Sibyl Quantitative — Risk Disclosure

Last updated: 2026-05-26

This Risk Disclosure describes the principal risks of using the Sibyl Quantitative service (the "Service") at sibyl-ai.com. By using the Service, you acknowledge that you have read, understood, and accepted these risks. This Disclosure is part of, and incorporated by reference into, the Terms of Service.

1. The Service Is Not Financial Advice

This is the most important section. Please read carefully.

Sibyl Quantitative is a research and analytics tool that produces machine-learning-based forecasts of price movement, AI-synthesized commentary ("Sibyl Take"), and related market-context data on a universe of approximately 200 publicly traded tickers (stocks, ETFs, cryptocurrencies).

Sibyl is not, and has never claimed to be:

• A registered investment advisor (RIA) under the Investment Advisers Act of 1940 or any state equivalent.
• A broker-dealer registered with the U.S. Securities and Exchange Commission, FINRA, or any equivalent regulator.
• A futures commission merchant or commodity trading advisor registered with the U.S. Commodity Futures Trading Commission or the National Futures Association.
• A bank, trust company, financial planner, or fiduciary.

Sibyl does not provide:

• Investment advice.
• Personalized investment recommendations tailored to your circumstances.
• Solicitations to buy, sell, or hold any security, derivative, cryptocurrency, or other asset.
• Tax advice, accounting advice, or legal advice.

Every forecast, signal, Sibyl Take synthesis, multi-horizon prediction, consensus tier, expected-move calculation, intraday signal, and any other Content produced by the Service is informational only. All trading and investment decisions are made solely by you, based on your own research, risk tolerance, financial situation, and consultation with your own licensed professionals.

You should consult a licensed financial advisor, tax professional, and/or attorney before acting on anything you see on the Service.

2. Trading Risk

All investing and trading involves risk, including the risk of total loss of capital.

2.1 Market risk

Markets are volatile, sometimes dramatically. Prices can move sharply in any direction, often without warning. The value of any security or asset you trade may fall to zero. Past performance, including any historical track record displayed in the Service (backtest accuracy, live hit rate, forward-test results), is not a guarantee or projection of future results. The future is genuinely uncertain.

2.2 Forecasting risk

Sibyl's forecasts are probabilistic outputs of machine-learning models trained on historical data. Forecasts are not predictions of what will happen; they are statistical guesses about what may happen, weighted by patterns the model identified during training. Specifically:

• Models can be wrong, sometimes dramatically.
• Forecasts may be biased by patterns in training data that no longer apply.
• Model performance during one market regime is not a guarantee of performance in another (regime shift risk).
• Even forecasts marked as "high conviction" or "unanimous" by the Service can be wrong.
• The accuracy figures shown in the Service (backtest accuracy, forward-test hit rate, per-horizon resolved accuracy) are historical measurements; they do not guarantee future accuracy.

2.3 AI synthesis risk

The "Sibyl Take" feature uses a large language model (currently Claude by Anthropic) to synthesize a single sentence of commentary per ticker. This synthesis is generated by a probabilistic AI model and can:

• Be factually wrong about the underlying data.
• Misinterpret the forecast.
• Reflect biases inherited from the AI provider's training data.
• Produce hallucinated or misleading commentary.

Do not rely on Sibyl Take as a substitute for your own analysis. Treat it as one input among many.

2.4 Beta risk

The Service is currently a beta product. Code, models, prompts, and data pipelines are in active development. Bugs, data errors, and incorrect outputs are possible. Notable beta-status risks include:

• Forecast data may be stale, missing, or incorrect for individual tickers.
• The system may be unavailable during scheduled or unscheduled maintenance.
• Auto-Trade may behave unexpectedly during code deployments.
• Model retrains may temporarily change forecast characteristics.

3. Options Trading Risk (When Applicable)

If you trade options based on signals from the Service (manually or via Auto-Trade), you accept additional risks specific to options:

• Time decay (theta) — option value erodes daily as expiration approaches, even if the underlying price does not move.
• Implied-volatility crush — option premiums can collapse rapidly when implied volatility drops, regardless of price direction.
• Assignment risk — short option positions can be assigned early, forcing you to deliver or take delivery of the underlying at the strike price.
• 100% loss possibility — purchased options can expire worthless. You may lose 100% of premium paid.
• Margin and unlimited loss — certain options strategies (naked calls, short straddles) can expose you to losses far greater than your initial investment.
• Complexity — options strategies that look profitable in theory may fail in practice due to bid-ask spreads, liquidity, or execution timing.

