.memlog.md 3.5 KB
topic: Wordle clone word-guessing game goal: Create a Wordle-like game: 5-letter words, 6 attempts, green/yellow/gray feedback — with unique twists mode: partner status: complete
updated: 2026-07-07T14:45
- (decision by user) 5-letter words are locked in — most common in language, right balance of challenge
- (idea by user) Result: show number of attempts taken
- (idea by user) Result: ranking among all players who guessed the same word
- (idea by user) Result: lifetime statistics with histogram by number of attempts
- (idea by user) Adaptive difficulty: track per-word average attempts globally, start mid-complexity, adjust word difficulty based on player performance vs averages
- (idea by coach) Words self-sort into difficulty tiers based on global average attempts — no pre-labeling needed
- (idea by coach) Hidden MMR/skill rating that drifts with performance vs global averages, surfaceable as badges
- (idea by user) Streak tracking with quality weighting — streaks on harder words are more impressive
- (idea by user) Standardized hint on 6th attempt for all players — fair since good players solve in ≤4, doesn't distort stats
- (idea by user) Optional weaker hint on 5th attempt — standardized across all players to keep stats comparable
- (idea by coach) Post-solve: show word definition, etymology, example sentence — learning layer on top of game
- (idea by user) Show a picture/illustration for simple words after solving
- (decision by user) Keyboard highlighting stays — it's only a visual helper, no extra information
- (decision by user) Green/yellow/gray feedback stays — necessary for gauging word distance
- (idea by user) Computer-guess subgame: player picks word, computer solves — teaches strategy and measures word complexity
- (idea by user) Real-time 2-player mode: post-v1, both guess same word simultaneously
- (decision by user) No penalty for fast guessing — doesn't add value
- (technique) finished SCAMPER Method
- (idea by user) Daily word first, then prompt to continue with more words — infinite play after the daily event
- (decision by user) No user-submitted words — can't trust community adequacy
- (idea by user) Variable word length mode (3-7 letters) — post-v1, daily word stays 5 letters
- (idea by user) Position-blind mode (yellow/gray only, no green) — possible hard-mode variant post-v1
- (technique) finished What If Scenarios
- (idea by user) Rotating target word each guess — chaotic variant, possible post-v1
- (decision by user) Word bank: only words with recent internet usage frequency (last ~5 years) — no obscure/archaic words
- (idea by user) Percentile ranking instead of absolute rank — more meaningful (top 15% vs #14,327)
- (idea by user) Pre-compute computer guesses with known algorithms as baseline — gives difficulty reference even without players, plus context for human performance
- (technique) finished Worst Possible Idea
- (decision by user) Green/yellow/gray position feedback is essential for 5-letter/6-attempt constraints — numeric score alone would need too many guesses
- (technique) finished Kill the Crown Jewel
- (insight) Computer baseline + adaptive difficulty + word curation are one system — same solver engine drives ranking, tiering, and solvability validation
- (insight) Percentile ranking + quality streaks = leaderboard without toxicity
- (insight) Post-solve learning + computer subgame: show optimal solver replay as teaching moment alongside definition
- (insight) Standardized late-game hints are a retention mechanic — catch failing players fairly