Repetition Code¶
A block code that maps each symbol to a constant run of copies, trading rate for distance-based error recovery.
Core Idea¶
A repetition code maps each symbol to \(n\) identical positions: \(a\mapsto a^n\). For binary symbols, the only valid blocks are \(0^n\) and \(1^n\). Their minimum Hamming distance is \(n\), so a nearest-word decoder corrects up to \(\lfloor(n-1)/2\rfloor\) arbitrary substitutions. A separate validity test can detect up to \(n-1\); majority decoding alone need not report that an error occurred. The rate is \(1/n\).[ref-87a7f3144fbe][ref-88457c511008]
Scope of Application¶
The code can protect bits in a noisy communication link or in storage over time. Odd-length binary blocks can be decoded by majority. The distance guarantee does not assume independent errors, although a binary-symmetric-channel probability calculation does.[ref-87a7f3144fbe][ref-e2252b592c04]
Clarity¶
Encoding \(1\) as \(111\) and receiving \(101\) illustrates two separate tasks: majority recovers \(1\), while codeword membership flags \(101\) as corrupted. The \(n-1\) detection figure is not an \(n-1\) correction guarantee. Quantum bit-flip repetition is an analogy with different encoding and syndrome operations.[ref-87a7f3144fbe][ref-88457c511008]
Manages Complexity¶
The code reduces reliability analysis to the spacing of its valid words. The cost is direct: \(n\) channel or storage positions carry only one source symbol. A more elaborate code can use its positions more efficiently, depending on the error model.[^ref-87a7f3144fbe]
Abstract Reasoning¶
Specify the alphabet, block length, valid constant words and error class. Compute distance before claiming a tolerance bound. Decide whether the receiver needs a recovered symbol, an error flag or both, and choose the corresponding operations. Then separately evaluate probable performance under the actual noise process.[ref-87a7f3144fbe][ref-88457c511008]
Knowledge Transfer¶
The same constant-word map works for transmission across space and storage across time. It is a specialized Error-Correcting Code, not a claim that channel errors are always independent. Physical correlations and failure rates do not transfer automatically. Over larger alphabets, nearest-word decoding may require an explicit tie rule rather than simple binary majority.[^ref-e2252b592c04]
[^ref-87a7f3144fbe]: Polyanskiy and Wu, Information Theory author manuscript, §17.2. [^ref-e2252b592c04]: University of Stuttgart, Error Control Coding course. [^ref-88457c511008]: University of Michigan lecture on error detecting and correcting codes.
Relationships to Other Abstractions¶
Current abstraction Repetition Code Domain-specific
Parents (1) — more general patterns this builds on
-
Repetition Code is a kind of Error-Correcting Code Domain-specific
Repetition codes are error-correcting codes.
Hierarchy path (1) — routes to 1 parentless root
- Repetition Code → Error-Correcting Code → Encoding And Decoding → Transformation → Function (Mapping)
Neighborhood in Abstraction Space¶
Repetition Code sits in a moderately populated region (46th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.
Family — Codes, Matrices & Combinatorial Problems (30 abstractions)
Nearest neighbors
- Zyablov Bound — 0.88
- Even code — 0.88
- Majority Logic Decoding — 0.88
- Error-Correcting Code — 0.87
- Low-Density Parity-Check Code — 0.85
Computed from structural-signature embeddings · 2026-10-08