AleksMouseTester v0.2.13

AleksMouseTester v0.2.13 — trust-hardening update

The scoring formula is unchanged (delivery-score-v3). This update is about trust: every number the app shows or exports is now provably the same number, with its validity attached — and the download is smaller and easier to verify.

What's new

• One score, everywhere — the result view, your Compare history, and the exported report now all read the score through one path. An invalid or legacy run shows "—" in every one of them; a stored verdict can't be overridden by a lucky recompute.
• Honest report export — exporting a run you loaded from history now uses the run's original stored data and score. Heads-up: re-exporting an old run can show a lower — true — score than an export made with an earlier version (the old export path silently dropped the worst-case tail metrics and scored without them).
• Two mice = no verdict — if a second device contributed to a capture (e.g. a touchpad twitch), the run is marked invalid (multiple devices) instead of silently mixing two devices into one measurement. Timing intervals are always built per device.
• Nothing is lost silently — if a run can't be saved (OneDrive / Controlled Folder Access / full disk), the app says so and offers a retry. Unreadable history files are counted in Compare instead of vanishing. History keeps the most recent 300 runs — that rule is now shown, and pruning is logged.
• Crash diagnostics — unhandled errors are written to a small rotating log (%LocalAppData%\MouseTester\logs), so a problem report can actually be diagnosed.
• Live headline — during a running test the score card now reads "Live / Measuring" instead of a stale "Ready".
• Smaller, cleaner download — the internal kernel-tracing machinery (~6 MB of native DLLs) is no longer shipped in the public build. BUILDINFO.txt states build_flavor=consumer.
• Better CSV export — every completed Quick Test writes a _metrics.csv (schema 3) next to its saved session: tail metrics (p95 / p99.9 / max gap / stutter & burst counts) plus the delivery score with its validity status and algorithm version.
• Power tweak safety — the USB power settings helper now takes an immutable restore point on first use, verifies every restore by reading the values back, and covers both AC and battery rails.

Your saved runs and settings carry over unchanged.

Download & integrity

• AleksMouseTester-v0.2.13-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed).
• SHA-256 of the .zip: e7842167900e30ef33491a77cc43f3a3c302ec8319f5487fc230959b77546400
• After unzipping, SHA256SUMS.txt lists the SHA-256 of every individual file — now in the standard format, so sha256sum -c SHA256SUMS.txt works out of the box (note: the file format changed vs earlier releases). THIRD-PARTY-NOTICES.txt lists the bundled open-source components.
• BUILDINFO.txt records the source commit, build flavor, and the pinned obfuscation tool version.

This release is a folder (the .exe plus its files), not a single .exe. Unzip and keep the whole folder together, then run AleksMouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway").

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

AleksMouseTester v0.2.12

AleksMouseTester v0.2.12 — clarity & honesty update

The measurement, scoring and your saved history are unchanged. This update makes the results clearer and more honest.

What's new

• Clearer Quick Test — one Start Quick Test button, a plain-language result, and the measurement scope shown right at the result: what it measures (how evenly your mouse's input arrives at the Windows raw-input boundary) and what it does not (USB polling rate, sensor or firmware latency).
• Honest invalid runs — a test with too few samples or a capture hiccup now shows "—" and "couldn't be evaluated reliably" instead of a misleading low score, in both the Quick Test result and your Compare history. (It could previously read as a scary "1.0".)
• Metric hover help — hover any Advanced / Lab metric for a plain-language explanation.
• Less UI interference during capture — reduced UI work while a test is running, so the app does less background drawing during measurement.

Your saved runs and settings carry over unchanged.

Download & integrity

• AleksMouseTester-v0.2.12-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed).
• SHA-256 of the .zip: 402bf83fde9f0d8540910170cf8f076eab44dc8d703fcddaaee3dcd62bec2f76
• After unzipping, SHA256SUMS.txt (in this folder) lists the SHA-256 of every individual file, and THIRD-PARTY-NOTICES.txt lists the bundled open-source components.

