A HUMAN VS. MACHINE EXPERIMENT

Outclick the Oracle.

You think you’re random. This 107-parameter brain disagrees. Prove it wrong in 30 clicks.

LIVE NEURAL NETWORK
UNPREDICTABILITY0
MOVE0 / 30
ESCAPE STREAK0
ORACLE HIT RATE—
I have already made my prediction. Go ahead. Surprise me. Pick any symbol. The Oracle predicts first, then learns from your click.
IT REMEMBERS
    Click, tap, or press 1 / 2 / 3. Surprising choices earn more points.

    Best on this device: —

    OPEN THE BLACK BOX

    Every guess. Every weight.

    Untrained brain

    Try repeating Orbit a few times, then switch. Watch the probability rise, the surprise spike, and the weights adapt.

    1. Predict Your last 3 clicks flow forward through the network. It commits to a bet.
    2. Reveal You click. Points = how unlikely it thought your choice was.
    3. Measure Loss = −ln(probability of your choice). Big surprise, big loss.
    4. Learn Backpropagation nudges all 107 parameters to make that choice likelier next time.

    These are the values before the last click’s update. Δ shows what learning changed afterward. No future guesses are revealed.

    01 · Remember

    x = one-hot(last 3 clicks)

    9 inputs · newest click first · no history = 0

    02 · Mix patterns

    zⱼ = Σᵢ W¹ⱼᵢxᵢ + b¹ⱼ
    hⱼ = tanh(zⱼ)

    8 hidden neurons · 72 weights + 8 biases

    03 · Make a bet

    aₖ = Σⱼ W²ₖⱼhⱼ + b²ₖ
    pₖ = exp(aₖ) / Σ exp(a)

    3 outputs · 24 weights + 3 biases

    04 · Learn from being wrong

    L = −ln(p of your choice)
    W ← W − 0.22 × ∂L/∂W

    Click a symbol to see its loss and a real gradient update.

    Input → hidden: all 72 weights + 8 biases
    Hidden → output: all 24 weights + 3 biases

    Each cell shows its weight and Δ. Cyan = positive; pink = negative; a gold outline marks the biggest changes from your last click. Rounded to 3 decimals; calculations use full precision. O / P / S = Orbit / Prism / Spark.

    Wait. Is it actually learning?

    Yes. Your last three choices become nine inputs to a neural network with eight tanh hidden neurons and three softmax outputs. After each move, backpropagation adjusts its weights using cross-entropy loss. The lines show real weights: cyan is positive, pink is negative; thicker means stronger. After each click you see the forward pass (white sparks) and then the learning pass (gold sparks flowing backward along the weights that changed most).

    The prediction is locked before your click. The diagram and percentages reveal that past prediction, never the next one. An escape means it guessed the wrong symbol. Points = round(100 × (1 − probability of your choice)); up to 3,000 points across 30 moves. Clicking truly at random averages about 2,000 points, so beating 2,000 means you modeled the model better than it modeled you. Streaks are for bragging rights and do not multiply points.

    Everyone gets the same initial brain for a given UTC day. Shared links preserve that challenge and the score to beat; restarting resets all learned weights. There is no timer. This is a playful experiment in pattern learning, not a test of intelligence or a measure of true randomness. Game choices stay in this tab; only your best score and sound preference are saved locally.