Methodology • Transparency • Limitations
Technology
A trustworthy indicator product should explain how it decides, what data it uses, and when it can fail. This page is your “audit trail”.
Pipeline
How signals are produced
High-level steps you can keep stable even as indicators evolve.
1) Normalize price/volume inputs
2) Detect regime (trend vs chop)
3) Compute momentum + volatility context
4) Apply risk guardrails + confidence labels
5) Output signals + explanations
Pseudo-code
Readable by design
if regime == TREND and momentum > threshold: signal = BULLISH reason = [structure, momentum, volatility] else if volatility is HIGH: signal = NEUTRAL guardrail = reduce_size else: signal = WAIT
Flow
Step-by-step signal flow
A straightforward pipeline you can audit.
Inputs
Price + volume streams
↓
Regime
Trend vs chop detection
↓
Signal
Momentum + confirmation
↓
Guardrails
Volatility-aware risk hints
Limitations
When indicators fail
A credible product names its weak spots.
Regime shifts
A strategy tuned for trend can degrade in chop.
Liquidity events
News spikes can invalidate recent structure.
Overfitting
Backtests can lie if assumptions are wrong.