The short version
A clever, genuinely novel twist on a classic abstract puzzle — the "computer" stalls committing to any secret code as long as your guesses could still be wrong — delivered through a clean, entirely tap-native color-picker interface that we confirmed works correctly across multiple full guess-and-feedback rounds.
Biggest strengths
- Confirmed working end to end via direct testing — tapping colour swatches to fill each peg, then tapping the arrow to submit, correctly produced distinct feedback marks (exact and colour-only matches) across multiple consecutive guesses on a real phone viewport
- A genuinely original mechanic on a very old puzzle format — the adversarial "cheating" AI keeps every prior clue consistent while stalling the actual answer as long as mathematically possible, which is a real constraint-satisfaction trick, not a reskin
- Deep configurability (colour count, peg count, five presets from "Simple" to "Yikes") for players who want to tune the challenge
Biggest weaknesses
- Bare-bones, form-like presentation -- plain HTML checkboxes and text, no visual polish beyond the coloured peg circles
- A few minutes per session by design (per the itch.io listing) -- a tight puzzle toy, not a game with much runtime beyond repeat plays
- The premise, while clever, is a single joke played once; there's no progression or unlocks to chase between sessions
Who it's for: Logic-puzzle fans who've played every Mastermind clone and want one that's actually rigged against them on purpose, honestly.
Our recommendation: recommended
7.3 / 10
CheaterMind takes a very old idea -- guess the secret code, get coloured-peg feedback -- and adds one sharp twist: the "computer" doesn't actually pick a secret code up front. It stays consistent with every clue it's already given you, but delays committing to a final answer for as long as mathematically possible, meaning your first guess is essentially never right, no matter how good it is. The name change from Mastermind is a trademark dodge, same as Shape Clear declining to call itself Candy Crush, not a sign of a lazy clone.
Controls (8.0). We tested this directly rather than trusting the source-reading claim that it's tap-native. Selecting a colour for each peg by tapping a swatch, then tapping the arrow to submit, worked cleanly across two full guesses on a real phone viewport -- and each submission produced genuinely different feedback marks (a mix of exact-match checks and colour-only tildes) rather than a static or broken response. That's real, working gameplay, not just a UI that renders.
Originality (8.0). The underlying constraint-satisfaction "cheating" AI is a legitimately clever idea applied to a public-domain puzzle format that's been cloned thousands of times without anyone doing much new with it. It's the kind of small, sharp idea that's worth a slot on the shelf on its own merits.
Presentation and value (6.0 / 6.5). This won't win any visual awards -- it's plain black text, default browser checkboxes, and a simple grid of coloured circles -- and a session runs only a few minutes by the author's own description. It's a puzzle toy in the purest sense: a clever mechanic wrapped in the minimum viable interface to test it.
For anyone who's grown tired of ordinary Mastermind clones, this is a real, working, and interesting variation -- confirmed to function correctly on the device most players will actually reach for it on.
Pros and cons
Pros
- Confirmed working end to end via direct testing — tapping colour swatches to fill each peg, then tapping the arrow to submit, correctly produced distinct feedback marks (exact and colour-only matches) across multiple consecutive guesses on a real phone viewport
- A genuinely original mechanic on a very old puzzle format — the adversarial "cheating" AI keeps every prior clue consistent while stalling the actual answer as long as mathematically possible, which is a real constraint-satisfaction trick, not a reskin
- Deep configurability (colour count, peg count, five presets from "Simple" to "Yikes") for players who want to tune the challenge
- Fits a real phone viewport cleanly with no layout issues once a viewport tag was added
- MIT licensed, verified directly from the source repository
Cons
- Bare-bones, form-like presentation -- plain HTML checkboxes and text, no visual polish beyond the coloured peg circles
- A few minutes per session by design (per the itch.io listing) -- a tight puzzle toy, not a game with much runtime beyond repeat plays
- The premise, while clever, is a single joke played once; there's no progression or unlocks to chase between sessions