What Exactly Are the 106 Independent Checks BotRefund Uses?

What the 106 checks cover

The 106 independent checks are a set of signals gathered from a visitor's browser, device, and behavior. They fall into a few broad categories:

• Browser fingerprinting – details like user agent, screen resolution, fonts, WebGL render data, and installed plugins.
• Hardware and GPU – information about the CPU, graphics card, and how they report concurrency and performance.
• Behavioral and biometric signals – mouse movements, click patterns, keyboard dynamics, scrolling, and timing.
• Network context – the IP address, connection type, and other network-derived clues.

Each check is a single data point. None of them is a bot verdict on its own. BotRefund uses them together to build a reliable picture of whether a visit is human or automated.

The checks are independent. That means they do not rely on the same underlying data. A bot that fakes one signal might still trip another. This independence is key to the accuracy of the system.

Category breakdown

This table gives a quick view of the 106 checks. But the real list is more detailed. Each category includes many individual signals.

Examples of checks in each category

Here are specific checks BotRefund uses. They come from its public bot detection pages and the homepage.

• Ghost click detection – catches click activity that happens without the natural sequence of human intent. (Click behavior)
• Honeypot trap interactions – watches for bots that respond to hidden or intentionally deceptive page elements. (Trap behavior)
• Robotic linear mouse movements – flags unnaturally straight pointer paths that rarely appear in real user sessions. (Pointer behavior)
• Absence of humanlike mouse tremor – looks for the tiny imperfections and jitter typical of human movement. (Motion behavior)
• Superhuman input speed (<1ms) – identifies interactions that happen faster than a person could realistically perform. (Speed behavior)
• Grid-aligned movement patterns – detects movement that snaps to precise lines or blocks instead of natural curves. (Path behavior)
• Absence of clicks or scrolling – highlights sessions that stay too static to match a real browsing journey. (Engagement behavior)
• Unnatural session durations – catches visit lengths that are too short, too long, or too uniform to be human. (Session behavior)

These are just a few. The full set includes many more like CPU Concurrency Lie, window.open Tamper, and Impossible Tab Speed. Each one is a separate independent check.

How a single check works

Take the CPU Concurrency Lie check as an example. A normal browser reports hardware, graphics, fonts, and operating-system details that naturally fit together for that device. The check looks for a mismatch that a real browsing session does not normally create. Virtual machines and spoofed profiles can claim one device while their graphics, fonts, audio, or processor behavior tells another story.

Similarly, the window.open Tamper check looks at how scripts interact with the browser. Real visitors produce imperfect, varied behavior: pauses, hesitation, natural movement, and interactions shaped by reading and decision-making. Scripts can send clicks and scrolls, but they struggle to reproduce the varied timing, movement, and hesitation of real people.

Impossible Tab Speed measures how quickly a visitor switches tabs. A bot can do this faster than any human. These checks are precise and measurable. They give BotRefund objective evidence about the visit.

Why a single anomaly is not a bot verdict

One anomaly alone is never enough. Privacy tools, travel, corporate networks, and unusual devices can produce unexpected behavior for genuine people. A real user on a VPN or a shared office network might trigger a few of these signals by accident.

BotRefund handles this by keeping each check as evidence—not a verdict. The checks are cross-referenced against other independent browser, network, device, and behavior data. Only when multiple signals tell the same story does the system lean toward a bot classification.

How the checks are combined

The real value comes from corroboration. BotRefund sends each signal into its prediction AI, which evaluates the complete picture across browser, network, device, and behavior evidence. By seeing how all signals fit together, it identifies a visit as bot or human with high accuracy.

In practice, this means a single strange reading might be dismissed if everything else looks normal. But if several independent checks point to the same conclusion—say, a spoofed GPU, superhuman input speed, and no mouse tremor—the model can be confident.

According to BotRefund, this approach achieves 99% accuracy. That accuracy comes from corroboration, not one browser tell.

Decision criteria: when to trust the checks

You might wonder when the checks are reliable enough to act on. BotRefund uses a few decision rules:

• Independence: Each check adds one objective fact. They are not duplicates of the same signal.
• Cross-checking: BotRefund tests whether other signals support the same story. If they do, the evidence is stronger.
• AI prediction: The model weighs the complete pattern instead of trusting a raw rule.

So a single anomaly is ignored. The system only acts when multiple independent signals agree. That keeps false positives low.

For an advertiser, this means you can trust the evidence when it points to a bot. The checks are designed to be specific enough to catch bots without flagging real users.

Why these checks matter for ad refunds

Bot clicks steal up to 20% of Google and Meta ad budgets. To recover that money, you need proof that the clicks were invalid. The 106 checks provide that evidence.

BotRefund uses the checks to detect every bot that clicks your ads and capture video proof for each one. That proof is then used to negotiate with Google and Meta for refunds. The more independent signals you have, the stronger your case.

The checks also help you understand why a visit is considered a bot. You can review the specific signals in your audit report.

Limitations and when these checks might not apply

No detection system is perfect. A determined bot can try to mimic human behavior, and some real users can look robotic—especially if they have motor impairments or use assistive technology.

BotRefund mitigates this by using many checks rather than relying on a single rule. That said, the 106 checks are designed for websites and ad click detection. They are not a universal anti-fraud solution for every scenario.

Also, these checks require JavaScript to run. If a visitor has JavaScript disabled, some checks cannot be performed. In that case, BotRefund uses whatever signals are still available and flags the session as potentially incomplete.

Frequently asked questions

Are all 106 checks applied to every visit?

Yes, BotRefund runs all applicable checks on each visit. Some checks may be skipped if the browser doesn't support a certain API, but the system tries to gather as many signals as possible.

How long does it take to run the checks?

The checks run in real time, typically within a second of the page load. They are lightweight and don't slow down the user experience.

Can a bot beat all 106 checks?

It's extremely difficult. The checks are independent, so a bot that mimics one signal might miss another. The cross-referencing approach makes it hard to trick every check at once.

Do these checks use cookies or storage?

Some checks use temporary data, but BotRefund is designed to respect privacy and relies mainly on signals that are already available in the browser.

What happens if a check flags a real user?

A single flag is ignored. The system only takes action when multiple independent checks agree. This keeps false positives low.

How do these checks support refund claims?

The checks produce timestamped evidence for each invalid click. That evidence is formatted into dispute reports and sent to Google or Meta during the refund negotiation.