The reasoning game where the mechanic IS the skill.
Most "brain training" apps are multiple-choice quizzes with confetti. You either already know the answer or you guess. Guessing isn't learning.
CORTEX is different. You weigh evidence on a tipping scale, build consequence chains, calibrate your probability intuition, and construct persuasive messages in the right order — the interaction itself teaches the concept.
Each mechanic maps directly to a reasoning skill. You don't learn about the concept. You perform it.
Sort evidence cards to tip the scale. Not all evidence is equal.
Tap the exact broken phrase. Precision, not guessing.
Order consequences into a timeline. "Then what? And then what?"
Set a probability slider, then see reality. Your intuition vs math.
Arrange argument blocks in persuasive order. Hook → Problem → Stakes → Solution → Evidence → Action.
Watch flawed reasoning get rewritten in real-time.
Kahoot-style colored answer blocks. Fast pattern recognition.
Pick a specific mechanic to practice, or shuffle everything together.
Every day, teens encounter survivorship bias, false dilemmas, anchoring tricks, and sunk cost traps — and they have zero training to detect them.
The apps that try to help are quizzes. Read a paragraph, pick A/B/C/D. If you don't already know the concept, you're just guessing. The "learning" happens in an explanation you read after you already got it wrong.
CORTEX flips this. The game mechanic teaches the concept. Sorting evidence on a scale is evidence weighing. Building a consequence chain is systems thinking. Setting a probability slider is calibration. You experience the skill before you're ever told its name.
Survivorship Bias · Anecdote ≠ Evidence · Ad Hominem · False Dilemma · Cherry Picking · Correlation ≠ Causation · Confounding Variables · Sunk Cost Fallacy · Confirmation Bias · Anchoring Bias · Availability Heuristic · Bandwagon Effect · Base Rate Neglect · Gambler's Fallacy · Goodhart's Law · Second-Order Effects · Cobra Effect · Perverse Incentives · Appeal to Popularity · Persuasive Structure · Evidence Hierarchy · Audience-First Persuasion · Hook → Stakes → Solution
From Duolingo — 3D buttons with bottom shadows, mascot with emotional states and speech bubbles, hearts, progress bar, Nunito font, minimal text.
From Kahoot — Each answer is a distinct bold color (red ●, blue ◆, green ▲, yellow ★). The color is the identity.
From Brilliant — "Aha moment" reveal with blur-to-sharp animation. Mascot as guide, not decoration.
What's new — Nobody combined these into a game where the interaction mechanic and the reasoning skill are the same thing.
git clone https://github.com/moranetz/iq.git
cd iq
npm install
npm run devOpen http://localhost:5173
npm run builddist/ folder is ready for Vercel, Netlify, or any static host.
React 18 · Vite · Canvas API (animated mascot) · Web Audio API (synthesized sounds, no audio files) · Web Vibration API (haptics on mobile) · CSS animations (confetti, screen shake, combo system)
Zero dependencies beyond React. 900 lines. One file.
- Persistent storage — XP, level, streaks survive across sessions
- API-generated questions — infinite fresh scenarios
- Daily challenge with cross-session streaks
- Spaced repetition — concepts you missed resurface more
- Social sharing — challenge a friend
- Drag-and-drop — true drag for chain building and evidence sorting
- Native iOS — SwiftUI port with Core Haptics
CORTEX — think smarter ✨🧠