How We Test Mice & Keyboards
Every recommendation on MouseTestHQ is backed by real measurements, not marketing specs. This page explains exactly how we test mice, keyboards, and peripherals — the equipment, the processes, and the criteria we score against.
We publish this so you can judge our conclusions yourself. If you disagree with a pick, you have the tools to verify.
Our testing philosophy
We follow three principles:
- Measure what matters. Sensor accuracy, polling rate stability, click latency, and switch feel — not headline specs.
- Test under real conditions. Hours of use, not benchmarks. A mouse that's "perfect on paper" but uncomfortable after 2 hours is a bad mouse.
- Be transparent. Every review includes what we tested, how long, and any limitations of our method.
Mouse testing equipment
| Tool | Purpose |
|---|---|
| High-speed camera (1000 fps) | Measure click latency from switch actuation to OS event |
| Digital scale (0.1g precision) | Weigh mice with and without cable |
| Calipers | Measure length, width, height, hump position |
| Force gauge (0-100g range) | Measure switch actuation force and keycap pull force |
| Decibel meter | Measure keyboard sound at 30cm distance |
| Custom mouse test rig | Controlled motion for sensor accuracy testing |
Prices were last checked in October 2026 and change often. Always confirm the current price with the retailer before buying.
How we measure mouse performance
1. Polling rate stability
We measure polling rate over 10 minutes of continuous mouse movement, sampling every 100ms. Stable mice hold their rated polling rate (usually 1000Hz) with less than 5% deviation. Unstable mice dip to 500Hz or 250Hz frequently.
Our tool: Our free Polling Rate Checker in your browser gives you the same data we collect.
2. Click latency
We measure the time from physical switch actuation to the click event reaching the operating system. Lower is better. Healthy modern mice: 5-8ms. Worn or cheap switches: 12ms+.
Method: High-speed camera (1000 fps) records switch press; software correlates the timestamp with the USB HID event. We measure 100 clicks per mouse and take the median.
3. Sensor accuracy
Modern sensors (HERO 2, Focus Pro 30K, PMW3950) are all excellent, but we verify on a controlled motion rig. The mouse moves at consistent speeds across a 30cm track; we measure reported movement vs actual.
Pass criteria: less than 1% deviation across all tested speeds (10-200 IPS).
4. Battery life
For wireless mice, we measure continuous-use battery life at rated polling rate. Real-world battery life is usually 80-90% of manufacturer claims due to LED lighting, higher polling, and other factors.
5. Wireless consistency
We test wireless mice in environments with 2.4GHz WiFi, Bluetooth devices, and microwaves (common home/office interference). The mouse must hold its rated polling rate without dropouts.
Keyboard testing equipment
1. Switch analysis
We disassemble sample switches and inspect:
- Spring weight (measured on calibrated force gauge)
- Actuation point (measured with custom rig)
- Stem wobble
- Factory lubing quality
2. Typing feel
Subjective but important. Our testing panel of 12 writers/coders/types 5,000+ words per keyboard across 2 weeks. We score on:
- Tactility (is the bump clear or mushy?)
- Sound (clean thock vs rattly vs pingy)
- Bottom-out feel (smooth vs harsh)
- Stabilizer smoothness (spacebar, enter, shift — common rattle points)
3. Anti-ghosting & rollover
We use our Keyboard Tester (same tool you have) to verify NKRO (N-key rollover). Modern keyboards should handle 10+ simultaneous keys; cheap boards often fail at 3-4 keys pressed together.
4. Sound measurement
We record keyboard sound at 30cm distance using a calibrated decibel meter, both at typing speed (60 WPM) and during fast gaming inputs. This gives a consistent, comparable sound profile.
Subjective scoring criteria
Some things can't be measured with equipment. We score subjectively on:
- Comfort — long-session fatigue (1-10 hours)
- Build quality — flex, creak, weight distribution
- Aesthetics — design language, RGB (if any), branding restraint
- Software — usability, required vs optional, cloud dependence
Our scoring rubric
Each product is scored on:
| Category | Weight | Type |
|---|---|---|
| Sensor / switch quality | 25% | Measured |
| Performance consistency | 20% | Measured |
| Comfort & ergonomics | 20% | Subjective |
| Build quality | 15% | Subjective |
| Value (price vs quality) | 10% | Calculated |
| Software & ecosystem | 10% | Subjective |
What we don't test
Honesty about limitations:
- We don't have an anechoic chamber for precise audio measurements — sound scores are relative, not absolute
- We can't test every hand size perfectly — we use 12 testers but the range is limited to 15.5-22cm hand length
- We don't test long-term durability (years) — reviews reflect 1-3 month testing windows. Some products need years to reveal issues (like double-clicking mechanical switches).
Editorial independence
We are not paid by manufacturers for coverage. Affiliate links generate revenue when you buy through them, but the commission rate never affects our rankings. A 1% commission product that's the best for the use case will rank above a 6% commission product that's mediocre.
Read our full affiliate disclosure and about page for more on our editorial standards.
DIY testing at home
You don't need our equipment to do useful testing. The free tools on this site give you the same objective measurements:
- CPS Test — measure your click speed and compare over time
- Polling Rate Checker — verify your mouse is delivering its rated rate
- Keyboard Tester — check anti-ghosting and rollover
- Latency Test — measure your reaction time baseline
If you have questions about our methodology or want to suggest a test we should add, get in touch.
Frequently asked questions
What equipment do you use to test mice?
For sensor accuracy we use a reference tracking test at fixed speeds, cross-checked against a second method so we are not relying on a single reading. For polling rate we sample the raw input event stream over multiple intervals so throttling is visible rather than hidden by averaging. For click latency we run at least 30 rounds per mouse and report the median and the spread, not just the best result, because a single fast click means nothing.
How long does each product take to test?
A full test cycle takes between 8 and 15 hours per mouse, spread across several sessions. That includes sensor accuracy at multiple speeds, polling rate stability, click latency rounds, a physical build and durability inspection, and at least a week of daily use before we write anything. Speed matters for the reader, but a rushed test produces numbers we cannot stand behind, so we do not publish results we have not properly measured.
How do you score and rank products?
Every product is scored against the same rubric across five categories: sensor accuracy and consistency, polling rate honesty, click latency, build quality and durability, and value for money. Scores are normalised so that a product cannot win by excelling at one thing and failing at four others. The full breakdown of what each category measures is on this page.
Do you retest products when they are updated?
Yes. Mice and keyboards get revised firmware, new sensor revisions, and sometimes silent design changes. When a product we recommend changes materially, we retest it and update the guide with a new date. Our best gaming mouse guide and the other buying guides all carry a last-updated date so you can see how current the data is.
Why do you publish the methodology publicly?
Because a testing claim you cannot check is not much better than a marketing claim. Publishing our protocol means you can judge whether our test is meaningful for your use case, and it means other people can repeat it. If our methodology has a flaw, the public write-up is where that gets found and fixed, rather than hidden.