Options are not suitable for all investors. You should not trade options unless you understand and accept these risks and have read your brokerage's Options Disclosure Document (typically titled "Characteristics and Risks of Standardized Options").

4. Auto-Trade Risk

The Service offers an optional Auto-Trade feature that, when you enable and configure it, places real orders through your linked Alpaca Securities LLC brokerage account based on signals produced by the Service.

By enabling Auto-Trade, you specifically acknowledge:

• Auto-Trade places real orders for real money. Losses are real.
• Sibyl is not your broker. Sibyl is a software system that generates signals and submits orders to Alpaca on your behalf, based on parameters you configured.
• Order execution, custody, clearing, and settlement are performed by Alpaca Securities LLC under your separate brokerage agreement with them.
• Sibyl does not guarantee that any order will be placed, filled, or filled at any particular price.
• Sibyl does not guarantee that the rules engine will behave correctly under all market conditions. Bugs, edge cases, and unexpected market behavior can cause orders to be placed, modified, or canceled in ways the rules engine does not anticipate.
• Auto-Trade can lose money — sometimes substantial amounts — even when individual signals are correct, because of slippage, fees, timing, and the random distribution of outcomes around any statistical edge.
• You are responsible for monitoring your brokerage account and the Auto-Trade activity, and for intervening if you see unexpected behavior.
• You are responsible for sufficient buying power in your brokerage account.

If you do not understand how Auto-Trade works, what parameters control its behavior, and what risks each parameter exposes you to, you should not enable Auto-Trade.

5. Cryptocurrency Risk (When Applicable)

The Service includes forecasts for certain cryptocurrencies (e.g., BTC-USD, ETH-USD, and others in the v6 universe). Cryptocurrencies carry additional risks:

• Extreme volatility — daily moves of 10%+ are normal.
• Limited regulatory protection compared to traditional securities.
• Custody and security risks at the exchange or wallet level.
• Potential regulatory action that affects price or accessibility.
• 24/7 markets — moves can occur outside regular trading hours when you are unavailable to respond.

6. Data Risk

The Service relies on third-party data feeds (Yahoo Finance, public news RSS, FinBERT sentiment, etc.) that may be:

• Inaccurate or stale.
• Subject to outages or rate limits.
• Reformatted in ways that break our processing.
• Subject to revisions (e.g., historical data corrections).

We make reasonable efforts to verify data, but we do not warrant data accuracy. Critical decisions should not rest solely on a single data point displayed in the Service.

7. Outage and Technology Risk

The Service is a software platform hosted on third-party infrastructure (DigitalOcean for the web app, RunPod for GPU compute). Outages, latency spikes, deployment errors, or third-party failures may:

• Delay or prevent forecasts from refreshing.
• Cause Auto-Trade to miss intended entry or exit signals.
• Make the Service temporarily unavailable.
• Result in stale or missing data displayed in the user interface.

Markets continue to move during outages. Do not rely on the Service to be available at any specific time. Always have a manual fallback for any time-sensitive trading activity.

8. Regulatory and Tax Risk

You are solely responsible for:

• Determining whether your use of the Service complies with the laws and regulations of your jurisdiction.
• Reporting and paying taxes on any gains realized through trading, including any auto-trades placed via Sibyl.
• Maintaining records adequate for tax reporting (Sibyl's Auto-Trade ledger may help but does not replace your brokerage's tax documents).

Regulatory environments evolve. Activities that are legal today may become regulated or prohibited tomorrow. Sibyl is not responsible for changes in your regulatory or tax situation.

9. No Fiduciary Relationship

Your use of the Service does not create a fiduciary, advisory, agency, partnership, joint-venture, or trust relationship between you and Sibyl. Sibyl owes you no fiduciary duty, no duty of suitability, no duty of best execution, and no duty to act in your best interest beyond the contractual obligations spelled out in the Terms of Service.

10. You Are Responsible for Your Decisions

By using the Service, you confirm that:

• You are aware of the risks of trading securities, derivatives, and cryptocurrencies.
• You can afford to lose the money you are trading.
• You will not blame Sibyl, or hold Sibyl responsible, for trading losses you incur.
• You have read, understood, and accepted these risks.
• You have not been induced to trade by any representation made by Sibyl beyond what is in the Terms, Privacy Policy, and this Risk Disclosure.

If you do not accept these risks, stop using the Service immediately.

11. Contact

Questions about this Risk Disclosure: legal@sibyl-ai.com General support: support@sibyl-ai.com

Sibyl Quantitative Los Angeles, California United States of America