INVESTOR FAQ

The foreign exchange market trades $7.5 trillion daily, the most liquid market in the world. It operates 24/5 with no single point of failure. FX pairs exhibit measurable equilibrium convergence on hourly timeframes, driven by central bank policy anchoring, carry trade dynamics, and institutional rebalancing. This structural convergence behavior is a well-documented phenomenon in academic literature.

The system models price as a stochastic equilibrium process from statistical physics. When price deviates far from its dynamic equilibrium, the system computes the exact first-passage probability, the likelihood of reaching the take-profit before the stop-loss. Trades only execute when that probability is high. This is not a heuristic. It is an analytical probability computation from stochastic calculus.

Three structural advantages: (1) Theory-driven, not data-mined. Indicators from stochastic processes, information theory, and state estimation transfer across regimes. (2) Path-dependent gating. Path-dependent risk decomposition provides 2.7x better filtering than direct PnL prediction. (3) Ensemble diversity. Five strategies with low signal correlation, selected by an evolutionary algorithm with enforced diversity constraints.

Six independent layers of validation: (1) Theory-driven indicators from established mathematics, not invented for trading. (2) Tightly constrained parameter space, not millions of parameters. (3) Strict temporal train/test split with out-of-sample performance exceeding in-sample. (4) Combinatorial Purged Cross-Validation: PBO = 0.0000. (5) Deflated Sharpe Ratio passes at 1% significance after correcting for thousands of trials. (6) Walk-forward CV: 100% of folds positive.

PBO (Probability of Backtest Overfitting) measures the chance that the best-performing in-sample strategy would underperform out-of-sample. PBO = 0.0000 means that across all combinatorial partitions of the data, the in-sample best strategy is positive out-of-sample in 100% of tests. In-sample/out-of-sample Spearman rank correlation: near-perfect. This is the strongest possible evidence against overfitting from a combinatorial perspective.

All trading systems carry risk. Key risks include: regime changes that reduce equilibrium convergence behavior, execution risk during volatile periods, model degradation over time, and the backtest-live gap. The system mitigates these through: 12-layer gate pipeline, quarter-Kelly position sizing (75% reduction in position volatility), circuit breakers, session P&L caps, and continuous monitoring. P99 max drawdown remains within strict limits across all pairs. The system is designed to survive losses, not avoid them.

The system is transparent about this. Live trading began February 2026. A specific exit mechanism was identified as the primary gap. Root cause isolated and resolved through dedicated research sessions. The fix has been validated with zero recurrence. This honest reporting demonstrates the rigor of the research process.

All performance metrics are net of 2.0 pips round-trip cost per trade (1.5 pips spread + 0.5 pips slippage). This is conservative for major pairs (EURUSD spread is typically 0.3 to 0.5 pips) but realistic for crosses during volatile hours. If your edge disappears when you add realistic costs, you never had an edge.

Neural Predictiva was designed, researched, engineered, and deployed by Abdalla Elzedy. 310+ research sessions, 11,760+ lines of research journal, 80+ analysis files. The system is seeking seed investment to scale the operation with dedicated quant research, engineering, and compliance capabilities.

Current: 5 FX pairs live, backtest-to-live gap identified and actively being closed through predictor recalibration. Near-term (1 to 4 weeks): Hold buffer walk-forward validation, shadow deployment, accumulating statistically meaningful post-deployment trade counts. Medium-term (1 to 3 months): Additional pairs, external data integration (options volatility surfaces, order flow analytics), macro filter threshold calibration using live data. Long-term (3 to 12 months): Multi-asset expansion (equity index futures, commodities), institutional infrastructure (co-location, compliance framework), advanced ML pipeline.

Neural Predictiva is currently a pre-seed, pre-entity project. The legal structure (LP, LLC, or fund vehicle) will be determined in consultation with investors and legal counsel to optimize for the investment thesis and regulatory requirements. The founder is open to structuring the entity in a way that aligns investor and operator incentives.

There is no fixed minimum at this stage. Neural Predictiva is seeking seed capital from sophisticated investors who understand quantitative trading and are comfortable with early-stage risk. Investment terms are negotiated individually. Reach out to discuss specifics.

The system's backtested Sharpe ratio exceeds 13.5 across 13+ years of out-of-sample data, with zero losing calendar years. However, past performance does not guarantee future results. The backtest-to-live gap is being actively closed. Detailed performance projections, capital deployment scenarios, and return simulations are available under NDA.

Three priorities. (1) Infrastructure: dedicated cloud servers, co-location, redundant data feeds, and institutional-grade execution. (2) Team: dedicated quant researcher, ML engineer, and compliance officer. (3) Capital deployment: scaling from current account size to institutional AUM across additional pairs and asset classes. A detailed use-of-funds breakdown is available upon request.

Fee structure is not yet finalized and will be established as part of the fundraising process. The founder's preference is for an alignment-first structure: meaningful skin in the game, performance fees tied to high-water marks, and hurdle rates that ensure investors are compensated before the operator. Specific terms will be negotiated.

Neural Predictiva is not currently registered as an investment adviser or fund manager. As part of the fundraising process, the appropriate regulatory filings will be completed in consultation with securities counsel. The roadmap includes building a full compliance framework as a near-term infrastructure priority.

The system operates a fully built macro intelligence layer across all daemons. It tracks four active trading sessions (Sydney, Tokyo, London, New York) with open/close countdowns and overlap detection, runs live regime state classification per pair, and maintains a binary macro filter gate that can block entries during extreme macro dislocation. The dashboard also surfaces COT positioning data, central bank rate decision calendars, commodity-currency correlations, and rate path scenarios. The core signal generation remains theory-driven (stochastic equilibrium process), not macro-dependent. Macro indicators were tested extensively during development and produced zero marginal improvement to signal quality, confirming the system's alpha comes from the mathematical structure, not from macro timing. The macro layer serves the operator with real-time market context and provides a conservative safety filter that rarely activates under normal conditions.