Analog Flow [KedArc Quant]

🧭Overview

AnalogFlow is an advanced analogue-based market projection engine that reconstructs future price tendencies by matching current price behavior to historical analogues in the same instrument. Instead of using traditional indicators (moving averages, RSI, or regression), AnalogFlow applies pattern vector similarity analysis — a data-driven technique that identifies historically similar sequences and aggregates their subsequent movements into a smooth, forward-looking curve.

Think of it as a market memory system:
If the current pattern looks like one we’ve seen before, how did price move afterward?

💡 Why AnalogFlow Is Unique

- Pattern-centric — it doesn’t rely on any standard indicator formula; it directly analyzes price movement vectors.
- Adaptive — it learns from the same instrument’s past behavior, making it self-calibrating to volatility and regime shifts.
- Non-repainting — the projection is generated on the latest completed bar and remains fixed until new data is available.
- Noise-resistant — the EMA-Blend engine smooths the projected trajectory, reducing random variance between analogues.

This isn’t a combination of existing tools — it’s an original implementation of analogue-based projection theory


⚙️ Inputs & Configuration

Pattern Bars
Number of bars in the reference pattern window 40
Projection Bars
Number of bars forward to project 30
Search Depth
Number of bars back to look for matching analogues 600
Distance Metric
Comparison method: `Euclidean`, `Manhattan`, or `Cosine` Euclidean
Matches
Number of top analogues to blend (1–5) Top-3
Build Mode
Projection type: `Cumulative`, `MeanStep`, or `EMA-Blend` EMA-Blend
EMA Blend Length
Smoothness of the projected path 15
Normalize Pattern
Enable Z-score normalization for shape matching true
Dissimilarity Mode
If true, finds inverse analogues (mean-reversion analysis) false
Line Color / Width
Style settings for projection curve Blue / 2

📈 How It Works with Past Data
- The system builds a memory bank of patterns from the last N bars
- It compares the latest Pattern Bars segment to each historical segment.
- It selects the Top-K most similar (or dissimilar) analogues.
- For each analogue, it retrieves what happened after that pattern historically.
- It averages or smooths those forward moves into a single composite forecast curve.
- The forecast (blue line) is drawn ahead of the current candle.

🔍 Output Explained

Blue Path:
The weighted mean future trajectory based on historical analogues.
Smoother when EMA-Blend mode is on.

Flat Section:
Indicates low directional consensus or equilibrium across analogues.

Upward / Downward Slope:
Historical tendency toward continuation or reversal following similar conditions.


📆 Recommended Timeframes

Scalping / Short-term
1m–5m --> Short `winLen` (20–30), small `ahead` (10–15)
Swing Trading
15m–1h --> Balanced settings (winLen=40–60, ahead=20–30)
Positional / Multi-Day
4h–1D --> Large windows (winLen=80–120, ahead=30–50)

🌍 Instrument Compatibility

✅ Works seamlessly on:

- Stocks / ETFs
- Indices
- Cryptocurrency
- Commodities (Gold, Crude, etc.)
- Futures & F&O (both intraday and positional)
- Forex

No symbol-specific calibration needed — it self-adapts to volatility.

🧭 How Traders Can Use It

Use Case

Forecast Context
- Identify likely short-term price path or drift direction.
Reversal Detection
- Flip `seekOpp` to true for mean-reversion pattern analysis.
Scenario Comparison
- Observe whether current regime tends to continue or stall.
Momentum Confirmation
- Combine with trend tools (e.g., EMA or MACD) for direction bias.
Backtesting Support
- Compare projected path vs. realized price to evaluate reliability.



❓ FAQ

Q1. Does AnalogFlow repaint?
→ No. It calculates only once per completed bar and projects forward. Future path remains static until a new bar closes.

Q2. Is it a neural network or AI model?
→ Not in the ML sense — it’s a deterministic analogue-matching engine using statistical distance metrics.

Q3. Why does the projection sometimes flatten?
→ That means similar historical setups had no clear consensus in direction (neutral expectation).

Q4. Can I use it for live trading signals?
→ AnalogFlow is not a signal generator. It provides *probabilistic context* for upcoming movement.

Q5. Does higher `scanDepth` improve accuracy?
→ Up to a point. More depth gives more analogues, but too much can dilute recency. Try 400–800.

📘 Glossary

Analogue
A past pattern similar to the current price behavior.
Distance Metric
Mathematical formula for pattern similarity.
Step Vector
Difference between consecutive closing prices.
EMA-Blend
Exponential smoothing of projected path.
Cumulative Mode
Adds sequential historical deltas directly.
Z-Score Normalization
Rescaling to mean 0, variance 1 for shape comparison.

🧩 Summary

AnalogFlow converts the market’s historical echoes into a structured, statistically weighted forward projection. It gives traders a contextual roadmap — not a signal — showing how similar past setups evolved, allowing better-informed entries, exits, and scenario planning across all asset classes.

⚠️ Disclaimer
This script is provided for educational purposes only.
Past performance does not guarantee future results.
Trading involves risk, and users should exercise caution and use proper risk management when applying this strategy.