Understanding virtual signal systems

What is a virtual signal system?

A virtual signal system is a rule-based model that takes in market data and outputs a directional opinion — buy, sell, or hold — with an attached confidence score. The word "virtual" matters: the system operates on paper, using simulated capital and real prices, with no actual money changing hands.

Coinblockers runs exactly this kind of system. The Neural Network composite score you see on the dashboard isn't a crystal ball. It's a weighted average of factors the model believes correlate with short-term price movement — and it's published in full so you can critique it.

How the composite score is built

The score starts from four input categories, each weighted by historical explanatory power:

Each input is z-scored against its own 90-day history before being combined. Z-scoring means a factor that's usually noisy doesn't suddenly dominate just because it spiked.

Why "confidence" doesn't mean "accuracy"

The confidence score (0–100%) represents how strongly the model's input signals agree with each other. A 90% confidence BUY means most of the 40+ sub-signals are pointing up — not that there's a 90% chance the coin goes up.

Historical data shows that high-confidence signals have a modestly better track record than coin flips, especially when multiple timeframes align. But "modestly better than random" is the honest claim — not "reliable alpha."

The role of the neural network validator

Sitting on top of the composite score is a second-stage neural network that acts as a gate. If the composite says BUY but the NN validator's own pattern-match disagrees, the signal gets suppressed. Think of it as a second opinion that can veto but not override.

The NN is trained on the site's own historical signal outcomes, so it learns which types of composite-score patterns have actually led to profitable virtual trades. It's always a work in progress.

What you should actually use this for

The signal system is an educational tool. Use it to:

• Understand which factors are driving a coin's score — click any coin to see the factor breakdown
• See whether technical and social signals agree or diverge — divergence is often more informative than agreement
• Track how virtual trades perform — the Virtual Engine page shows every entry and exit

What you should not do: treat a BUY signal as investment advice, or infer that Coinblockers has any special predictive edge. The system is published precisely so you can evaluate it skeptically.

The honesty constraint

Every rule, weight, and threshold in the scoring system is public. If the methodology page doesn't explain something you see in the score, that's a bug, not a feature. File an issue on GitHub or use the contact form.