Omniscius v2.5
Two years of research. Hundreds of iterations. One mission: data-driven trading signals that actually work.
Omniscius v2.5 is the machine learning engine powering every AlgoZilla trading signal. It represents the culmination of over two years of quantitative research, backtesting, and iterative refinement — from early prototypes with basic indicator crossovers to a sophisticated ensemble classifier trained on 170+ technical features.
The name Omniscius (Latin: “all-knowing”) reflects our ambition, not a claim of perfection. The model doesn’t predict the future — it identifies statistical edges in historical patterns and applies them to current market conditions with rigorous risk management.
What sets Omniscius apart is its methodology. Every signal is generated through walk-forward validation: the model is trained on past data and tested exclusively on unseen future data. This prevents overfitting — the most common failure mode in algorithmic trading. If a strategy only works on the data it was trained on, it’s worthless. Omniscius is validated across three complete Bitcoin halving cycles (2017–2025).
The model is retrained every two weeks, per coin. Market dynamics change — new correlations emerge, old ones decay. Bi-weekly retraining ensures the model adapts to current conditions rather than relying on stale patterns.
We believe in continuous improvement. Omniscius v2.5 is not the final version — it’s the current best. Every iteration is rigorously tested against the previous version before deployment. We never ship changes that don’t demonstrably improve walk-forward performance.
How It Works
170+ Features
Every coin is analyzed across momentum, trend, volume, volatility, sentiment, microstructure, and proprietary regime indicators. Each base indicator is expanded into delta, divergence, and multi-timeframe variants.
Explore indicators →Ensemble Classifier
A LightGBM gradient boosting ensemble combines all features into a single confidence score. The model learns which indicators matter for each coin in each market regime — no static rules.
Regime Detection
A proprietary regime classifier identifies whether the market is in a bull, bear, sideways, or transitional phase. Entry thresholds, exit sensitivity, and risk parameters all adapt automatically.
Walk-Forward Validation
Walk-forward validation is the gold standard in quantitative finance. The concept is simple: train on the past, test on the future. Never use the same data twice.
Omniscius is validated across 3 Bitcoin halving cycles. Each cycle represents fundamentally different market conditions — from the 2017 ICO mania to the 2022 bear market to the 2024 ETF-driven rally. A model that performs well across all three isn’t lucky — it has found genuine statistical edges.
We publish our backtest metrics transparently: Sharpe ratio, win rate, maximum drawdown, and number of trades. These aren’t cherry-picked — they’re averages across all validation folds.
Continuous Improvement
Omniscius v2.5 is the result of hundreds of experiments. From v0.1 (a simple RSI crossover) through v1.x (multi-indicator with regime detection) to v2.5 (full ensemble with walk-forward validation and proprietary features).
Our research process follows strict scientific methodology: one variable per test, independent validation sets, and statistical significance testing before any change is accepted.
Every improvement is documented and version-controlled. When we deploy a new version, we publish the changelog below so you know exactly what changed and why.
Current Performance
*Walk-forward backtest results. Past performance does not guarantee future returns.
*Walk-forward backtest results (2017–2025). Past performance does not guarantee future returns.
Changelog
Every significant model update is documented here. Minor retrains are not listed.
v2.5 — Production Release
Major architecture upgrade. Walk-forward validation across 3 halving cycles. 170+ features including proprietary regime detection. Bi-weekly per-coin retraining. Deployed to production on all 56 coins.
- ✓ Full ensemble classifier with LightGBM
- ✓ Proprietary regime detection engine
- ✓ Walk-forward validation (3 halving cycles)
- ✓ Bi-weekly per-coin retraining
- ✓ 30 coin coverage
- ✓ Real-time Telegram delivery
v1.8 — Regime-Aware Thresholds
Introduced adaptive entry and exit thresholds based on market regime. Bull markets use aggressive entries; bear markets require higher conviction. Reduced false signals by 23%.
v1.5 — Multi-Timeframe Expansion
Added multi-timeframe analysis (2H, 4H, 8H, 12H, 24H) for all base indicators. Feature count expanded from 60 to 170+. Significant improvement in trend-following accuracy.
v1.0 — First Public Release
Initial public release with BTC, ETH, and XRP coverage. Basic ensemble classifier with 60 features. Walk-forward validation introduced. Monthly retraining cycle.
v0.x — Research & Development
Two years of research, prototyping, and backtesting. From simple indicator crossovers (v0.1) through machine learning experiments (v0.3) to the ensemble architecture that became v1.0. Hundreds of experiments, most of them failures — each one teaching us what doesn’t work.
See Omniscius in Action
View live trading signals generated by Omniscius v2.5.