This release is a folder (the .exe plus its files), not a single .exe. Unzip and keep the whole folder together, then run AleksMouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway").

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

AleksMouseTester v0.2.11

AleksMouseTester v0.2.11

Renamed: MouseTester → AleksMouseTester

This project is now AleksMouseTester — an independent tool with its own focus: how evenly your mouse's input is actually delivered to the host (timing stability, honest validity flags, the highest stable rate as your sweet spot). The rename makes clear it is not the older, unrelated MouseTester by microe1.

• Purely a name change from v0.2.10 — the measurement, scoring and features are unchanged; only the visible name moved (window/dialogs, the executable, this repo).
• Your saved history carries over. The data folder is unchanged, so all your existing runs and settings are still there after updating.
• The download is now AleksMouseTester.exe inside AleksMouseTester-v0.2.11-win-x64.zip.

---

A free Windows tool that measures how evenly your mouse's input is delivered to the host — honest that this is host-delivery timing, not a device polling-rate or latency claim.

Highlights

• Quick Test with a plain-language delivery score
• Advanced / Lab metrics + a live interval graph (double-click it for a labeled full-screen view)
• Compare your runs, find your stable sweet-spot rate (the rate that's consistently clean, not the highest one that managed it once), and double-click any run to reopen it — values, ratings, and graphs
• Optional power-saving check / disable / restore (reversible, admin-gated)

Download

• AleksMouseTester-v0.2.11-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed)
• SHA256: b5f52377235100d302070f4c368d8af81abfbfadeb97d8095bf7d6021d40beed

Unzip (keep the whole folder) and run AleksMouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway"). See How this was built in the README for AI-assistance, closed-source status, and what the SHA-256 does and doesn't prove.

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

MouseTester v0.2.10

MouseTester v0.2.10

A fix-focused release — mostly things that keep the tool honest about itself.

What's fixed

• The "sweet spot" recommendation is now honest. Compare ranks each rate by its typical (median) delivery across your runs, so it points you to the rate that's consistently clean — not the highest rate that happened to nail it once. On a system where 8000 Hz is sometimes great but often stutters, it now correctly recommends the steadier rate (the whole point: the highest stable rate, not the highest number).
• Per-mouse USB power saving now reads the real state. The "Check power saving" panel used to show a stale "On" for your mouse even after you disabled it (a custom power plan could trigger this). It now reads the exact same Windows state it writes, verifies the change actually took effect (no more false "done"), refreshes the panel immediately, and shows "Mixed" when a composite mouse's interfaces disagree.
• Reopened runs no longer hide their worst spike. Double-clicking a saved run on the Compare page now always shows its worst gap and worst batch in the graphs — the earlier downsampling could drop the single biggest outlier, and a saved graph should never look cleaner than the real run.
• Power tweaks are truly reversible. The optional power tweak now changes nothing if it can't save a restore point first, so "Restore settings" always has something to restore.
• The tool stays out of its own measurement. Power-saving actions are blocked while a test or capture is running, so checking/changing power settings can never disturb a live reading.

---

A free Windows tool that measures how evenly your mouse's input is delivered to the host — honest that this is host-delivery timing, not a device polling-rate or latency claim.

Highlights

• Quick Test with a plain-language delivery score
• Advanced / Lab metrics + a live interval graph (double-click it for a labeled full-screen view)
• Compare your runs, find your stable sweet-spot rate (higher Hz isn't automatically better), and double-click any run to reopen it — values, ratings, and graphs
• Optional power-saving check / disable / restore (reversible, admin-gated)

Download

• MouseTester-v0.2.10-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed)
• SHA256: d499e35e50596f2ac0e0e22bb3449e199b0683097f39b7aeb09b61a54d52e3c1

Unzip (keep the whole folder) and run MouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway"). Signing is planned. See How this was built in the README for AI-assistance, closed-source status, and what the SHA-256 does and doesn't prove.

