Infinity Algo AI System Disclosure

SECTION 1 – PURPOSE AND FUNCTION OF THE AI SYSTEM

Provider: Infinity Algo, Czech Republic. For questions or concerns about this AI system, contact us at [email protected].

The "AI Clustering" feature in Infinity Algo is an optimization algorithm, not a predictive one. Its purpose is to help traders adapt to changing market conditions by analyzing recent historical data to find which indicator settings would have performed most effectively in the immediate past.

It functions as an automated strategy tester that runs in the background of your TradingView chart. The system does not engage in "self-learning" or evolve on its own; it simply performs a new, isolated analysis during each cycle. The system does not create or manipulate audio, image, or video content; Art 52 deep-fake rules therefore do not apply.

Regulatory Classification: This AI system falls outside the scope of high-risk AI systems as defined in Annex III of the EU AI Act. It does not make decisions about credit, employment, law enforcement, biometric identification, critical infrastructure, education, or other high-risk areas specified in the regulation. It is purely a technical analysis tool for financial market data.

This disclosure fulfills our transparency obligations under Article 50 of the EU AI Act for limited-risk AI systems. Global Context: While this disclosure focuses on EU requirements, we apply similar transparency standards worldwide. Our risk controls also align with the voluntary NIST AI Risk Management Framework 1.0 and the UK's pro-innovation AI principles. Users in other jurisdictions should review local AI and financial regulations as applicable.

SECTION 2 – HOW THE AI SYSTEM WORKS

When the "AI Clustering" feature is enabled, the indicator performs the following steps periodically. This recalculation is triggered once every `X` bars, based on the "AI Update Frequency" setting you control (default is 50 bars).

Technical Process: The system uses K-means clustering algorithm for initial parameter grouping on historical OHLCV data, followed by rule-based scoring and optimization. All signal generation thereafter is rule-based. For more details on our AI logic, please see our Disclaimer Section 6.

1. Parameter Simulation: The system simulates hundreds of trading strategies by testing different combinations of core indicator settings against recent historical price data (by default, the last 500 bars). The parameters tested include:
   • Sensitivity: Testing periods within user-selected ranges (e.g., "Fast (10-14)", "Balanced (10-20)", etc.).
   • Upper/Lower Thresholds: Testing a predefined set of values for buy and sell thresholds (e.g., Upper: `[66, 68, 70, 74, 76]`, Lower: `[26, 28, 30, 34, 36]`).
2. Performance Scoring: Each simulated combination is scored based on a user-selected performance metric. This allows you to define what "best" means for your trading style. The available metrics are:
   • Profit Factor * Sqrt(Trades)
   • Average Profit
   • Win Rate
   • Total Profit
   • Profit Factor
3. Optimal Configuration Selection: The system identifies the single combination of settings that achieved the highest score on the chosen performance metric over the recent historical data.
4. Signal Adaptation: The indicator then uses this "optimal" configuration to generate the "AI" Buy and Sell signals on your live chart until the next recalculation cycle.

The AI Dashboard on the chart displays the settings that the system has currently identified as optimal. Because this entire process runs locally within TradingView's environment, complex simulations (e.g., low "AI Update Frequency") can increase chart redraw times or cause indicator timeouts. We recommend using this feature on timeframes of 15 minutes or higher for a smoother experience.

SECTION 2.1 - KEY PERFORMANCE CHARACTERISTICS & HUMAN OVERSIGHT

The performance of the AI optimizer is characterized by the following typical metrics:

• Mean Profit Factor: On a 4-hour BTC/USD chart, the default settings have historically identified configurations with a simulated profit factor of approximately 1.35 over the lookback period. This metric is illustrative only based on historical backtesting and does not predict future performance. Real-time profit factors may deviate significantly depending on market conditions, asset volatility, and timing of trades. Past performance is not indicative of future results.
• Optimization Speed: A standard optimization cycle on a 4-hour chart typically completes in approximately 0.8 seconds under normal conditions.
• Reliability: The operational failure-rate (e.g., script timeout errors) is below 1% on timeframes of 15 minutes or higher under typical market conditions.
• Market Condition Variability: Performance characteristics may vary significantly across different market conditions. The AI optimizer may perform differently in trending vs. ranging markets, high vs. low volatility periods, and during major news events or market disruptions. These variations can result in substantial differences from historical performance metrics.

