HangOpt
A dual-phase word-guessing solver: a greedy probability phase followed by a Minimum Volume Enclosing Ellipsoid (MVEE) optimization phase.
- C++
- Optimization
- MVEE
- Probability
Most word-guessers are frequency tables. HangOpt treats the problem as optimization in two phases:
- Phase 1, greedy. A gamma function picks the letter with the maximum probability sum across all positions, until the first correct letter lands.
- Phase 2, MVEE. It computes the Minimum Volume Enclosing Ellipsoid of the remaining probability distribution and optimizes within it to find the rest.
Pure C++, no external dependencies, a small project built to think carefully about probability, geometry, and constrained search.
Ongoing, I’m still refining the algorithm and the interface.