A price
is a claim.
A prediction-market contract pays $1 if the event happens and $0 if it does not. The price is therefore a claim about how likely the event is.
Break-even is simple: over many trades, the event must happen at least as often as the price implies just to cover your cost, and slightly more often once fees are included.
The favorite-longshot
tilt.
If prices were perfectly accurate, a 10¢ contract would hit 10% of the time. Studies of betting markets have often found a tilt: longshots hit less often than their prices imply, and favorites hit slightly more often. Findings for prediction markets are mixed and vary by market and period.
At 1.00, prices are perfectly calibrated. Higher values mean cheap contracts hit less often than their prices imply. This is a simple model, so treat its shape as illustrative.
A steady
habit.
Now suppose you buy the same contract again and again. Each thin gray line below is one possible sequence of results under the settings above; the dashed line is the average result.
The stats come from 300 simulated runs. Your own results would follow just one path, and short runs are dominated by luck, which is why a small average loss per trade is easy to miss.
Fees hit
cheap contracts hardest.
A fixed fee per contract is a bigger share of a cheap contract's price. The table shows the expected result per $100 staked, using your tilt and fee settings.
Share
your settings.
The link includes your slider settings, so anyone who opens it sees your exact scenario.
ML: Simulations
Three quick experiments with randomness. Everything runs in your browser: no accounts, no tracking, no market data.
Dice Simulator
Fair six-sided dice. Switching between 1 and 2 dice starts a new experiment.
Small samples bounce around: 10 rolls can look lopsided purely by chance. As trials grow, observed frequencies tend to settle near the theoretical probabilities (the law of large numbers), though individual counts never become exactly equal. With two dice, 7 is the most likely sum (6 of 36 combinations) while 2 and 12 are the least likely (1 of 36 each).
Observed vs. theoretical (% of trials)
ObservedTheoretical
Probability Wheel
Set the chance of red; blue gets the rest. Each spin is independent of every spin before it.
Moving the slider after you have results resets the experiment, because expected counts depend on the probability. Repeating the same experiment gives different results: that variation is normal randomness, not a bug. Batches are computed instantly without animating each spin.
Observed vs. expected (counts)
ObservedExpected
Catch the Threshold
An arcade game with a fictional, randomly generated price line. It is not market data, not a trading simulator, and not a prediction of any real price. Press Catch! (or Space / Enter, or tap the chart) the moment the line first crosses the dashed target. 5 rounds; later rounds are faster and choppier.
How scoring works. Each round is worth up to 100 points. Press at the moment the line first crosses the dashed target: the closer you are, the more you score. Closeness is measured against the length of the round.
✔ Perfect: 100 points, within 1.5% · ✔ Great: 75 points, within 4% · ✔ Good: 50 points, within 8% · ~ Close: 25 points, within 15% · ✖ Miss: 0 points, further off or no press.
After each round you will see how many milliseconds early or late you were. Leaving this tab resets the game. If your device is set to reduce motion, rounds run slower.
These simulations use random numbers generated in your browser and fictional data. They are for informational and educational purposes only, not financial, investment, or legal advice.