How to Optimally Solve Sanuli
I decided to vibe code a bit and added Sanuli Solver+, an algorithmic enhancement to Sanuli Solver for solving Sanuli (the Finnish Wordle clone) optimally. Instead of just brute-forcing a solution, the tool now suggests the best next guesses based on the current game stage.
Three Strategies
As you play and get feedback on your guesses, the number of possible words that match your clues decreases. The optimal way to solve Sanuli depends on how many candidate words remain:
Entropy (100+ words): Measure information gain to distinguish between possibilities. This works best when you have many options, as it efficiently cuts the search space by half.
Expected Remaining Words (20-100 words): Calculate how many candidates will remain after each guess and pick the one that minimizes this count. This is faster to compute than entropy and works well for mid-game situations.
Minimax (Under 20 words): Minimize the worst-case scenario. Once you're down to few options, this guarantees you can solve the puzzle within the remaining guesses, even if you get unlucky with feedback patterns.
Finnish Advantages
Standard Wordle algorithms miss Finnish's unique structure. These linguistic features matter when solving optimally:
- Vowel Harmony: Back vowels {a, o, u} don't mix with front vowels {ä, ö, y}. Testing both harmonies early eliminates large candidate pools.
- Gemination: Doubled consonants and vowels appear frequently. Testing duplicates earlier than in English strategies provides more information.
- Rare Consonant Clusters: Multiple consonants rarely start words, reducing plausible patterns.
Optimal Play
Start with high-entropy openers like KASTI or KILTA to test vowel harmony. Test gemination early, then shift to minimax pruning once candidates drop below 20. Recommended first guesses: KASTI • KILTA • KILSA • KARSI • SILTA.