IMPORTANT PERFORMANCE DISCLAIMER: All performance metrics listed above are based on historical simulations and backtesting only. They do not account for real-world trading costs such as spreads, slippage, commissions, or network delays. There is no guarantee that the AI system will achieve similar performance in live trading conditions, and substantial losses may occur.

Human Oversight and Control

You always remain in full control of the AI system. You can:

• Disable the AI: Turn off "AI Clustering" at any time using the "AI Enabled" toggle in the indicator's settings.
• Control Frequency: Adjust the "AI Update Frequency" to determine how often the system recalculates.
• Override Parameters: Manually change any indicator setting to override the AI's suggestions at any time. We recommend you review the AI's chosen settings after each recalculation to ensure they align with your strategy.

Clear AI Identification: When AI Clustering is active, all trading signals are clearly labeled as "AI Buy" or "AI Sell" on your chart. Additionally, the optional AI Dashboard displays the current AI-optimized settings, ensuring you always know when AI-driven parameters are in use.

Signal-Only System: The Infinity Algo indicator is a signal generation tool only. It displays visual alerts on your TradingView chart but has no capability to access your trading account, place orders, or execute trades. Any trade execution—whether manual or automated through third-party tools—is entirely separate from our system.

Over-Reliance Warning: Do not rely solely on AI-generated signals for trading decisions. Over-reliance on AI without applying human judgment, independent analysis, and proper risk management may significantly increase trading risks. Always combine AI signals with your own market analysis and strategy.

SECTION 3 – TRAINING DATA

The AI system does not use external or private data sets. The "training data" consists solely of the historical OHLCV (Open, High, Low, Close, Volume) market data available directly on the user's TradingView chart, typically from the last 500 bars.

The optimizer processes no personal data within the meaning of GDPR art 4(1). Crucially, all computations occur locally in your TradingView environment—no data is ever sent to Infinity Algo servers or any external "black box" model. The system analyzes historical data to find patterns that can inform the optimal indicator settings. No personal user data is ever accessed or used in this process.

SECTION 4 – LIMITATIONS AND RISKS

It is crucial for users to understand the limitations of this AI system:

Historical Limitation (Past ≠ Future)

• The system's recommendations are based entirely on what would have worked in the past. Past performance is not an indicator or guarantee of future results, and there is no guarantee that the selected "optimal" settings will be profitable in live market conditions.

Non-Predictive Nature

• The AI does not predict future price movements. It is a reactive tool that optimizes settings based on historical analysis. It is not a "self-learning" or continuously evolving model.

Curve-Fitting Dangers

• As with any backtesting, there is a risk that the settings identified as "optimal" are simply "curve-fitted" to past data and may not perform well in the future. This risk is significantly higher in illiquid markets or during anomalous events (e.g., flash crashes, market manipulation, extreme volatility spikes), where historical data may not represent typical market conditions.
• Low-Volume Asset Risk: The curve-fitting risk increases substantially when using the AI system on low-volume or thinly traded assets, where historical patterns may be less reliable and more susceptible to random noise or market manipulation.
• Market Regime Changes: The AI system may fail to adapt quickly to fundamental changes in market structure, regulatory environment, or macroeconomic conditions that make historical patterns obsolete.

Exclusion of Real-World Costs

• Because the AI's backtests are simulations, they do not include factors like exchange fees, slippage, or network latency. Live trading results may differ materially from the simulated outcomes.

Third-Party Automation & Foreseeable Misuse Warning

Important: While our AI system itself does not execute trades, users may choose to connect our signals to third-party automation tools (e.g., 3Commas, WunderTrading, custom bots). If you do this:

• You are fully responsible for the automated execution and its consequences
• The AI's historical optimization was NOT designed for unsupervised automated trading
• You MUST implement independent risk-management checks in your automation
• Unsupervised automation may lead to over-trading, excessive fees, and liquidation risk
• We strongly recommend manual review of all signals before execution

The AI system is a tool for analysis and decision-support, not a replacement for proper risk management and independent judgment. All trading decisions and their outcomes are the sole responsibility of the user. For a full breakdown of risks, please see our Disclaimer.

