A ternary code that beats the table: A3(15,10) ≥ 31
Can new lower bounds for ternary codes be found by beating entries of Andries Brouwer's table of A3(n,d), the largest ternary code of length n and minimum Hamming distance d, for n ≤ 16?
Argus's locked verifier, an independent script and an in-browser check anyone can run. The audit agent agreed with the verdict, with notes: the supported hypothesis was registered just after the run that found the code, as the paper says.
A3(15,10) ≥ 31 (was 27; upper bound 39)
- Argus's locked verifier
- An independent script
- A codeword file anyone can check in the browser
The audit agent agreed with the verdict.
- Sources on record
- 21
- Findings
- 76
- Runtime
- 10h 39m
Check it yourselfYour browser compares all 465 pairs of codewords. Nothing is precomputed.
31 words of length 15 over {0, 1, 2}; any two must differ in at least 10 places.
000011202210121000112022101210001120221012100011202210121000112022101210000120221012100001202210121000011022101210000112221012100001120210121000011202101210000112022012100001120221121000011202210210000112022101100001120221012
011221120202221112211202022210122112020222101221120202221011211202022210112112020222101122120202221011221202022210112211020222101122112202221011221120022210112211202222101122112020221011221120202210112211202022101122112020222
111111111111111and the row of ones
Or check it with your own code: Download the 31 codewords (.txt), one word per line.