Risk Disclosure
The risks and modelling limitations you should understand before using Predct's parlay pricing, edge and break-even outputs.
- Effective
- 6 August 2026
- Last updated
- 6 August 2026
Predct simulates prediction-market parlays. Even without real money at stake, it is important to understand what the numbers on screen do and do not mean before you let them influence any decision you make elsewhere.
1. No real money, no real positions
Nothing on Predct is a wager, an order, a contract, or a position in any real market. No transaction is routed anywhere. Virtual credits cannot be deposited, withdrawn, cashed out or redeemed.
2. Model risk
Our pricing engine estimates the joint probability of multiple legs using assumed correlation between markets, clusters and categories. Correlation is estimated, not observed; it changes over time, and it can be badly wrong precisely when it matters most. A parlay that looks favourably priced under our assumptions can be unfavourable under different, equally reasonable ones.
3. Edge and expected value are estimates
Edge, expected value and break-even figures depend entirely on the probabilities fed into them. Small probability errors compound sharply across multiple legs: a parlay is a product, so a 3% error on each of five legs is not a 3% error on the result.
4. The price always carries a margin
Every quote is derived from whichever joint probability is less favourable to you — the market price or the Predct model — and then reduced by the stated platform margin. The mathematical consequence is that the long-run expected return on any simulated parlay is below the amount staked, for every possible combination of legs, sides and correlations. Simulated outcomes are also resolved from the same probabilities the price was built on, so no construction, hedge or correlation pattern produces a positive expected return.
5. Multi-leg structures amplify variance
Adding legs increases the payout multiplier and reduces the probability of winning. Long parlays typically have high variance and long losing streaks even when they carry a genuine edge. Streaks in the simulation are not evidence of skill.
6. Simulated data
Market questions, probabilities, price history and volumes shown here are illustrative and may be generated. Performance in a simulation, including leaderboards, is not indicative of results in any real market.
7. Behavioural risk
Simulated trading can normalise risk-taking and encourage chasing losses. If practising here starts to feel compulsive, or you find yourself carrying these habits into real-money contexts, read our Responsible Play page.
8. Technical risk
Outages, browser storage limits, data loss, calculation errors and display errors can all occur. Do not rely on Predct as a system of record for anything that matters.