Analysis Transparency

Move accuracy is derived from centipawn loss (CPL): the difference between the engine's best move and the move played. Lower CPL means better practical accuracy.

This creates gentle penalties for small slips and steep penalties for large tactical misses. It mirrors how mistakes feel in real games.

Chessomy classifies each move by how much evaluation is lost from that position. The goal is legibility: players should understand what happened without decoding jargon.

Some moves receive special tags when context matters more than raw CPL. These labels help explain notable tactical or practical moments.

Not every move carries equal importance. Critical positions are weighted more than routine recaptures. Chessomy also applies a mercy rule to prevent one collapse from making an otherwise solid game look artificially hopeless. The intent is honest reflection, not score inflation.

Fast and Balanced are practical defaults for everyday review. Deep, Ultra, and Mega search further, but require more time and battery. On mobile, lower presets are usually the best tradeoff. On desktop, higher depths are more comfortable for long study sessions.

• Engine evaluations can shift at higher depth, especially in sharp tactical positions.
• Any single-number summary hides nuance; use labels and lines together, not in isolation.
• Browser performance differs by device, so analysis speed and depth consistency can vary.
• Chessomy explains decisions, but it does not replace human judgment or coaching context.