretation confidence. A one-point change in a single session is not automatic evidence of progress.
Data quality describes how usable the recorded trials are. Confidence describes the strength of the evidence supporting an interpretation. High confidence does not mean high skill, and a low skill score does not inherently mean the measurement is faulty.
Warnings, missing stages and too few valid trials limit the conclusions that can be drawn. Complete the measurement rather than treating missing data as zero. Algorithm version and visual conditions form part of the report context.
Diagnostics, practice and warm-up serve different purposes. The following scope describes the full Steam build presented on this site.
Measure four control pillars, check data quality and find a practical next step.
Switching, stopping, tracking and consistency contribute to the Aim Score. Reaction is reported separately. A full result requires completing the diagnostic protocol.
Short rounds for reaction, flick shots, precision and tracking.
Choose the mode, difficulty, round duration and target parameters. A training score describes that session and does not replace the full diagnostic test.
Session comparisons, trends and before-and-after practice views.
Comparability depends on settings, hardware, protocol and measurement quality. The statistics laboratory provides additional indicators when the amount and type of data support them.
A target-game catalogue, saved configurations and sensitivity conversions.
DPI, sensitivity, ADS, FOV and the game model define the context. eDPI and cm/360 have model-dependent limitations. Apply all in-game changes manually.
An experimental comparison of 95%, 100% and 105% control scaling.
Six blocks, two per variant, use fixed conditions. Variants and partial results remain hidden until the comparison ends. A difference suggests a hypothesis that must be checked with a full retest.
Find a recurring movement pattern and a suitable exercise.
Analyses include overshoot, undershoot, late braking, corrections and target loss. Movement traces and guidance connect to practice, a plan and a comparable retest. A full profile requires sufficient data.
Six exercise families for attention, tracking and stimulus search.
Smooth tracking, multiple-object tracking, attention shifting, peripheral awareness, target search and distraction resistance use a separate profile. This module does not use an eye tracker or provide a medical diagnosis.
A 5-, 8- or 12-minute routine to prepare for a session.
Activation, flick, precision, tracking, direction changes and a readiness check form six stages. Warm-up history and readiness remain separate from the Aim Score.
17 languages (Steam language by default), readability, sound, reports and local files.
The interface theme and test-field appearance are configured separately. Settings include text scale, contrast and reduced effects, while data tools provide exports and backups.
Match Analysis has been in the app since version 1.11 as a beta. The in-app test measures your aim under repeatable conditions. Match Analysis shows how much of that survives in a real game, with pressure, movement and an opponent who shoots back.
How to record a match. In Match Analysis choose the game and start recording before the match. Play normally. The app records your mouse movement through Windows Raw Input and, if you tick the option, the screen. The screen is recorded by Windows, with the same mechanism as Xbox Game Bar. Stop recording after the match; the analysis takes a moment and runs on your computer.
What you get. The numbers and a few short clips with the moments worth watching in slow motion. Every clip has a hand-speed chart synchronised with the picture and a one-sentence description. In Counter-Strike 2 you also get kills, deaths and headshots from Valve's official Game State Integration.
| Measure | What it means |
|---|---|
| End of the first movement | Where your first movement stopped relative to where you shot: in mm of hand movement and in degrees of in-game turn. |
| Overshoot | The movement ended past the point where you shot and you had to come back. |
| Pull-in and corrections | How long the fix after the first movement took (ms) and how many corrections there were. |
| Shot on the move | The shot went off before your hand had stopped. |
| Mouse lift | A break in the middle of a turn because you ran out of mousepad. |
How to read the result. One match is a small sample: a few rounds on another map can move the numbers more than training does. Compare several matches in the same mode and draw conclusions from a repeated pattern, not from one round. Comparison with other players is optional and only works after you opt in. Version 2.0 added a match summary, deaths without a shot and a match-versus-test comparison.
Match Analysis has no overlay, does not read game memory and sends nothing to the game. The recording stays on your computer, and only short clips are kept after the match. Before playing in a tournament or league, check its rules on programs running in the background.
Example match, clips and safety questions: CS2 Match Analysis.
Choose a mode that matches your priority, then set the level, round time and target parameters. Reaction focuses on responding to a stimulus, flick on rapid acquisition, precision on controlled hits, and tracking on following a moving target.
Compare rounds with the same mode, level, duration and conditions. Session points follow the rules of the chosen exercise and are not the Aim Score scale. Short sets with controlled execution are more useful for interpretation than comparing records from unrelated tasks.
Start by recording the settings you actually use. The target-game catalog and saved profiles help reproduce that context. A conversion is only as reliable as its control model and inputs.
Quick cm/360, eDPI and sensitivity conversion when you change DPI or game: cm/360 & eDPI calculator.
Change one parameter at a time and retain the previous value. Apply a recommendation manually, then run a comparable retest. Lower or higher sensitivity alone is not evidence of a better fit.
Arithmetic example: base sensitivity 0.42 × multiplier 0.82 = 0.3444; applying a further ADS multiplier of 0.82 gives 0.282408. This illustrates multipliers independently of the settings shown in the screenshot.
History retains sessions and their context. The player-progress view helps you follow scores and compare before and after training. The statistical laboratory supports a broader reading when a suitable measurement series is available.
Keep hardware, DPI, sensitivity, profile, protocol version and task conditions consistent. The application flags mismatches; not every saved session belongs to the same series. Equivalent A/B forms support retesting but do not guarantee improvement.
Before retesting, record what you practised or changed. Review pillars, quality and repeatability, not just the overall score. A single better session may also reflect warm-up, fatigue or natural variation.
Version 2.0 supports 17 interface languages and uses your Steam language by default. The interface theme and training-field appearance are separate settings. Choose text size, contrast, background and target colour for readability, not just appearance.
Motion and effect settings help reduce discomfort. Sound and messages support the task flow. Keep important conditions consistent across sessions you intend to compare; changing target appearance can affect detection difficulty.
A report combines the score, pillars, quality, confidence, context and recommended next step. PDF and TXT export preserve the summary outside the application view. Comparisons depend on having suitable earlier data.
Read mouse guidance and suggested DPI, sensitivity and ADS alongside their reasoning and limitations. Declared weight, grip or device model are not laboratory measurements. A single session cannot prove that buying new hardware will improve your results.
Review a report before sharing it. It may contain a player name and hardware setup. The material shown on the website is an example, not a report about the visitor.
The application saves settings and results locally. Settings provides data management, export, backup and restore tools. Before resetting history or moving to another computer, make a backup and check that it contains the required files.
Local storage is not evidence of Steam Cloud support. This site does not claim cloud synchronisation. Details about the website, contact and existing Google integration are described in the privacy policy.
Start with a complete measurement. Select one priority and follow its assigned plan. Keep settings and rest intervals consistent, record meaningful changes, then repeat the test under comparable conditions.
Check whether the target pillar and execution pattern improved, not just the points total. Collect more data when warnings or inconsistent sessions limit interpretation. A bug report should include the app version, reproduction steps and a report without unnecessary personal information.
On Steam, for Windows 10 and 11. It has been available since October 15, 2026; the current version is 2.0.
No. FPS in the name refers to first-person shooters. The application provides aim diagnostics and practice, not frame-rate optimization.
Partly. This site has a free mini aim test: 20 targets in your browser with an instant result. The full measurement (crosshair control test, 3D arena, CS2 match analysis and retest) needs the Windows app.
No. Results include player behaviour and the conditions of the whole session. They are not isolated measurements of a sensor, monitor or hardware latency.
On Steam since Oct 15, 2026 · version 2.0
Buy on Steam