Legal
Prediction and Risk Disclaimer
Draft for legal review. Bracketed items must be completed, and jurisdiction-dependent provisions confirmed by qualified counsel, before this document is relied on. Nothing here is legal advice.
1. What our predictions are
NFL ATS Confluence publishes probabilistic estimates about which side of the point spread is more likely to cover, based on market behaviour, public betting splits, line movement, and historical research. They are estimates, not statements of fact, and not instructions to wager.
2. Historical results and what they mean
Historical results, including backtests and previously observed ATS accuracy, describe past performance under specific conditions. Such results do not guarantee that the model will achieve the same accuracy in future NFL matchups.
Different kinds of results mean different things:
- Historical backtesting applies the current model rules to games that have already been played. It shows what those rules would have produced, with the benefit of hindsight.
- In-sample results are measured on the same games used to choose or tune the model, and tend to look better than genuine out-of-sample performance.
- Out-of-sample validation applies fixed rules to games that played no part in building the model. It is a stronger test, but still historical.
- Live tracked predictions are frozen before kickoff and graded afterwards. They are the most honest record of what the model actually published.
- Future projections are unknown. No result, from any of the categories above, establishes what will happen next.
Every performance figure we publish is tied to the sample it was measured on, and we publish the sample size and conditions alongside it. Where a figure comes from a small sample, it is presented as such.
3. Data and market limitations
- Market prices change continuously, and a line may move between our last observation and your decision.
- Signal strength changes as money arrives; a strong read in the morning may be weak by kickoff.
- Different data providers report different splits, prices, and opening lines for the same game.
- Data feeds may be delayed, incomplete, revised, or temporarily unavailable.
- Predictions may be updated as new information arrives; the record shows when a prediction was frozen.
- Backtests can be affected by selection bias, survivorship of rules that happened to work, stale opening lines, and data corrections made after the fact.
4. Your responsibility
You are responsible for your own decisions. We do not guarantee financial results, returns, winnings, or any specific outcome, and we do not manage money or place wagers.
If you use our information in connection with wagering, that decision is yours. It is subject to the laws of your jurisdiction, the rules of the operator you use, and your own financial circumstances. Only wager what you can afford to lose, and only where wagering is legal.
5. No advisory relationship
We are not your financial adviser, investment adviser, bookmaker, or agent. Nothing in the Service is personalised financial, legal, or tax advice.