SECTION 5 – RECOMMENDED BEST PRACTICES

To use the AI Clustering feature effectively, we recommend the following risk control measures:

• Validate on a Demo Account: Always test the AI-optimized settings on a paper trading or demo account before deploying them with live capital.
• Use Independent Risk Management: Continue to use standard risk management tools, such as stop-loss orders and appropriate position sizing, regardless of how "optimal" the AI settings appear.
• Combine with Independent Analysis: Do not rely solely on AI signals—always combine them with your own market analysis, fundamental research, and trading strategy.
• Re-evaluate Periodically: Market conditions change. Periodically re-evaluate the AI's settings and the lookback window to ensure outdated data is not biasing the results, especially after major market events.
• Monitor Performance Closely: Track the real-world performance of AI-generated signals compared to historical backtests to identify when the system may be failing to adapt to current market conditions.
• Understand Market Context: Be aware of current market conditions (trending vs. ranging, high vs. low volatility) and how they might affect the AI's performance relative to its historical optimization period.
• Avoid Over-Optimization: Don't constantly adjust settings based on short-term performance. Give the AI system time to perform across different market conditions before making significant changes.

Environmental Consideration: The local computation may increase device energy use and processing load during optimizations. Consider this when using the system frequently on mobile devices or older hardware.

SECTION 6 – BIAS MITIGATION AND TECHNICAL DETAILS

Bias Mitigation: Because the AI optimizer only evaluates parameter combinations against objective, mathematical performance metrics (e.g., Profit Factor, Win Rate), it is not subject to the demographic, social, or cognitive biases that can affect human-labeled data sets. However, the system may still exhibit market-specific biases:

• Historical Bias: The system inherently favors patterns that worked in the recent past, which may not persist in future market conditions
• Asset-Specific Bias: Performance may vary significantly between different asset classes (crypto vs. forex vs. stocks) due to varying market microstructures
• Volatility Bias: The system may over-optimize for recent volatility patterns, potentially failing during regime changes
• Volume Bias: On low-volume assets, the system may identify false patterns due to limited data or market inefficiencies

Data Integrity and Feed Dependencies: The AI system's effectiveness depends entirely on the quality and completeness of TradingView's data feed. Data irregularities, such as historical corrections, missing bars, delayed data, or incomplete market sessions may significantly impact optimization results. Users should verify data quality and completeness before relying on AI-generated settings, particularly when switching between data providers or during market hours when data feeds may be delayed.

If you use custom data imports in TradingView, ensure their accuracy, as the AI relies entirely on the chart's available data for optimization.

Pine Script Version: The AI Clustering feature is implemented in Pine Script v6. Please ensure your TradingView chart environment is up to date to avoid potential compatibility issues.

Post-Deployment Monitoring

We continuously log anonymised optimisation failures (e.g., script timeouts) to monitor the system's performance across different market conditions. We commit to patching any material issues that are identified within 30 days of confirmation, unless a critical vulnerability requires faster action. A public changelog of updates is available at docs.infinityalgo.com/.

User Reporting: Report suspected biases, errors, or malfunctions via [email protected]; we investigate all reports as part of our ongoing monitoring and improvement process.

SECTION 7 – CONTACT AND COMPLAINTS

If you have any questions about our AI system, or if you believe the system has malfunctioned or violated your rights, please contact our dedicated AI compliance address at [email protected]. We commit to acknowledging receipt of your complaint within 72 hours and providing a substantive response within 14 days.

SECTION 8 – SEVERABILITY

If any part of this disclosure is deemed unlawful, void, or unenforceable, such provision shall nonetheless be enforceable to the fullest extent permitted by applicable law. The unenforceable portion shall be deemed to be severed from this disclosure; such determination shall not affect the validity and enforceability of any other remaining provisions.