Tool

Avalanche Tester

For a well-designed cryptographic hash, flipping any single input bit should flip about half the output bits, with no detectable bias. Use the single-flip mode to see one cascade end-to-end; use the statistical mode to watch hundreds of random flips converge to that ideal.

Input

11 bytes · 88 bits

Algorithm

Bit position to flip: 0 / 87

Byte 0, bit 7 (within character h)

Background

The avalanche criterion (Feistel, 1973) requires that a single input bit flip propagate widely across the output.
The stronger strict avalanche criterion (SAC) (Webster & Tavares, 1985) requires every output bit to flip with probability exactly 0.5 when any one input bit is flipped.
Heatmap bars near teal indicate ideal behavior. Bars drifting toward amber or red mean some output bits are biased , a sign of a weak hash or a reduced-round variant.
None of MD5, SHA-1, SHA-256, SHA-512, SHA-3, BLAKE2, or BLAKE3 should show measurable bias at this scale. Try MD5 vs SHA-256: both look avalanche-correct; MD5's collision weakness comes from elsewhere (differential paths in the compression function), not from a missing avalanche property.

FAQ