About

Precise timekeeping is assumed everywhere, from network protocols to human scheduling. This tool deliberately breaks that assumption. It applies controlled distortion algorithms to the current system time and renders the fuzzed result on both analog and digital displays. Six distinct fuzzing modes are available: Gaussian noise injects random offsets sampled from a normal distribution with configurable σ (standard deviation in seconds); drift accumulates a random walk that slowly deviates from true time; elastic time modulates clock speed sinusoidally so minutes stretch and compress; stutter freezes the display then catches up in bursts; quantize snaps time to the nearest n minutes; and reverse periodically runs the clock backward. The tool is useful for testing time-dependent UI, stress-testing scheduling logic, or simply visualizing how unreliable time perception can be. Note: the fuzzer operates on display only. It does not modify your system clock.

Formulas

The Gaussian noise mode samples from a standard normal distribution using the Box-Muller transform, then scales by intensity:

offset = σ ⋅ cos(2πu₁) ⋅ √−2 ⋅ ln(u₂)

where u₁, u₂ are uniform random values in (0, 1), and σ is the user-selected intensity in seconds. The fuzzed display time becomes:

t_display = t_real + offset

For drift mode, a Wiener process accumulates error over discrete ticks:

D_n = D_n−1 + σ_drift ⋅ z

where z ∼ N(0, 1) and σ_drift controls step size. For elastic time, the effective speed factor at time t is:

s(t) = 1 + A ⋅ sin(2πtT)

where A ∈ [0, 0.9] is amplitude and T is the oscillation period. Analog clock hand angles are computed as θ = (value ÷ max) × 2π − π÷2, where value is the fuzzed time component and max is 12 for hours, 60 for minutes and seconds.

Reference Data

Fuzz Mode	Algorithm	Intensity Range	Typical Use Case	Visual Effect
Gaussian Noise	Box-Muller normal distribution	σ = 1 - 300 s	Simulating NTP jitter	Jittery seconds, occasional minute jumps
Drift	Random walk (Wiener process)	0.1 - 5.0 s/tick	Testing clock synchronization	Gradual, cumulative deviation
Elastic	Sinusoidal speed modulation	Period 10 - 120 s	Perception experiments	Seconds stretch and compress rhythmically
Stutter	Freeze/catch-up cycles	Freeze 1 - 10 s	Network lag simulation	Clock freezes then jumps forward
Quantize	Floor to nearest n minutes	n = 1 - 30 min	Low-resolution displays	Time snaps in discrete steps
Reverse	Periodic direction inversion	Every 5 - 60 s	Art installations, disorientation	Clock runs backwards briefly
Combined	Layered modes	Per-mode settings	Chaos testing	Unpredictable compound distortion
Box-Muller	z = cos(2πu₁) ⋅ √−2ln(u₂)	Generates normally distributed random numbers from uniform samples
Wiener Step	W_t+1 = W_t + σ ⋅ z	Each tick adds a normally distributed increment to the accumulated drift
Elastic Speed	s(t) = 1 + A ⋅ sin(2πt ÷ T)	Speed oscillates between 1−A and 1+A
Clock Angle	θ = (value ÷ max) × 2π − π÷2	Converts time units to radians for analog hand positioning

Frequently Asked Questions

No. The tool operates entirely in the browser's rendering layer. It reads your system time via Date.now(), applies a mathematical offset, and displays the distorted result. Your operating system clock, NTP synchronization, and all other applications remain unaffected.

Simple random offset uses a uniform distribution, meaning a 5-second offset is as likely as a 300-second offset within the range. Gaussian noise uses the Box-Muller transform to generate normally distributed values where small offsets near zero are far more probable than extreme ones. This produces more naturalistic jitter, similar to real-world clock drift observed in NTP measurements where σ typically ranges from 1 to 50 ms.

The drift value D_n is clamped to ±43200 s (12 hours) to keep the analog clock meaningful. If you need extreme drift for testing, reduce the intensity or reset periodically using the Reset button. The accumulated drift persists across page reloads via localStorage.

The current implementation applies one mode at a time for clarity of observation. However, the drift mode inherently combines with the base time since it accumulates a persistent offset. To layer effects, switch between modes without resetting - the drift offset carries over and compounds with the new mode's distortion.

Elastic mode modulates clock speed sinusoidally. When sin(2πt ÷ T) is negative, the effective speed drops below 1.0, making displayed time advance slower than real time. At the trough, speed reaches 1 − A. With amplitude A = 0.9, the clock crawls at 10% normal speed during the slowest phase.

The Canvas 2D rendering uses requestAnimationFrame (typically 60 fps) and computes hand angles from the fuzzed millisecond timestamp. The second hand moves smoothly rather than ticking discretely. Under all fuzz modes, the analog and digital displays are driven by the same fuzzed timestamp, so they always agree. Rendering precision is limited only by canvas pixel resolution.