Skip to content

Latest commit

 

History

History
31 lines (20 loc) · 1.54 KB

README.md

File metadata and controls

31 lines (20 loc) · 1.54 KB

Ideal Wordle Sequence

Given the desire to cover the entire alphabet efficiently (finding correct letters without redundancy), what words are absolutely best in sequence? This repo answers this question by weighing each word in the subset by the ubiquity of its letters, as well as the local ubiquity of each letter in each position. This sorts words into four general, rank-discrete categories:

  1. (highest) Common letters in the right place
  2. (higher) Common letters in the wrong place
  3. (moderate) Uncommon letters in the right place
  4. (low) Uncommon letters in the wrong place

This ranking naturally arises from double-counting ubiquity within the local ubiquity of each letter (more common letters are on average more common in a given position than uncommmon letters). To penalize redundancy, letters from words chosen by highest rank have their weight set to zero, and the ranking is recalculated. This has the nice feature of penalizing local redundancy while allowing for nonlocal redundancy, were it in an otherwise high-ranking word.

This routine finds the following set to be absolutely best:

  1. Tares
  2. Colin
  3. Dumpy
  4. Bight
  5. Flake

given words bekah and frike are forbidden. This constitutes what I believe to be the best sequence to use, particularly in games like squabble and sedecordle. Do what you will with this massively overpowered knowledge.


License

See LICENSE.md


Cheers,

Servinjesus1