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

MouseTester v0.2.9

MouseTester v0.2.9

What's new in this version

• Open any past run. On the Compare page, double-click a run to load it back into the full UI — its values, ratings, score breakdown, metric band, and graphs. New runs store a compact trace so the time-series graphs reload too; older runs (recorded before this version) show the saved distribution graph.
• Fix: correct reference line when switching runs. The red target line on the Timing-Distribution graph used to keep the previous run's rate when you clicked between saved runs. It now always follows the loaded run's own target Hz.
• Lighter on your system while measuring. Reduced the app's own overhead during a capture — higher-priority capture thread, garbage-collection pauses suppressed for the measurement window, fewer lock reads, and no redundant post-capture passes. An honest measurement tool should disturb the very timing it measures as little as possible; this matters most at 8000 Hz.

---

A free Windows tool that measures how evenly your mouse's input is delivered to the host — honest that this is host-delivery timing, not a device polling-rate or latency claim.

Highlights

• Quick Test with a plain-language delivery score
• Advanced / Lab metrics + a live interval graph (double-click it for a labeled full-screen view)
• Compare your runs, find your stable sweet-spot rate (higher Hz isn't automatically better), and double-click any run to reopen it — values, ratings, and graphs
• Optional power-saving check / disable / restore (reversible, admin-gated)

Download

• MouseTester-v0.2.9-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed)
• SHA256: a0b48d6b3f2a83efac4ce8d1cee9ca454e41ba0d8b7d80d4157a9844d4ff0210

Unzip (keep the whole folder) and run MouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway"). Signing is planned.

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

MouseTester v0.2.8

MouseTester v0.2.8

What's new in this version

• Better across monitor resolutions. The app now opens maximized, so the graph has real room on smaller / Full-HD (1080p) screens — the earlier default window left it cramped.
• Full-screen graph. Double-click any graph to open it full screen. It keeps the scroll-wheel zoom and adds dropdowns to switch the graph / X-axis. Press Esc (or double-click again) to close it.
• Labeled axes in the full-screen graph — titled X and Y axes with the exact value ticks (there's finally room for them at full size).

---

A free Windows tool that measures how evenly your mouse's input is delivered to the host — honest that this is host-delivery timing, not a device polling-rate or latency claim.

Highlights

• Quick Test with a plain-language delivery score
• Advanced / Lab metrics + a live interval graph (double-click it for a labeled full-screen view)
• Compare your runs and find your stable sweet-spot rate (higher Hz isn't automatically better)
• Optional power-saving check / disable / restore (reversible, admin-gated)

Download

• MouseTester-v0.2.8-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed)
• SHA256: bd5ecb07c5a2db80a8f55708d45473db18d8fdd1468ab59713f6f78d7242aa27

Unzip (keep the whole folder) and run MouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway"). Signing is planned.

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.

MouseTester v0.2.7

MouseTester v0.2.7

A free Windows tool that measures how evenly your mouse's input is delivered to the host — honest that this is host-delivery timing, not a device polling-rate or latency claim.

Highlights

• Quick Test with a plain-language delivery score
• Advanced / Lab metrics + a live interval graph
• Compare your runs and find your stable sweet-spot rate (higher Hz isn't automatically better)
• Optional power-saving check / disable / restore (reversible, admin-gated)

Download

• MouseTester-v0.2.7-win-x64.zip — Windows 10/11 x64, self-contained (no .NET install needed)
• SHA256: c6de30fc7be7c72a3526a1d1a7fdb444be123e101faa784e80891bee1a4068df

Unzip (keep the whole folder) and run MouseTester.exe.

Not code-signed yet — Windows SmartScreen may warn on first run ("More info" → "Run anyway"). Signing is planned.

Honest status

Internally credible, not yet validated across many mice / PCs. Feedback from different hardware is very welcome.

Reports Windows host raw-input delivery timing, not physical USB/HID polling timing.